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="#globalvars">Global Variables</a></li>
26 <li><a href="#functionstructure">Functions</a></li>
27 <li><a href="#paramattrs">Parameter Attributes</a></li>
28 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
29 <li><a href="#datalayout">Data Layout</a></li>
32 <li><a href="#typesystem">Type System</a>
34 <li><a href="#t_primitive">Primitive Types</a>
36 <li><a href="#t_classifications">Type Classifications</a></li>
39 <li><a href="#t_derived">Derived Types</a>
41 <li><a href="#t_array">Array Type</a></li>
42 <li><a href="#t_function">Function Type</a></li>
43 <li><a href="#t_pointer">Pointer Type</a></li>
44 <li><a href="#t_struct">Structure Type</a></li>
45 <li><a href="#t_pstruct">Packed Structure Type</a></li>
46 <li><a href="#t_vector">Vector Type</a></li>
47 <li><a href="#t_opaque">Opaque Type</a></li>
52 <li><a href="#constants">Constants</a>
54 <li><a href="#simpleconstants">Simple Constants</a>
55 <li><a href="#aggregateconstants">Aggregate Constants</a>
56 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
57 <li><a href="#undefvalues">Undefined Values</a>
58 <li><a href="#constantexprs">Constant Expressions</a>
61 <li><a href="#othervalues">Other Values</a>
63 <li><a href="#inlineasm">Inline Assembler Expressions</a>
66 <li><a href="#instref">Instruction Reference</a>
68 <li><a href="#terminators">Terminator Instructions</a>
70 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
71 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
72 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
73 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
74 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
75 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
78 <li><a href="#binaryops">Binary Operations</a>
80 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
81 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
82 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
83 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
84 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
85 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
86 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
87 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
88 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
91 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
93 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
94 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
95 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
96 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
97 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
98 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
101 <li><a href="#vectorops">Vector Operations</a>
103 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
104 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
105 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
108 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
110 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
111 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
112 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
113 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
114 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
115 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
118 <li><a href="#convertops">Conversion Operations</a>
120 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
121 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
122 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
123 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
124 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
125 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
126 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
127 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
128 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
129 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
130 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
131 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
133 <li><a href="#otherops">Other Operations</a>
135 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
136 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
137 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
138 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
139 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
140 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
145 <li><a href="#intrinsics">Intrinsic Functions</a>
147 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
149 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
150 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
151 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
154 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
156 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
157 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
158 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
161 <li><a href="#int_codegen">Code Generator Intrinsics</a>
163 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
164 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
165 <li><a href="#i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
166 <li><a href="#i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
167 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
168 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
169 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
172 <li><a href="#int_libc">Standard C Library Intrinsics</a>
174 <li><a href="#i_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
175 <li><a href="#i_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
176 <li><a href="#i_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
177 <li><a href="#i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
178 <li><a href="#i_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
181 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
183 <li><a href="#i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
184 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
185 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
186 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
189 <li><a href="#int_debugger">Debugger intrinsics</a></li>
194 <div class="doc_author">
195 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
196 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
199 <!-- *********************************************************************** -->
200 <div class="doc_section"> <a name="abstract">Abstract </a></div>
201 <!-- *********************************************************************** -->
203 <div class="doc_text">
204 <p>This document is a reference manual for the LLVM assembly language.
205 LLVM is an SSA based representation that provides type safety,
206 low-level operations, flexibility, and the capability of representing
207 'all' high-level languages cleanly. It is the common code
208 representation used throughout all phases of the LLVM compilation
212 <!-- *********************************************************************** -->
213 <div class="doc_section"> <a name="introduction">Introduction</a> </div>
214 <!-- *********************************************************************** -->
216 <div class="doc_text">
218 <p>The LLVM code representation is designed to be used in three
219 different forms: as an in-memory compiler IR, as an on-disk bytecode
220 representation (suitable for fast loading by a Just-In-Time compiler),
221 and as a human readable assembly language representation. This allows
222 LLVM to provide a powerful intermediate representation for efficient
223 compiler transformations and analysis, while providing a natural means
224 to debug and visualize the transformations. The three different forms
225 of LLVM are all equivalent. This document describes the human readable
226 representation and notation.</p>
228 <p>The LLVM representation aims to be light-weight and low-level
229 while being expressive, typed, and extensible at the same time. It
230 aims to be a "universal IR" of sorts, by being at a low enough level
231 that high-level ideas may be cleanly mapped to it (similar to how
232 microprocessors are "universal IR's", allowing many source languages to
233 be mapped to them). By providing type information, LLVM can be used as
234 the target of optimizations: for example, through pointer analysis, it
235 can be proven that a C automatic variable is never accessed outside of
236 the current function... allowing it to be promoted to a simple SSA
237 value instead of a memory location.</p>
241 <!-- _______________________________________________________________________ -->
242 <div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
244 <div class="doc_text">
246 <p>It is important to note that this document describes 'well formed'
247 LLVM assembly language. There is a difference between what the parser
248 accepts and what is considered 'well formed'. For example, the
249 following instruction is syntactically okay, but not well formed:</p>
252 %x = <a href="#i_add">add</a> i32 1, %x
255 <p>...because the definition of <tt>%x</tt> does not dominate all of
256 its uses. The LLVM infrastructure provides a verification pass that may
257 be used to verify that an LLVM module is well formed. This pass is
258 automatically run by the parser after parsing input assembly and by
259 the optimizer before it outputs bytecode. The violations pointed out
260 by the verifier pass indicate bugs in transformation passes or input to
263 <!-- Describe the typesetting conventions here. --> </div>
265 <!-- *********************************************************************** -->
266 <div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
267 <!-- *********************************************************************** -->
269 <div class="doc_text">
271 <p>LLVM uses three different forms of identifiers, for different
275 <li>Named values are represented as a string of characters with a '%' prefix.
276 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
277 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
278 Identifiers which require other characters in their names can be surrounded
279 with quotes. In this way, anything except a <tt>"</tt> character can be used
282 <li>Unnamed values are represented as an unsigned numeric value with a '%'
283 prefix. For example, %12, %2, %44.</li>
285 <li>Constants, which are described in a <a href="#constants">section about
286 constants</a>, below.</li>
289 <p>LLVM requires that values start with a '%' sign for two reasons: Compilers
290 don't need to worry about name clashes with reserved words, and the set of
291 reserved words may be expanded in the future without penalty. Additionally,
292 unnamed identifiers allow a compiler to quickly come up with a temporary
293 variable without having to avoid symbol table conflicts.</p>
295 <p>Reserved words in LLVM are very similar to reserved words in other
296 languages. There are keywords for different opcodes
297 ('<tt><a href="#i_add">add</a></tt>',
298 '<tt><a href="#i_bitcast">bitcast</a></tt>',
299 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
300 href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
301 and others. These reserved words cannot conflict with variable names, because
302 none of them start with a '%' character.</p>
304 <p>Here is an example of LLVM code to multiply the integer variable
305 '<tt>%X</tt>' by 8:</p>
310 %result = <a href="#i_mul">mul</a> i32 %X, 8
313 <p>After strength reduction:</p>
316 %result = <a href="#i_shl">shl</a> i32 %X, i8 3
319 <p>And the hard way:</p>
322 <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
323 <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
324 %result = <a href="#i_add">add</a> i32 %1, %1
327 <p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
328 important lexical features of LLVM:</p>
332 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
335 <li>Unnamed temporaries are created when the result of a computation is not
336 assigned to a named value.</li>
338 <li>Unnamed temporaries are numbered sequentially</li>
342 <p>...and it also shows a convention that we follow in this document. When
343 demonstrating instructions, we will follow an instruction with a comment that
344 defines the type and name of value produced. Comments are shown in italic
349 <!-- *********************************************************************** -->
350 <div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
351 <!-- *********************************************************************** -->
353 <!-- ======================================================================= -->
354 <div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
357 <div class="doc_text">
359 <p>LLVM programs are composed of "Module"s, each of which is a
360 translation unit of the input programs. Each module consists of
361 functions, global variables, and symbol table entries. Modules may be
362 combined together with the LLVM linker, which merges function (and
363 global variable) definitions, resolves forward declarations, and merges
364 symbol table entries. Here is an example of the "hello world" module:</p>
366 <pre><i>; Declare the string constant as a global constant...</i>
367 <a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
368 href="#globalvars">constant</a> <a href="#t_array">[13 x i8 ]</a> c"hello world\0A\00" <i>; [13 x i8 ]*</i>
370 <i>; External declaration of the puts function</i>
371 <a href="#functionstructure">declare</a> i32 %puts(i8 *) <i>; i32(i8 *)* </i>
373 <i>; Global variable / Function body section separator</i>
376 <i>; Definition of main function</i>
377 define i32 %main() { <i>; i32()* </i>
378 <i>; Convert [13x i8 ]* to i8 *...</i>
380 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* %.LC0, i64 0, i64 0 <i>; i8 *</i>
382 <i>; Call puts function to write out the string to stdout...</i>
384 href="#i_call">call</a> i32 %puts(i8 * %cast210) <i>; i32</i>
386 href="#i_ret">ret</a> i32 0<br>}<br></pre>
388 <p>This example is made up of a <a href="#globalvars">global variable</a>
389 named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
390 function, and a <a href="#functionstructure">function definition</a>
391 for "<tt>main</tt>".</p>
393 <p>In general, a module is made up of a list of global values,
394 where both functions and global variables are global values. Global values are
395 represented by a pointer to a memory location (in this case, a pointer to an
396 array of char, and a pointer to a function), and have one of the following <a
397 href="#linkage">linkage types</a>.</p>
399 <p>Due to a limitation in the current LLVM assembly parser (it is limited by
400 one-token lookahead), modules are split into two pieces by the "implementation"
401 keyword. Global variable prototypes and definitions must occur before the
402 keyword, and function definitions must occur after it. Function prototypes may
403 occur either before or after it. In the future, the implementation keyword may
404 become a noop, if the parser gets smarter.</p>
408 <!-- ======================================================================= -->
409 <div class="doc_subsection">
410 <a name="linkage">Linkage Types</a>
413 <div class="doc_text">
416 All Global Variables and Functions have one of the following types of linkage:
421 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
423 <dd>Global values with internal linkage are only directly accessible by
424 objects in the current module. In particular, linking code into a module with
425 an internal global value may cause the internal to be renamed as necessary to
426 avoid collisions. Because the symbol is internal to the module, all
427 references can be updated. This corresponds to the notion of the
428 '<tt>static</tt>' keyword in C.
431 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
433 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
434 the same name when linkage occurs. This is typically used to implement
435 inline functions, templates, or other code which must be generated in each
436 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
437 allowed to be discarded.
440 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
442 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
443 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
444 used for globals that may be emitted in multiple translation units, but that
445 are not guaranteed to be emitted into every translation unit that uses them.
446 One example of this are common globals in C, such as "<tt>int X;</tt>" at
450 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
452 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
453 pointer to array type. When two global variables with appending linkage are
454 linked together, the two global arrays are appended together. This is the
455 LLVM, typesafe, equivalent of having the system linker append together
456 "sections" with identical names when .o files are linked.
459 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
460 <dd>The semantics of this linkage follow the ELF model: the symbol is weak
461 until linked, if not linked, the symbol becomes null instead of being an
466 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
468 <dd>If none of the above identifiers are used, the global is externally
469 visible, meaning that it participates in linkage and can be used to resolve
470 external symbol references.
474 The next two types of linkage are targeted for Microsoft Windows platform
475 only. They are designed to support importing (exporting) symbols from (to)
480 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
482 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
483 or variable via a global pointer to a pointer that is set up by the DLL
484 exporting the symbol. On Microsoft Windows targets, the pointer name is
485 formed by combining <code>_imp__</code> and the function or variable name.
488 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
490 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
491 pointer to a pointer in a DLL, so that it can be referenced with the
492 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
493 name is formed by combining <code>_imp__</code> and the function or variable
499 <p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
500 variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
501 variable and was linked with this one, one of the two would be renamed,
502 preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
503 external (i.e., lacking any linkage declarations), they are accessible
504 outside of the current module.</p>
505 <p>It is illegal for a function <i>declaration</i>
506 to have any linkage type other than "externally visible", <tt>dllimport</tt>,
507 or <tt>extern_weak</tt>.</p>
511 <!-- ======================================================================= -->
512 <div class="doc_subsection">
513 <a name="callingconv">Calling Conventions</a>
516 <div class="doc_text">
518 <p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
519 and <a href="#i_invoke">invokes</a> can all have an optional calling convention
520 specified for the call. The calling convention of any pair of dynamic
521 caller/callee must match, or the behavior of the program is undefined. The
522 following calling conventions are supported by LLVM, and more may be added in
526 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
528 <dd>This calling convention (the default if no other calling convention is
529 specified) matches the target C calling conventions. This calling convention
530 supports varargs function calls and tolerates some mismatch in the declared
531 prototype and implemented declaration of the function (as does normal C).
534 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
536 <dd>This calling convention attempts to make calls as fast as possible
537 (e.g. by passing things in registers). This calling convention allows the
538 target to use whatever tricks it wants to produce fast code for the target,
539 without having to conform to an externally specified ABI. Implementations of
540 this convention should allow arbitrary tail call optimization to be supported.
541 This calling convention does not support varargs and requires the prototype of
542 all callees to exactly match the prototype of the function definition.
545 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
547 <dd>This calling convention attempts to make code in the caller as efficient
548 as possible under the assumption that the call is not commonly executed. As
549 such, these calls often preserve all registers so that the call does not break
550 any live ranges in the caller side. This calling convention does not support
551 varargs and requires the prototype of all callees to exactly match the
552 prototype of the function definition.
555 <dt><b>"<tt>cc <<em>n</em>></tt>" - Numbered convention</b>:</dt>
557 <dd>Any calling convention may be specified by number, allowing
558 target-specific calling conventions to be used. Target specific calling
559 conventions start at 64.
563 <p>More calling conventions can be added/defined on an as-needed basis, to
564 support pascal conventions or any other well-known target-independent
569 <!-- ======================================================================= -->
570 <div class="doc_subsection">
571 <a name="visibility">Visibility Styles</a>
574 <div class="doc_text">
577 All Global Variables and Functions have one of the following visibility styles:
581 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
583 <dd>On ELF, default visibility means that the declaration is visible to other
584 modules and, in shared libraries, means that the declared entity may be
585 overridden. On Darwin, default visibility means that the declaration is
586 visible to other modules. Default visibility corresponds to "external
587 linkage" in the language.
590 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
592 <dd>Two declarations of an object with hidden visibility refer to the same
593 object if they are in the same shared object. Usually, hidden visibility
594 indicates that the symbol will not be placed into the dynamic symbol table,
595 so no other module (executable or shared library) can reference it
603 <!-- ======================================================================= -->
604 <div class="doc_subsection">
605 <a name="globalvars">Global Variables</a>
608 <div class="doc_text">
610 <p>Global variables define regions of memory allocated at compilation time
611 instead of run-time. Global variables may optionally be initialized, may have
612 an explicit section to be placed in, and may
613 have an optional explicit alignment specified. A
614 variable may be defined as a global "constant," which indicates that the
615 contents of the variable will <b>never</b> be modified (enabling better
616 optimization, allowing the global data to be placed in the read-only section of
617 an executable, etc). Note that variables that need runtime initialization
618 cannot be marked "constant" as there is a store to the variable.</p>
621 LLVM explicitly allows <em>declarations</em> of global variables to be marked
622 constant, even if the final definition of the global is not. This capability
623 can be used to enable slightly better optimization of the program, but requires
624 the language definition to guarantee that optimizations based on the
625 'constantness' are valid for the translation units that do not include the
629 <p>As SSA values, global variables define pointer values that are in
630 scope (i.e. they dominate) all basic blocks in the program. Global
631 variables always define a pointer to their "content" type because they
632 describe a region of memory, and all memory objects in LLVM are
633 accessed through pointers.</p>
635 <p>LLVM allows an explicit section to be specified for globals. If the target
636 supports it, it will emit globals to the section specified.</p>
638 <p>An explicit alignment may be specified for a global. If not present, or if
639 the alignment is set to zero, the alignment of the global is set by the target
640 to whatever it feels convenient. If an explicit alignment is specified, the
641 global is forced to have at least that much alignment. All alignments must be
644 <p>For example, the following defines a global with an initializer, section,
648 %G = constant float 1.0, section "foo", align 4
654 <!-- ======================================================================= -->
655 <div class="doc_subsection">
656 <a name="functionstructure">Functions</a>
659 <div class="doc_text">
661 <p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
662 an optional <a href="#linkage">linkage type</a>, an optional
663 <a href="#visibility">visibility style</a>, an optional
664 <a href="#callingconv">calling convention</a>, a return type, an optional
665 <a href="#paramattrs">parameter attribute</a> for the return type, a function
666 name, a (possibly empty) argument list (each with optional
667 <a href="#paramattrs">parameter attributes</a>), an optional section, an
668 optional alignment, an opening curly brace, a list of basic blocks, and a
671 LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
672 optional <a href="#linkage">linkage type</a>, an optional
673 <a href="#visibility">visibility style</a>, an optional
674 <a href="#callingconv">calling convention</a>, a return type, an optional
675 <a href="#paramattrs">parameter attribute</a> for the return type, a function
676 name, a possibly empty list of arguments, and an optional alignment.</p>
678 <p>A function definition contains a list of basic blocks, forming the CFG for
679 the function. Each basic block may optionally start with a label (giving the
680 basic block a symbol table entry), contains a list of instructions, and ends
681 with a <a href="#terminators">terminator</a> instruction (such as a branch or
682 function return).</p>
684 <p>The first basic block in a program is special in two ways: it is immediately
685 executed on entrance to the function, and it is not allowed to have predecessor
686 basic blocks (i.e. there can not be any branches to the entry block of a
687 function). Because the block can have no predecessors, it also cannot have any
688 <a href="#i_phi">PHI nodes</a>.</p>
690 <p>LLVM functions are identified by their name and type signature. Hence, two
691 functions with the same name but different parameter lists or return values are
692 considered different functions, and LLVM will resolve references to each
695 <p>LLVM allows an explicit section to be specified for functions. If the target
696 supports it, it will emit functions to the section specified.</p>
698 <p>An explicit alignment may be specified for a function. If not present, or if
699 the alignment is set to zero, the alignment of the function is set by the target
700 to whatever it feels convenient. If an explicit alignment is specified, the
701 function is forced to have at least that much alignment. All alignments must be
706 <!-- ======================================================================= -->
707 <div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
708 <div class="doc_text">
709 <p>The return type and each parameter of a function type may have a set of
710 <i>parameter attributes</i> associated with them. Parameter attributes are
711 used to communicate additional information about the result or parameters of
712 a function. Parameter attributes are considered to be part of the function
713 type so two functions types that differ only by the parameter attributes
714 are different function types.</p>
716 <p>Parameter attributes are simple keywords that follow the type specified. If
717 multiple parameter attributes are needed, they are space separated. For
719 %someFunc = i16 (i8 sext %someParam) zext
720 %someFunc = i16 (i8 zext %someParam) zext</pre>
721 <p>Note that the two function types above are unique because the parameter has
722 a different attribute (sext in the first one, zext in the second). Also note
723 that the attribute for the function result (zext) comes immediately after the
726 <p>Currently, only the following parameter attributes are defined:</p>
728 <dt><tt>zext</tt></dt>
729 <dd>This indicates that the parameter should be zero extended just before
730 a call to this function.</dd>
731 <dt><tt>sext</tt></dt>
732 <dd>This indicates that the parameter should be sign extended just before
733 a call to this function.</dd>
734 <dt><tt>inreg</tt></dt>
735 <dd>This indicates that the parameter should be placed in register (if
736 possible) during assembling function call. Support for this attribute is
738 <dt><tt>sret</tt></dt>
739 <dd>This indicates that the parameter specifies the address of a structure
740 that is the return value of the function in the source program.
746 <!-- ======================================================================= -->
747 <div class="doc_subsection">
748 <a name="moduleasm">Module-Level Inline Assembly</a>
751 <div class="doc_text">
753 Modules may contain "module-level inline asm" blocks, which corresponds to the
754 GCC "file scope inline asm" blocks. These blocks are internally concatenated by
755 LLVM and treated as a single unit, but may be separated in the .ll file if
756 desired. The syntax is very simple:
759 <div class="doc_code"><pre>
760 module asm "inline asm code goes here"
761 module asm "more can go here"
764 <p>The strings can contain any character by escaping non-printable characters.
765 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
770 The inline asm code is simply printed to the machine code .s file when
771 assembly code is generated.
775 <!-- ======================================================================= -->
776 <div class="doc_subsection">
777 <a name="datalayout">Data Layout</a>
780 <div class="doc_text">
781 <p>A module may specify a target specific data layout string that specifies how
782 data is to be laid out in memory. The syntax for the data layout is simply:<br/>
783 <pre> target datalayout = "<i>layout specification</i>"
785 The <i>layout specification</i> consists of a list of specifications separated
786 by the minus sign character ('-'). Each specification starts with a letter
787 and may include other information after the letter to define some aspect of the
788 data layout. The specifications accepted are as follows: </p>
791 <dd>Specifies that the target lays out data in big-endian form. That is, the
792 bits with the most significance have the lowest address location.</dd>
794 <dd>Specifies that hte target lays out data in little-endian form. That is,
795 the bits with the least significance have the lowest address location.</dd>
796 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
797 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
798 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
799 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
801 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
802 <dd>This specifies the alignment for an integer type of a given bit
803 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
804 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
805 <dd>This specifies the alignment for a vector type of a given bit
807 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
808 <dd>This specifies the alignment for a floating point type of a given bit
809 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
811 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
812 <dd>This specifies the alignment for an aggregate type of a given bit
815 <p>When constructing the data layout for a given target, LLVM starts with a
816 default set of specifications which are then (possibly) overriden by the
817 specifications in the <tt>datalayout</tt> keyword. The default specifications
818 are given in this list:</p>
820 <li><tt>E</tt> - big endian</li>
821 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
822 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
823 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
824 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
825 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
826 <li><tt>i64:32:64</tt> - i64 has abi alignment of 32-bits but preferred
827 alignment of 64-bits</li>
828 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
829 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
830 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
831 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
832 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
834 <p>When llvm is determining the alignment for a given type, it uses the
837 <li>If the type sought is an exact match for one of the specifications, that
838 specification is used.</li>
839 <li>If no match is found, and the type sought is an integer type, then the
840 smallest integer type that is larger than the bitwidth of the sought type is
841 used. If none of the specifications are larger than the bitwidth then the the
842 largest integer type is used. For example, given the default specifications
843 above, the i7 type will use the alignment of i8 (next largest) while both
844 i65 and i256 will use the alignment of i64 (largest specified).</li>
845 <li>If no match is found, and the type sought is a vector type, then the
846 largest vector type that is smaller than the sought vector type will be used
847 as a fall back. This happens because <128 x double> can be implemented in
848 terms of 64 <2 x double>, for example.</li>
852 <!-- *********************************************************************** -->
853 <div class="doc_section"> <a name="typesystem">Type System</a> </div>
854 <!-- *********************************************************************** -->
856 <div class="doc_text">
858 <p>The LLVM type system is one of the most important features of the
859 intermediate representation. Being typed enables a number of
860 optimizations to be performed on the IR directly, without having to do
861 extra analyses on the side before the transformation. A strong type
862 system makes it easier to read the generated code and enables novel
863 analyses and transformations that are not feasible to perform on normal
864 three address code representations.</p>
868 <!-- ======================================================================= -->
869 <div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
870 <div class="doc_text">
871 <p>The primitive types are the fundamental building blocks of the LLVM
872 system. The current set of primitive types is as follows:</p>
874 <table class="layout">
879 <tr><th>Type</th><th>Description</th></tr>
880 <tr><td><tt>void</tt></td><td>No value</td></tr>
881 <tr><td><tt>i8</tt></td><td>8-bit value</td></tr>
882 <tr><td><tt>i32</tt></td><td>32-bit value</td></tr>
883 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
884 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
891 <tr><th>Type</th><th>Description</th></tr>
892 <tr><td><tt>i1</tt></td><td>True or False value</td></tr>
893 <tr><td><tt>i16</tt></td><td>16-bit value</td></tr>
894 <tr><td><tt>i64</tt></td><td>64-bit value</td></tr>
895 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
903 <!-- _______________________________________________________________________ -->
904 <div class="doc_subsubsection"> <a name="t_classifications">Type
905 Classifications</a> </div>
906 <div class="doc_text">
907 <p>These different primitive types fall into a few useful
910 <table border="1" cellspacing="0" cellpadding="4">
912 <tr><th>Classification</th><th>Types</th></tr>
914 <td><a name="t_integer">integer</a></td>
915 <td><tt>i1, i8, i16, i32, i64</tt></td>
918 <td><a name="t_floating">floating point</a></td>
919 <td><tt>float, double</tt></td>
922 <td><a name="t_firstclass">first class</a></td>
923 <td><tt>i1, i8, i16, i32, i64, float, double, <br/>
924 <a href="#t_pointer">pointer</a>,<a href="#t_vector">vector</a></tt>
930 <p>The <a href="#t_firstclass">first class</a> types are perhaps the
931 most important. Values of these types are the only ones which can be
932 produced by instructions, passed as arguments, or used as operands to
933 instructions. This means that all structures and arrays must be
934 manipulated either by pointer or by component.</p>
937 <!-- ======================================================================= -->
938 <div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
940 <div class="doc_text">
942 <p>The real power in LLVM comes from the derived types in the system.
943 This is what allows a programmer to represent arrays, functions,
944 pointers, and other useful types. Note that these derived types may be
945 recursive: For example, it is possible to have a two dimensional array.</p>
949 <!-- _______________________________________________________________________ -->
950 <div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
952 <div class="doc_text">
956 <p>The array type is a very simple derived type that arranges elements
957 sequentially in memory. The array type requires a size (number of
958 elements) and an underlying data type.</p>
963 [<# elements> x <elementtype>]
966 <p>The number of elements is a constant integer value; elementtype may
967 be any type with a size.</p>
970 <table class="layout">
973 <tt>[40 x i32 ]</tt><br/>
974 <tt>[41 x i32 ]</tt><br/>
975 <tt>[40 x i8]</tt><br/>
978 Array of 40 32-bit integer values.<br/>
979 Array of 41 32-bit integer values.<br/>
980 Array of 40 8-bit integer values.<br/>
984 <p>Here are some examples of multidimensional arrays:</p>
985 <table class="layout">
988 <tt>[3 x [4 x i32]]</tt><br/>
989 <tt>[12 x [10 x float]]</tt><br/>
990 <tt>[2 x [3 x [4 x i16]]]</tt><br/>
993 3x4 array of 32-bit integer values.<br/>
994 12x10 array of single precision floating point values.<br/>
995 2x3x4 array of 16-bit integer values.<br/>
1000 <p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1001 length array. Normally, accesses past the end of an array are undefined in
1002 LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1003 As a special case, however, zero length arrays are recognized to be variable
1004 length. This allows implementation of 'pascal style arrays' with the LLVM
1005 type "{ i32, [0 x float]}", for example.</p>
1009 <!-- _______________________________________________________________________ -->
1010 <div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
1011 <div class="doc_text">
1013 <p>The function type can be thought of as a function signature. It
1014 consists of a return type and a list of formal parameter types.
1015 Function types are usually used to build virtual function tables
1016 (which are structures of pointers to functions), for indirect function
1017 calls, and when defining a function.</p>
1019 The return type of a function type cannot be an aggregate type.
1022 <pre> <returntype> (<parameter list>)<br></pre>
1023 <p>...where '<tt><parameter list></tt>' is a comma-separated list of type
1024 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1025 which indicates that the function takes a variable number of arguments.
1026 Variable argument functions can access their arguments with the <a
1027 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
1029 <table class="layout">
1031 <td class="left"><tt>i32 (i32)</tt></td>
1032 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1034 </tr><tr class="layout">
1035 <td class="left"><tt>float (i16 sext, i32 *) *
1037 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1038 an <tt>i16</tt> that should be sign extended and a
1039 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
1042 </tr><tr class="layout">
1043 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1044 <td class="left">A vararg function that takes at least one
1045 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1046 which returns an integer. This is the signature for <tt>printf</tt> in
1053 <!-- _______________________________________________________________________ -->
1054 <div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
1055 <div class="doc_text">
1057 <p>The structure type is used to represent a collection of data members
1058 together in memory. The packing of the field types is defined to match
1059 the ABI of the underlying processor. The elements of a structure may
1060 be any type that has a size.</p>
1061 <p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1062 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1063 field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1066 <pre> { <type list> }<br></pre>
1068 <table class="layout">
1071 <tt>{ i32, i32, i32 }</tt><br/>
1072 <tt>{ float, i32 (i32) * }</tt><br/>
1075 a triple of three <tt>i32</tt> values<br/>
1076 A pair, where the first element is a <tt>float</tt> and the second element
1077 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
1078 that takes an <tt>i32</tt>, returning an <tt>i32</tt>.<br/>
1084 <!-- _______________________________________________________________________ -->
1085 <div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1087 <div class="doc_text">
1089 <p>The packed structure type is used to represent a collection of data members
1090 together in memory. There is no padding between fields. Further, the alignment
1091 of a packed structure is 1 byte. The elements of a packed structure may
1092 be any type that has a size.</p>
1093 <p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1094 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1095 field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1098 <pre> < { <type list> } > <br></pre>
1100 <table class="layout">
1103 <tt> < { i32, i32, i32 } > </tt><br/>
1104 <tt> < { float, i32 (i32) * } > </tt><br/>
1107 a triple of three <tt>i32</tt> values<br/>
1108 A pair, where the first element is a <tt>float</tt> and the second element
1109 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
1110 that takes an <tt>i32</tt>, returning an <tt>i32</tt>.<br/>
1116 <!-- _______________________________________________________________________ -->
1117 <div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
1118 <div class="doc_text">
1120 <p>As in many languages, the pointer type represents a pointer or
1121 reference to another object, which must live in memory.</p>
1123 <pre> <type> *<br></pre>
1125 <table class="layout">
1128 <tt>[4x i32]*</tt><br/>
1129 <tt>i32 (i32 *) *</tt><br/>
1132 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
1133 four <tt>i32</tt> values<br/>
1134 A <a href="#t_pointer">pointer</a> to a <a
1135 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
1142 <!-- _______________________________________________________________________ -->
1143 <div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
1144 <div class="doc_text">
1148 <p>A vector type is a simple derived type that represents a vector
1149 of elements. Vector types are used when multiple primitive data
1150 are operated in parallel using a single instruction (SIMD).
1151 A vector type requires a size (number of
1152 elements) and an underlying primitive data type. Vectors must have a power
1153 of two length (1, 2, 4, 8, 16 ...). Vector types are
1154 considered <a href="#t_firstclass">first class</a>.</p>
1159 < <# elements> x <elementtype> >
1162 <p>The number of elements is a constant integer value; elementtype may
1163 be any integer or floating point type.</p>
1167 <table class="layout">
1170 <tt><4 x i32></tt><br/>
1171 <tt><8 x float></tt><br/>
1172 <tt><2 x i64></tt><br/>
1175 Vector of 4 32-bit integer values.<br/>
1176 Vector of 8 floating-point values.<br/>
1177 Vector of 2 64-bit integer values.<br/>
1183 <!-- _______________________________________________________________________ -->
1184 <div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1185 <div class="doc_text">
1189 <p>Opaque types are used to represent unknown types in the system. This
1190 corresponds (for example) to the C notion of a foward declared structure type.
1191 In LLVM, opaque types can eventually be resolved to any type (not just a
1192 structure type).</p>
1202 <table class="layout">
1208 An opaque type.<br/>
1215 <!-- *********************************************************************** -->
1216 <div class="doc_section"> <a name="constants">Constants</a> </div>
1217 <!-- *********************************************************************** -->
1219 <div class="doc_text">
1221 <p>LLVM has several different basic types of constants. This section describes
1222 them all and their syntax.</p>
1226 <!-- ======================================================================= -->
1227 <div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
1229 <div class="doc_text">
1232 <dt><b>Boolean constants</b></dt>
1234 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1235 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
1238 <dt><b>Integer constants</b></dt>
1240 <dd>Standard integers (such as '4') are constants of the <a
1241 href="#t_integer">integer</a> type. Negative numbers may be used with
1245 <dt><b>Floating point constants</b></dt>
1247 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1248 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
1249 notation (see below). Floating point constants must have a <a
1250 href="#t_floating">floating point</a> type. </dd>
1252 <dt><b>Null pointer constants</b></dt>
1254 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
1255 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1259 <p>The one non-intuitive notation for constants is the optional hexadecimal form
1260 of floating point constants. For example, the form '<tt>double
1261 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
1262 4.5e+15</tt>'. The only time hexadecimal floating point constants are required
1263 (and the only time that they are generated by the disassembler) is when a
1264 floating point constant must be emitted but it cannot be represented as a
1265 decimal floating point number. For example, NaN's, infinities, and other
1266 special values are represented in their IEEE hexadecimal format so that
1267 assembly and disassembly do not cause any bits to change in the constants.</p>
1271 <!-- ======================================================================= -->
1272 <div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1275 <div class="doc_text">
1276 <p>Aggregate constants arise from aggregation of simple constants
1277 and smaller aggregate constants.</p>
1280 <dt><b>Structure constants</b></dt>
1282 <dd>Structure constants are represented with notation similar to structure
1283 type definitions (a comma separated list of elements, surrounded by braces
1284 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* %G }</tt>",
1285 where "<tt>%G</tt>" is declared as "<tt>%G = external global i32</tt>". Structure constants
1286 must have <a href="#t_struct">structure type</a>, and the number and
1287 types of elements must match those specified by the type.
1290 <dt><b>Array constants</b></dt>
1292 <dd>Array constants are represented with notation similar to array type
1293 definitions (a comma separated list of elements, surrounded by square brackets
1294 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
1295 constants must have <a href="#t_array">array type</a>, and the number and
1296 types of elements must match those specified by the type.
1299 <dt><b>Vector constants</b></dt>
1301 <dd>Vector constants are represented with notation similar to vector type
1302 definitions (a comma separated list of elements, surrounded by
1303 less-than/greater-than's (<tt><></tt>)). For example: "<tt>< i32 42,
1304 i32 11, i32 74, i32 100 ></tt>". VEctor constants must have <a
1305 href="#t_vector">vector type</a>, and the number and types of elements must
1306 match those specified by the type.
1309 <dt><b>Zero initialization</b></dt>
1311 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1312 value to zero of <em>any</em> type, including scalar and aggregate types.
1313 This is often used to avoid having to print large zero initializers (e.g. for
1314 large arrays) and is always exactly equivalent to using explicit zero
1321 <!-- ======================================================================= -->
1322 <div class="doc_subsection">
1323 <a name="globalconstants">Global Variable and Function Addresses</a>
1326 <div class="doc_text">
1328 <p>The addresses of <a href="#globalvars">global variables</a> and <a
1329 href="#functionstructure">functions</a> are always implicitly valid (link-time)
1330 constants. These constants are explicitly referenced when the <a
1331 href="#identifiers">identifier for the global</a> is used and always have <a
1332 href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1338 %Z = global [2 x i32*] [ i32* %X, i32* %Y ]
1343 <!-- ======================================================================= -->
1344 <div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1345 <div class="doc_text">
1346 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1347 no specific value. Undefined values may be of any type and be used anywhere
1348 a constant is permitted.</p>
1350 <p>Undefined values indicate to the compiler that the program is well defined
1351 no matter what value is used, giving the compiler more freedom to optimize.
1355 <!-- ======================================================================= -->
1356 <div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1359 <div class="doc_text">
1361 <p>Constant expressions are used to allow expressions involving other constants
1362 to be used as constants. Constant expressions may be of any <a
1363 href="#t_firstclass">first class</a> type and may involve any LLVM operation
1364 that does not have side effects (e.g. load and call are not supported). The
1365 following is the syntax for constant expressions:</p>
1368 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1369 <dd>Truncate a constant to another type. The bit size of CST must be larger
1370 than the bit size of TYPE. Both types must be integers.</dd>
1372 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1373 <dd>Zero extend a constant to another type. The bit size of CST must be
1374 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1376 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1377 <dd>Sign extend a constant to another type. The bit size of CST must be
1378 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
1380 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1381 <dd>Truncate a floating point constant to another floating point type. The
1382 size of CST must be larger than the size of TYPE. Both types must be
1383 floating point.</dd>
1385 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1386 <dd>Floating point extend a constant to another type. The size of CST must be
1387 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1389 <dt><b><tt>fp2uint ( CST to TYPE )</tt></b></dt>
1390 <dd>Convert a floating point constant to the corresponding unsigned integer
1391 constant. TYPE must be an integer type. CST must be floating point. If the
1392 value won't fit in the integer type, the results are undefined.</dd>
1394 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
1395 <dd>Convert a floating point constant to the corresponding signed integer
1396 constant. TYPE must be an integer type. CST must be floating point. If the
1397 value won't fit in the integer type, the results are undefined.</dd>
1399 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
1400 <dd>Convert an unsigned integer constant to the corresponding floating point
1401 constant. TYPE must be floating point. CST must be of integer type. If the
1402 value won't fit in the floating point type, the results are undefined.</dd>
1404 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
1405 <dd>Convert a signed integer constant to the corresponding floating point
1406 constant. TYPE must be floating point. CST must be of integer type. If the
1407 value won't fit in the floating point type, the results are undefined.</dd>
1409 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1410 <dd>Convert a pointer typed constant to the corresponding integer constant
1411 TYPE must be an integer type. CST must be of pointer type. The CST value is
1412 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1414 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1415 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1416 pointer type. CST must be of integer type. The CST value is zero extended,
1417 truncated, or unchanged to make it fit in a pointer size. This one is
1418 <i>really</i> dangerous!</dd>
1420 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
1421 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1422 identical (same number of bits). The conversion is done as if the CST value
1423 was stored to memory and read back as TYPE. In other words, no bits change
1424 with this operator, just the type. This can be used for conversion of
1425 vector types to any other type, as long as they have the same bit width. For
1426 pointers it is only valid to cast to another pointer type.
1429 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1431 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1432 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1433 instruction, the index list may have zero or more indexes, which are required
1434 to make sense for the type of "CSTPTR".</dd>
1436 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1438 <dd>Perform the <a href="#i_select">select operation</a> on
1441 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1442 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1444 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1445 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
1447 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1449 <dd>Perform the <a href="#i_extractelement">extractelement
1450 operation</a> on constants.
1452 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1454 <dd>Perform the <a href="#i_insertelement">insertelement
1455 operation</a> on constants.</dd>
1458 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1460 <dd>Perform the <a href="#i_shufflevector">shufflevector
1461 operation</a> on constants.</dd>
1463 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1465 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1466 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1467 binary</a> operations. The constraints on operands are the same as those for
1468 the corresponding instruction (e.g. no bitwise operations on floating point
1469 values are allowed).</dd>
1473 <!-- *********************************************************************** -->
1474 <div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1475 <!-- *********************************************************************** -->
1477 <!-- ======================================================================= -->
1478 <div class="doc_subsection">
1479 <a name="inlineasm">Inline Assembler Expressions</a>
1482 <div class="doc_text">
1485 LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1486 Module-Level Inline Assembly</a>) through the use of a special value. This
1487 value represents the inline assembler as a string (containing the instructions
1488 to emit), a list of operand constraints (stored as a string), and a flag that
1489 indicates whether or not the inline asm expression has side effects. An example
1490 inline assembler expression is:
1494 i32 (i32) asm "bswap $0", "=r,r"
1498 Inline assembler expressions may <b>only</b> be used as the callee operand of
1499 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1503 %X = call i32 asm "<a href="#i_bswap">bswap</a> $0", "=r,r"(i32 %Y)
1507 Inline asms with side effects not visible in the constraint list must be marked
1508 as having side effects. This is done through the use of the
1509 '<tt>sideeffect</tt>' keyword, like so:
1513 call void asm sideeffect "eieio", ""()
1516 <p>TODO: The format of the asm and constraints string still need to be
1517 documented here. Constraints on what can be done (e.g. duplication, moving, etc
1518 need to be documented).
1523 <!-- *********************************************************************** -->
1524 <div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1525 <!-- *********************************************************************** -->
1527 <div class="doc_text">
1529 <p>The LLVM instruction set consists of several different
1530 classifications of instructions: <a href="#terminators">terminator
1531 instructions</a>, <a href="#binaryops">binary instructions</a>,
1532 <a href="#bitwiseops">bitwise binary instructions</a>, <a
1533 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1534 instructions</a>.</p>
1538 <!-- ======================================================================= -->
1539 <div class="doc_subsection"> <a name="terminators">Terminator
1540 Instructions</a> </div>
1542 <div class="doc_text">
1544 <p>As mentioned <a href="#functionstructure">previously</a>, every
1545 basic block in a program ends with a "Terminator" instruction, which
1546 indicates which block should be executed after the current block is
1547 finished. These terminator instructions typically yield a '<tt>void</tt>'
1548 value: they produce control flow, not values (the one exception being
1549 the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1550 <p>There are six different terminator instructions: the '<a
1551 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1552 instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1553 the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1554 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1555 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1559 <!-- _______________________________________________________________________ -->
1560 <div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1561 Instruction</a> </div>
1562 <div class="doc_text">
1564 <pre> ret <type> <value> <i>; Return a value from a non-void function</i>
1565 ret void <i>; Return from void function</i>
1568 <p>The '<tt>ret</tt>' instruction is used to return control flow (and a
1569 value) from a function back to the caller.</p>
1570 <p>There are two forms of the '<tt>ret</tt>' instruction: one that
1571 returns a value and then causes control flow, and one that just causes
1572 control flow to occur.</p>
1574 <p>The '<tt>ret</tt>' instruction may return any '<a
1575 href="#t_firstclass">first class</a>' type. Notice that a function is
1576 not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1577 instruction inside of the function that returns a value that does not
1578 match the return type of the function.</p>
1580 <p>When the '<tt>ret</tt>' instruction is executed, control flow
1581 returns back to the calling function's context. If the caller is a "<a
1582 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1583 the instruction after the call. If the caller was an "<a
1584 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1585 at the beginning of the "normal" destination block. If the instruction
1586 returns a value, that value shall set the call or invoke instruction's
1589 <pre> ret i32 5 <i>; Return an integer value of 5</i>
1590 ret void <i>; Return from a void function</i>
1593 <!-- _______________________________________________________________________ -->
1594 <div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1595 <div class="doc_text">
1597 <pre> br i1 <cond>, label <iftrue>, label <iffalse><br> br label <dest> <i>; Unconditional branch</i>
1600 <p>The '<tt>br</tt>' instruction is used to cause control flow to
1601 transfer to a different basic block in the current function. There are
1602 two forms of this instruction, corresponding to a conditional branch
1603 and an unconditional branch.</p>
1605 <p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1606 single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
1607 unconditional form of the '<tt>br</tt>' instruction takes a single
1608 '<tt>label</tt>' value as a target.</p>
1610 <p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
1611 argument is evaluated. If the value is <tt>true</tt>, control flows
1612 to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1613 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
1615 <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
1616 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
1618 <!-- _______________________________________________________________________ -->
1619 <div class="doc_subsubsection">
1620 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1623 <div class="doc_text">
1627 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
1632 <p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1633 several different places. It is a generalization of the '<tt>br</tt>'
1634 instruction, allowing a branch to occur to one of many possible
1640 <p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1641 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1642 an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1643 table is not allowed to contain duplicate constant entries.</p>
1647 <p>The <tt>switch</tt> instruction specifies a table of values and
1648 destinations. When the '<tt>switch</tt>' instruction is executed, this
1649 table is searched for the given value. If the value is found, control flow is
1650 transfered to the corresponding destination; otherwise, control flow is
1651 transfered to the default destination.</p>
1653 <h5>Implementation:</h5>
1655 <p>Depending on properties of the target machine and the particular
1656 <tt>switch</tt> instruction, this instruction may be code generated in different
1657 ways. For example, it could be generated as a series of chained conditional
1658 branches or with a lookup table.</p>
1663 <i>; Emulate a conditional br instruction</i>
1664 %Val = <a href="#i_zext">zext</a> i1 %value to i32
1665 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
1667 <i>; Emulate an unconditional br instruction</i>
1668 switch i32 0, label %dest [ ]
1670 <i>; Implement a jump table:</i>
1671 switch i32 %val, label %otherwise [ i32 0, label %onzero
1673 i32 2, label %ontwo ]
1677 <!-- _______________________________________________________________________ -->
1678 <div class="doc_subsubsection">
1679 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1682 <div class="doc_text">
1687 <result> = invoke [<a href="#callingconv">cconv</a>] <ptr to function ty> %<function ptr val>(<function args>)
1688 to label <normal label> unwind label <exception label>
1693 <p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1694 function, with the possibility of control flow transfer to either the
1695 '<tt>normal</tt>' label or the
1696 '<tt>exception</tt>' label. If the callee function returns with the
1697 "<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1698 "normal" label. If the callee (or any indirect callees) returns with the "<a
1699 href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1700 continued at the dynamically nearest "exception" label.</p>
1704 <p>This instruction requires several arguments:</p>
1708 The optional "cconv" marker indicates which <a href="callingconv">calling
1709 convention</a> the call should use. If none is specified, the call defaults
1710 to using C calling conventions.
1712 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1713 function value being invoked. In most cases, this is a direct function
1714 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1715 an arbitrary pointer to function value.
1718 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1719 function to be invoked. </li>
1721 <li>'<tt>function args</tt>': argument list whose types match the function
1722 signature argument types. If the function signature indicates the function
1723 accepts a variable number of arguments, the extra arguments can be
1726 <li>'<tt>normal label</tt>': the label reached when the called function
1727 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1729 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1730 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1736 <p>This instruction is designed to operate as a standard '<tt><a
1737 href="#i_call">call</a></tt>' instruction in most regards. The primary
1738 difference is that it establishes an association with a label, which is used by
1739 the runtime library to unwind the stack.</p>
1741 <p>This instruction is used in languages with destructors to ensure that proper
1742 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1743 exception. Additionally, this is important for implementation of
1744 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
1748 %retval = invoke i32 %Test(i32 15) to label %Continue
1749 unwind label %TestCleanup <i>; {i32}:retval set</i>
1750 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Test(i32 15) to label %Continue
1751 unwind label %TestCleanup <i>; {i32}:retval set</i>
1756 <!-- _______________________________________________________________________ -->
1758 <div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1759 Instruction</a> </div>
1761 <div class="doc_text">
1770 <p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1771 at the first callee in the dynamic call stack which used an <a
1772 href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1773 primarily used to implement exception handling.</p>
1777 <p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1778 immediately halt. The dynamic call stack is then searched for the first <a
1779 href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1780 execution continues at the "exceptional" destination block specified by the
1781 <tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1782 dynamic call chain, undefined behavior results.</p>
1785 <!-- _______________________________________________________________________ -->
1787 <div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1788 Instruction</a> </div>
1790 <div class="doc_text">
1799 <p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1800 instruction is used to inform the optimizer that a particular portion of the
1801 code is not reachable. This can be used to indicate that the code after a
1802 no-return function cannot be reached, and other facts.</p>
1806 <p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1811 <!-- ======================================================================= -->
1812 <div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
1813 <div class="doc_text">
1814 <p>Binary operators are used to do most of the computation in a
1815 program. They require two operands, execute an operation on them, and
1816 produce a single value. The operands might represent
1817 multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
1818 The result value of a binary operator is not
1819 necessarily the same type as its operands.</p>
1820 <p>There are several different binary operators:</p>
1822 <!-- _______________________________________________________________________ -->
1823 <div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1824 Instruction</a> </div>
1825 <div class="doc_text">
1827 <pre> <result> = add <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1830 <p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
1832 <p>The two arguments to the '<tt>add</tt>' instruction must be either <a
1833 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1834 This instruction can also take <a href="#t_vector">vector</a> versions of the values.
1835 Both arguments must have identical types.</p>
1837 <p>The value produced is the integer or floating point sum of the two
1840 <pre> <result> = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
1843 <!-- _______________________________________________________________________ -->
1844 <div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1845 Instruction</a> </div>
1846 <div class="doc_text">
1848 <pre> <result> = sub <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1851 <p>The '<tt>sub</tt>' instruction returns the difference of its two
1853 <p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1854 instruction present in most other intermediate representations.</p>
1856 <p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
1857 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
1859 This instruction can also take <a href="#t_vector">vector</a> versions of the values.
1860 Both arguments must have identical types.</p>
1862 <p>The value produced is the integer or floating point difference of
1863 the two operands.</p>
1865 <pre> <result> = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
1866 <result> = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
1869 <!-- _______________________________________________________________________ -->
1870 <div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1871 Instruction</a> </div>
1872 <div class="doc_text">
1874 <pre> <result> = mul <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1877 <p>The '<tt>mul</tt>' instruction returns the product of its two
1880 <p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
1881 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
1883 This instruction can also take <a href="#t_vector">vector</a> versions of the values.
1884 Both arguments must have identical types.</p>
1886 <p>The value produced is the integer or floating point product of the
1888 <p>Because the operands are the same width, the result of an integer
1889 multiplication is the same whether the operands should be deemed unsigned or
1892 <pre> <result> = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
1895 <!-- _______________________________________________________________________ -->
1896 <div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
1898 <div class="doc_text">
1900 <pre> <result> = udiv <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1903 <p>The '<tt>udiv</tt>' instruction returns the quotient of its two
1906 <p>The two arguments to the '<tt>udiv</tt>' instruction must be
1907 <a href="#t_integer">integer</a> values. Both arguments must have identical
1908 types. This instruction can also take <a href="#t_vector">vector</a> versions
1909 of the values in which case the elements must be integers.</p>
1911 <p>The value produced is the unsigned integer quotient of the two operands. This
1912 instruction always performs an unsigned division operation, regardless of
1913 whether the arguments are unsigned or not.</p>
1915 <pre> <result> = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
1918 <!-- _______________________________________________________________________ -->
1919 <div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
1921 <div class="doc_text">
1923 <pre> <result> = sdiv <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1926 <p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
1929 <p>The two arguments to the '<tt>sdiv</tt>' instruction must be
1930 <a href="#t_integer">integer</a> values. Both arguments must have identical
1931 types. This instruction can also take <a href="#t_vector">vector</a> versions
1932 of the values in which case the elements must be integers.</p>
1934 <p>The value produced is the signed integer quotient of the two operands. This
1935 instruction always performs a signed division operation, regardless of whether
1936 the arguments are signed or not.</p>
1938 <pre> <result> = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
1941 <!-- _______________________________________________________________________ -->
1942 <div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
1943 Instruction</a> </div>
1944 <div class="doc_text">
1946 <pre> <result> = fdiv <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1949 <p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
1952 <p>The two arguments to the '<tt>div</tt>' instruction must be
1953 <a href="#t_floating">floating point</a> values. Both arguments must have
1954 identical types. This instruction can also take <a href="#t_vector">vector</a>
1955 versions of the values in which case the elements must be floating point.</p>
1957 <p>The value produced is the floating point quotient of the two operands.</p>
1959 <pre> <result> = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
1962 <!-- _______________________________________________________________________ -->
1963 <div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
1965 <div class="doc_text">
1967 <pre> <result> = urem <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1970 <p>The '<tt>urem</tt>' instruction returns the remainder from the
1971 unsigned division of its two arguments.</p>
1973 <p>The two arguments to the '<tt>urem</tt>' instruction must be
1974 <a href="#t_integer">integer</a> values. Both arguments must have identical
1977 <p>This instruction returns the unsigned integer <i>remainder</i> of a division.
1978 This instruction always performs an unsigned division to get the remainder,
1979 regardless of whether the arguments are unsigned or not.</p>
1981 <pre> <result> = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
1985 <!-- _______________________________________________________________________ -->
1986 <div class="doc_subsubsection"> <a name="i_srem">'<tt>srem</tt>'
1987 Instruction</a> </div>
1988 <div class="doc_text">
1990 <pre> <result> = srem <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1993 <p>The '<tt>srem</tt>' instruction returns the remainder from the
1994 signed division of its two operands.</p>
1996 <p>The two arguments to the '<tt>srem</tt>' instruction must be
1997 <a href="#t_integer">integer</a> values. Both arguments must have identical
2000 <p>This instruction returns the <i>remainder</i> of a division (where the result
2001 has the same sign as the divisor), not the <i>modulus</i> (where the
2002 result has the same sign as the dividend) of a value. For more
2003 information about the difference, see <a
2004 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
2007 <pre> <result> = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
2011 <!-- _______________________________________________________________________ -->
2012 <div class="doc_subsubsection"> <a name="i_frem">'<tt>frem</tt>'
2013 Instruction</a> </div>
2014 <div class="doc_text">
2016 <pre> <result> = frem <ty> <var1>, <var2> <i>; yields {ty}:result</i>
2019 <p>The '<tt>frem</tt>' instruction returns the remainder from the
2020 division of its two operands.</p>
2022 <p>The two arguments to the '<tt>frem</tt>' instruction must be
2023 <a href="#t_floating">floating point</a> values. Both arguments must have
2024 identical types.</p>
2026 <p>This instruction returns the <i>remainder</i> of a division.</p>
2028 <pre> <result> = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
2032 <!-- ======================================================================= -->
2033 <div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2034 Operations</a> </div>
2035 <div class="doc_text">
2036 <p>Bitwise binary operators are used to do various forms of
2037 bit-twiddling in a program. They are generally very efficient
2038 instructions and can commonly be strength reduced from other
2039 instructions. They require two operands, execute an operation on them,
2040 and produce a single value. The resulting value of the bitwise binary
2041 operators is always the same type as its first operand.</p>
2044 <!-- _______________________________________________________________________ -->
2045 <div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2046 Instruction</a> </div>
2047 <div class="doc_text">
2049 <pre> <result> = shl <ty> <var1>, <var2> <i>; yields {ty}:result</i>
2052 <p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2053 the left a specified number of bits.</p>
2055 <p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
2056 href="#t_integer">integer</a> type.</p>
2058 <p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
2059 <h5>Example:</h5><pre>
2060 <result> = shl i32 4, %var <i>; yields {i32}: 4 << %var</i>
2061 <result> = shl i32 4, 2 <i>; yields {i32}: 16</i>
2062 <result> = shl i32 1, 10 <i>; yields {i32}: 1024</i>
2065 <!-- _______________________________________________________________________ -->
2066 <div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2067 Instruction</a> </div>
2068 <div class="doc_text">
2070 <pre> <result> = lshr <ty> <var1>, <var2> <i>; yields {ty}:result</i>
2074 <p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2075 operand shifted to the right a specified number of bits.</p>
2078 <p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
2079 <a href="#t_integer">integer</a> type.</p>
2082 <p>This instruction always performs a logical shift right operation. The most
2083 significant bits of the result will be filled with zero bits after the
2088 <result> = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2089 <result> = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2090 <result> = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2091 <result> = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
2095 <!-- _______________________________________________________________________ -->
2096 <div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2097 Instruction</a> </div>
2098 <div class="doc_text">
2101 <pre> <result> = ashr <ty> <var1>, <var2> <i>; yields {ty}:result</i>
2105 <p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2106 operand shifted to the right a specified number of bits.</p>
2109 <p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
2110 <a href="#t_integer">integer</a> type.</p>
2113 <p>This instruction always performs an arithmetic shift right operation,
2114 The most significant bits of the result will be filled with the sign bit
2115 of <tt>var1</tt>.</p>
2119 <result> = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2120 <result> = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2121 <result> = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2122 <result> = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
2126 <!-- _______________________________________________________________________ -->
2127 <div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2128 Instruction</a> </div>
2129 <div class="doc_text">
2131 <pre> <result> = and <ty> <var1>, <var2> <i>; yields {ty}:result</i>
2134 <p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2135 its two operands.</p>
2137 <p>The two arguments to the '<tt>and</tt>' instruction must be <a
2138 href="#t_integer">integer</a> values. Both arguments must have
2139 identical types.</p>
2141 <p>The truth table used for the '<tt>and</tt>' instruction is:</p>
2143 <div style="align: center">
2144 <table border="1" cellspacing="0" cellpadding="4">
2175 <pre> <result> = and i32 4, %var <i>; yields {i32}:result = 4 & %var</i>
2176 <result> = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2177 <result> = and i32 4, 8 <i>; yields {i32}:result = 0</i>
2180 <!-- _______________________________________________________________________ -->
2181 <div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
2182 <div class="doc_text">
2184 <pre> <result> = or <ty> <var1>, <var2> <i>; yields {ty}:result</i>
2187 <p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2188 or of its two operands.</p>
2190 <p>The two arguments to the '<tt>or</tt>' instruction must be <a
2191 href="#t_integer">integer</a> values. Both arguments must have
2192 identical types.</p>
2194 <p>The truth table used for the '<tt>or</tt>' instruction is:</p>
2196 <div style="align: center">
2197 <table border="1" cellspacing="0" cellpadding="4">
2228 <pre> <result> = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2229 <result> = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2230 <result> = or i32 4, 8 <i>; yields {i32}:result = 12</i>
2233 <!-- _______________________________________________________________________ -->
2234 <div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2235 Instruction</a> </div>
2236 <div class="doc_text">
2238 <pre> <result> = xor <ty> <var1>, <var2> <i>; yields {ty}:result</i>
2241 <p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2242 or of its two operands. The <tt>xor</tt> is used to implement the
2243 "one's complement" operation, which is the "~" operator in C.</p>
2245 <p>The two arguments to the '<tt>xor</tt>' instruction must be <a
2246 href="#t_integer">integer</a> values. Both arguments must have
2247 identical types.</p>
2249 <p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
2251 <div style="align: center">
2252 <table border="1" cellspacing="0" cellpadding="4">
2284 <pre> <result> = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2285 <result> = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2286 <result> = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2287 <result> = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
2291 <!-- ======================================================================= -->
2292 <div class="doc_subsection">
2293 <a name="vectorops">Vector Operations</a>
2296 <div class="doc_text">
2298 <p>LLVM supports several instructions to represent vector operations in a
2299 target-independent manner. This instructions cover the element-access and
2300 vector-specific operations needed to process vectors effectively. While LLVM
2301 does directly support these vector operations, many sophisticated algorithms
2302 will want to use target-specific intrinsics to take full advantage of a specific
2307 <!-- _______________________________________________________________________ -->
2308 <div class="doc_subsubsection">
2309 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2312 <div class="doc_text">
2317 <result> = extractelement <n x <ty>> <val>, i32 <idx> <i>; yields <ty></i>
2323 The '<tt>extractelement</tt>' instruction extracts a single scalar
2324 element from a vector at a specified index.
2331 The first operand of an '<tt>extractelement</tt>' instruction is a
2332 value of <a href="#t_vector">vector</a> type. The second operand is
2333 an index indicating the position from which to extract the element.
2334 The index may be a variable.</p>
2339 The result is a scalar of the same type as the element type of
2340 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2341 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2342 results are undefined.
2348 %result = extractelement <4 x i32> %vec, i32 0 <i>; yields i32</i>
2353 <!-- _______________________________________________________________________ -->
2354 <div class="doc_subsubsection">
2355 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2358 <div class="doc_text">
2363 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, i32 <idx> <i>; yields <n x <ty>></i>
2369 The '<tt>insertelement</tt>' instruction inserts a scalar
2370 element into a vector at a specified index.
2377 The first operand of an '<tt>insertelement</tt>' instruction is a
2378 value of <a href="#t_vector">vector</a> type. The second operand is a
2379 scalar value whose type must equal the element type of the first
2380 operand. The third operand is an index indicating the position at
2381 which to insert the value. The index may be a variable.</p>
2386 The result is a vector of the same type as <tt>val</tt>. Its
2387 element values are those of <tt>val</tt> except at position
2388 <tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2389 exceeds the length of <tt>val</tt>, the results are undefined.
2395 %result = insertelement <4 x i32> %vec, i32 1, i32 0 <i>; yields <4 x i32></i>
2399 <!-- _______________________________________________________________________ -->
2400 <div class="doc_subsubsection">
2401 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2404 <div class="doc_text">
2409 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <n x i32> <mask> <i>; yields <n x <ty>></i>
2415 The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2416 from two input vectors, returning a vector of the same type.
2422 The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2423 with types that match each other and types that match the result of the
2424 instruction. The third argument is a shuffle mask, which has the same number
2425 of elements as the other vector type, but whose element type is always 'i32'.
2429 The shuffle mask operand is required to be a constant vector with either
2430 constant integer or undef values.
2436 The elements of the two input vectors are numbered from left to right across
2437 both of the vectors. The shuffle mask operand specifies, for each element of
2438 the result vector, which element of the two input registers the result element
2439 gets. The element selector may be undef (meaning "don't care") and the second
2440 operand may be undef if performing a shuffle from only one vector.
2446 %result = shufflevector <4 x i32> %v1, <4 x i32> %v2,
2447 <4 x i32> <i32 0, i32 4, i32 1, i32 5> <i>; yields <4 x i32></i>
2448 %result = shufflevector <4 x i32> %v1, <4 x i32> undef,
2449 <4 x i32> <i32 0, i32 1, i32 2, i32 3> <i>; yields <4 x i32></i> - Identity shuffle.
2454 <!-- ======================================================================= -->
2455 <div class="doc_subsection">
2456 <a name="memoryops">Memory Access and Addressing Operations</a>
2459 <div class="doc_text">
2461 <p>A key design point of an SSA-based representation is how it
2462 represents memory. In LLVM, no memory locations are in SSA form, which
2463 makes things very simple. This section describes how to read, write,
2464 allocate, and free memory in LLVM.</p>
2468 <!-- _______________________________________________________________________ -->
2469 <div class="doc_subsubsection">
2470 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2473 <div class="doc_text">
2478 <result> = malloc <type>[, i32 <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
2483 <p>The '<tt>malloc</tt>' instruction allocates memory from the system
2484 heap and returns a pointer to it.</p>
2488 <p>The '<tt>malloc</tt>' instruction allocates
2489 <tt>sizeof(<type>)*NumElements</tt>
2490 bytes of memory from the operating system and returns a pointer of the
2491 appropriate type to the program. If "NumElements" is specified, it is the
2492 number of elements allocated. If an alignment is specified, the value result
2493 of the allocation is guaranteed to be aligned to at least that boundary. If
2494 not specified, or if zero, the target can choose to align the allocation on any
2495 convenient boundary.</p>
2497 <p>'<tt>type</tt>' must be a sized type.</p>
2501 <p>Memory is allocated using the system "<tt>malloc</tt>" function, and
2502 a pointer is returned.</p>
2507 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
2509 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
2510 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
2511 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
2512 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
2513 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
2517 <!-- _______________________________________________________________________ -->
2518 <div class="doc_subsubsection">
2519 <a name="i_free">'<tt>free</tt>' Instruction</a>
2522 <div class="doc_text">
2527 free <type> <value> <i>; yields {void}</i>
2532 <p>The '<tt>free</tt>' instruction returns memory back to the unused
2533 memory heap to be reallocated in the future.</p>
2537 <p>'<tt>value</tt>' shall be a pointer value that points to a value
2538 that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2543 <p>Access to the memory pointed to by the pointer is no longer defined
2544 after this instruction executes.</p>
2549 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
2550 free [4 x i8]* %array
2554 <!-- _______________________________________________________________________ -->
2555 <div class="doc_subsubsection">
2556 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2559 <div class="doc_text">
2564 <result> = alloca <type>[, i32 <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
2569 <p>The '<tt>alloca</tt>' instruction allocates memory on the current
2570 stack frame of the procedure that is live until the current function
2571 returns to its caller.</p>
2575 <p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(<type>)*NumElements</tt>
2576 bytes of memory on the runtime stack, returning a pointer of the
2577 appropriate type to the program. If "NumElements" is specified, it is the
2578 number of elements allocated. If an alignment is specified, the value result
2579 of the allocation is guaranteed to be aligned to at least that boundary. If
2580 not specified, or if zero, the target can choose to align the allocation on any
2581 convenient boundary.</p>
2583 <p>'<tt>type</tt>' may be any sized type.</p>
2587 <p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
2588 memory is automatically released when the function returns. The '<tt>alloca</tt>'
2589 instruction is commonly used to represent automatic variables that must
2590 have an address available. When the function returns (either with the <tt><a
2591 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
2592 instructions), the memory is reclaimed.</p>
2597 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
2598 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
2599 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
2600 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
2604 <!-- _______________________________________________________________________ -->
2605 <div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2606 Instruction</a> </div>
2607 <div class="doc_text">
2609 <pre> <result> = load <ty>* <pointer><br> <result> = volatile load <ty>* <pointer><br></pre>
2611 <p>The '<tt>load</tt>' instruction is used to read from memory.</p>
2613 <p>The argument to the '<tt>load</tt>' instruction specifies the memory
2614 address from which to load. The pointer must point to a <a
2615 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
2616 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
2617 the number or order of execution of this <tt>load</tt> with other
2618 volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2621 <p>The location of memory pointed to is loaded.</p>
2623 <pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
2625 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
2626 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
2629 <!-- _______________________________________________________________________ -->
2630 <div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2631 Instruction</a> </div>
2632 <div class="doc_text">
2634 <pre> store <ty> <value>, <ty>* <pointer> <i>; yields {void}</i>
2635 volatile store <ty> <value>, <ty>* <pointer> <i>; yields {void}</i>
2638 <p>The '<tt>store</tt>' instruction is used to write to memory.</p>
2640 <p>There are two arguments to the '<tt>store</tt>' instruction: a value
2641 to store and an address in which to store it. The type of the '<tt><pointer></tt>'
2642 operand must be a pointer to the type of the '<tt><value></tt>'
2643 operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
2644 optimizer is not allowed to modify the number or order of execution of
2645 this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2646 href="#i_store">store</a></tt> instructions.</p>
2648 <p>The contents of memory are updated to contain '<tt><value></tt>'
2649 at the location specified by the '<tt><pointer></tt>' operand.</p>
2651 <pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
2653 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
2654 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
2658 <!-- _______________________________________________________________________ -->
2659 <div class="doc_subsubsection">
2660 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2663 <div class="doc_text">
2666 <result> = getelementptr <ty>* <ptrval>{, <ty> <idx>}*
2672 The '<tt>getelementptr</tt>' instruction is used to get the address of a
2673 subelement of an aggregate data structure.</p>
2677 <p>This instruction takes a list of integer operands that indicate what
2678 elements of the aggregate object to index to. The actual types of the arguments
2679 provided depend on the type of the first pointer argument. The
2680 '<tt>getelementptr</tt>' instruction is used to index down through the type
2681 levels of a structure or to a specific index in an array. When indexing into a
2682 structure, only <tt>i32</tt> integer constants are allowed. When indexing
2683 into an array or pointer, only integers of 32 or 64 bits are allowed, and will
2684 be sign extended to 64-bit values.</p>
2686 <p>For example, let's consider a C code fragment and how it gets
2687 compiled to LLVM:</p>
2701 define i32 *foo(struct ST *s) {
2702 return &s[1].Z.B[5][13];
2706 <p>The LLVM code generated by the GCC frontend is:</p>
2709 %RT = type { i8 , [10 x [20 x i32]], i8 }
2710 %ST = type { i32, double, %RT }
2714 define i32* %foo(%ST* %s) {
2716 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
2723 <p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
2724 on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
2725 and <a href="#t_array">array</a> types can use a 32-bit or 64-bit
2726 <a href="#t_integer">integer</a> type but the value will always be sign extended
2727 to 64-bits. <a href="#t_struct">Structure</a> types, require <tt>i32</tt>
2728 <b>constants</b>.</p>
2730 <p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
2731 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
2732 }</tt>' type, a structure. The second index indexes into the third element of
2733 the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
2734 i8 }</tt>' type, another structure. The third index indexes into the second
2735 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
2736 array. The two dimensions of the array are subscripted into, yielding an
2737 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
2738 to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
2740 <p>Note that it is perfectly legal to index partially through a
2741 structure, returning a pointer to an inner element. Because of this,
2742 the LLVM code for the given testcase is equivalent to:</p>
2745 define i32* %foo(%ST* %s) {
2746 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
2747 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
2748 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
2749 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
2750 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
2755 <p>Note that it is undefined to access an array out of bounds: array and
2756 pointer indexes must always be within the defined bounds of the array type.
2757 The one exception for this rules is zero length arrays. These arrays are
2758 defined to be accessible as variable length arrays, which requires access
2759 beyond the zero'th element.</p>
2761 <p>The getelementptr instruction is often confusing. For some more insight
2762 into how it works, see <a href="GetElementPtr.html">the getelementptr
2768 <i>; yields [12 x i8]*:aptr</i>
2769 %aptr = getelementptr {i32, [12 x i8]}* %sptr, i64 0, i32 1
2773 <!-- ======================================================================= -->
2774 <div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
2776 <div class="doc_text">
2777 <p>The instructions in this category are the conversion instructions (casting)
2778 which all take a single operand and a type. They perform various bit conversions
2782 <!-- _______________________________________________________________________ -->
2783 <div class="doc_subsubsection">
2784 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
2786 <div class="doc_text">
2790 <result> = trunc <ty> <value> to <ty2> <i>; yields ty2</i>
2795 The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
2800 The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
2801 be an <a href="#t_integer">integer</a> type, and a type that specifies the size
2802 and type of the result, which must be an <a href="#t_integer">integer</a>
2803 type. The bit size of <tt>value</tt> must be larger than the bit size of
2804 <tt>ty2</tt>. Equal sized types are not allowed.</p>
2808 The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
2809 and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
2810 larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
2811 It will always truncate bits.</p>
2815 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
2816 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
2817 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
2821 <!-- _______________________________________________________________________ -->
2822 <div class="doc_subsubsection">
2823 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
2825 <div class="doc_text">
2829 <result> = zext <ty> <value> to <ty2> <i>; yields ty2</i>
2833 <p>The '<tt>zext</tt>' instruction zero extends its operand to type
2838 <p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
2839 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
2840 also be of <a href="#t_integer">integer</a> type. The bit size of the
2841 <tt>value</tt> must be smaller than the bit size of the destination type,
2845 <p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
2846 bits until it reaches the size of the destination type, <tt>ty2</tt>. When the
2847 the operand and the type are the same size, no bit filling is done and the
2848 cast is considered a <i>no-op cast</i> because no bits change (only the type
2851 <p>When zero extending from i1, the result will always be either 0 or 1.</p>
2855 %X = zext i32 257 to i64 <i>; yields i64:257</i>
2856 %Y = zext i1 true to i32 <i>; yields i32:1</i>
2860 <!-- _______________________________________________________________________ -->
2861 <div class="doc_subsubsection">
2862 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
2864 <div class="doc_text">
2868 <result> = sext <ty> <value> to <ty2> <i>; yields ty2</i>
2872 <p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
2876 The '<tt>sext</tt>' instruction takes a value to cast, which must be of
2877 <a href="#t_integer">integer</a> type, and a type to cast it to, which must
2878 also be of <a href="#t_integer">integer</a> type. The bit size of the
2879 <tt>value</tt> must be smaller than the bit size of the destination type,
2884 The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
2885 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
2886 the type <tt>ty2</tt>. When the the operand and the type are the same size,
2887 no bit filling is done and the cast is considered a <i>no-op cast</i> because
2888 no bits change (only the type changes).</p>
2890 <p>When sign extending from i1, the extension always results in -1 or 0.</p>
2894 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
2895 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
2899 <!-- _______________________________________________________________________ -->
2900 <div class="doc_subsubsection">
2901 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
2904 <div class="doc_text">
2909 <result> = fptrunc <ty> <value> to <ty2> <i>; yields ty2</i>
2913 <p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
2918 <p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
2919 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
2920 cast it to. The size of <tt>value</tt> must be larger than the size of
2921 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
2922 <i>no-op cast</i>.</p>
2925 <p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
2926 <a href="#t_floating">floating point</a> type to a smaller
2927 <a href="#t_floating">floating point</a> type. If the value cannot fit within
2928 the destination type, <tt>ty2</tt>, then the results are undefined.</p>
2932 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
2933 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
2937 <!-- _______________________________________________________________________ -->
2938 <div class="doc_subsubsection">
2939 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
2941 <div class="doc_text">
2945 <result> = fpext <ty> <value> to <ty2> <i>; yields ty2</i>
2949 <p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
2950 floating point value.</p>
2953 <p>The '<tt>fpext</tt>' instruction takes a
2954 <a href="#t_floating">floating point</a> <tt>value</tt> to cast,
2955 and a <a href="#t_floating">floating point</a> type to cast it to. The source
2956 type must be smaller than the destination type.</p>
2959 <p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
2960 <a href="t_floating">floating point</a> type to a larger
2961 <a href="t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
2962 used to make a <i>no-op cast</i> because it always changes bits. Use
2963 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
2967 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
2968 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
2972 <!-- _______________________________________________________________________ -->
2973 <div class="doc_subsubsection">
2974 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
2976 <div class="doc_text">
2980 <result> = fp2uint <ty> <value> to <ty2> <i>; yields ty2</i>
2984 <p>The '<tt>fp2uint</tt>' converts a floating point <tt>value</tt> to its
2985 unsigned integer equivalent of type <tt>ty2</tt>.
2989 <p>The '<tt>fp2uint</tt>' instruction takes a value to cast, which must be a
2990 <a href="#t_floating">floating point</a> value, and a type to cast it to, which
2991 must be an <a href="#t_integer">integer</a> type.</p>
2994 <p> The '<tt>fp2uint</tt>' instruction converts its
2995 <a href="#t_floating">floating point</a> operand into the nearest (rounding
2996 towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
2997 the results are undefined.</p>
2999 <p>When converting to i1, the conversion is done as a comparison against
3000 zero. If the <tt>value</tt> was zero, the i1 result will be <tt>false</tt>.
3001 If the <tt>value</tt> was non-zero, the i1 result will be <tt>true</tt>.</p>
3005 %X = fp2uint double 123.0 to i32 <i>; yields i32:123</i>
3006 %Y = fp2uint float 1.0E+300 to i1 <i>; yields i1:true</i>
3007 %X = fp2uint float 1.04E+17 to i8 <i>; yields undefined:1</i>
3011 <!-- _______________________________________________________________________ -->
3012 <div class="doc_subsubsection">
3013 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
3015 <div class="doc_text">
3019 <result> = fptosi <ty> <value> to <ty2> <i>; yields ty2</i>
3023 <p>The '<tt>fptosi</tt>' instruction converts
3024 <a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
3029 <p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
3030 <a href="#t_floating">floating point</a> value, and a type to cast it to, which
3031 must also be an <a href="#t_integer">integer</a> type.</p>
3034 <p>The '<tt>fptosi</tt>' instruction converts its
3035 <a href="#t_floating">floating point</a> operand into the nearest (rounding
3036 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3037 the results are undefined.</p>
3039 <p>When converting to i1, the conversion is done as a comparison against
3040 zero. If the <tt>value</tt> was zero, the i1 result will be <tt>false</tt>.
3041 If the <tt>value</tt> was non-zero, the i1 result will be <tt>true</tt>.</p>
3045 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
3046 %Y = fptosi float 1.0E-247 to i1 <i>; yields i1:true</i>
3047 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
3051 <!-- _______________________________________________________________________ -->
3052 <div class="doc_subsubsection">
3053 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
3055 <div class="doc_text">
3059 <result> = uitofp <ty> <value> to <ty2> <i>; yields ty2</i>
3063 <p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
3064 integer and converts that value to the <tt>ty2</tt> type.</p>
3068 <p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be an
3069 <a href="#t_integer">integer</a> value, and a type to cast it to, which must
3070 be a <a href="#t_floating">floating point</a> type.</p>
3073 <p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
3074 integer quantity and converts it to the corresponding floating point value. If
3075 the value cannot fit in the floating point value, the results are undefined.</p>
3080 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
3081 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
3085 <!-- _______________________________________________________________________ -->
3086 <div class="doc_subsubsection">
3087 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
3089 <div class="doc_text">
3093 <result> = sitofp <ty> <value> to <ty2> <i>; yields ty2</i>
3097 <p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
3098 integer and converts that value to the <tt>ty2</tt> type.</p>
3101 <p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be an
3102 <a href="#t_integer">integer</a> value, and a type to cast it to, which must be
3103 a <a href="#t_floating">floating point</a> type.</p>
3106 <p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
3107 integer quantity and converts it to the corresponding floating point value. If
3108 the value cannot fit in the floating point value, the results are undefined.</p>
3112 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
3113 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
3117 <!-- _______________________________________________________________________ -->
3118 <div class="doc_subsubsection">
3119 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3121 <div class="doc_text">
3125 <result> = ptrtoint <ty> <value> to <ty2> <i>; yields ty2</i>
3129 <p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3130 the integer type <tt>ty2</tt>.</p>
3133 <p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
3134 must be a <a href="t_pointer">pointer</a> value, and a type to cast it to
3135 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
3138 <p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3139 <tt>ty2</tt> by interpreting the pointer value as an integer and either
3140 truncating or zero extending that value to the size of the integer type. If
3141 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3142 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
3143 are the same size, then nothing is done (<i>no-op cast</i>).</p>
3147 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit</i>
3148 %Y = ptrtoint i32* %x to i64 <i>; yields zero extend on 32-bit</i>
3152 <!-- _______________________________________________________________________ -->
3153 <div class="doc_subsubsection">
3154 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3156 <div class="doc_text">
3160 <result> = inttoptr <ty> <value> to <ty2> <i>; yields ty2</i>
3164 <p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3165 a pointer type, <tt>ty2</tt>.</p>
3168 <p>The '<tt>inttoptr</tt>' instruction takes an <a href="i_integer">integer</a>
3169 value to cast, and a type to cast it to, which must be a
3170 <a href="#t_pointer">pointer</a> type.
3173 <p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3174 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
3175 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3176 size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3177 the size of a pointer then a zero extension is done. If they are the same size,
3178 nothing is done (<i>no-op cast</i>).</p>
3182 %X = inttoptr i32 255 to i32* <i>; yields zero extend on 64-bit</i>
3183 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit </i>
3184 %Y = inttoptr i16 0 to i32* <i>; yields zero extend on 32-bit</i>
3188 <!-- _______________________________________________________________________ -->
3189 <div class="doc_subsubsection">
3190 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
3192 <div class="doc_text">
3196 <result> = bitcast <ty> <value> to <ty2> <i>; yields ty2</i>
3200 <p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3201 <tt>ty2</tt> without changing any bits.</p>
3204 <p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
3205 a first class value, and a type to cast it to, which must also be a <a
3206 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
3207 and the destination type, <tt>ty2</tt>, must be identical. If the source
3208 type is a pointer, the destination type must also be a pointer.</p>
3211 <p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
3212 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3213 this conversion. The conversion is done as if the <tt>value</tt> had been
3214 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3215 converted to other pointer types with this instruction. To convert pointers to
3216 other types, use the <a href="#i_inttoptr">inttoptr</a> or
3217 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
3221 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
3222 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
3223 %Z = bitcast <2xint> %V to i64; <i>; yields i64: %V</i>
3227 <!-- ======================================================================= -->
3228 <div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3229 <div class="doc_text">
3230 <p>The instructions in this category are the "miscellaneous"
3231 instructions, which defy better classification.</p>
3234 <!-- _______________________________________________________________________ -->
3235 <div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3237 <div class="doc_text">
3239 <pre> <result> = icmp <cond> <ty> <var1>, <var2>
3240 <i>; yields {i1}:result</i>
3243 <p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
3244 of its two integer operands.</p>
3246 <p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3247 the condition code which indicates the kind of comparison to perform. It is not
3248 a value, just a keyword. The possibilities for the condition code are:
3250 <li><tt>eq</tt>: equal</li>
3251 <li><tt>ne</tt>: not equal </li>
3252 <li><tt>ugt</tt>: unsigned greater than</li>
3253 <li><tt>uge</tt>: unsigned greater or equal</li>
3254 <li><tt>ult</tt>: unsigned less than</li>
3255 <li><tt>ule</tt>: unsigned less or equal</li>
3256 <li><tt>sgt</tt>: signed greater than</li>
3257 <li><tt>sge</tt>: signed greater or equal</li>
3258 <li><tt>slt</tt>: signed less than</li>
3259 <li><tt>sle</tt>: signed less or equal</li>
3261 <p>The remaining two arguments must be <a href="#t_integer">integer</a> or
3262 <a href="#t_pointer">pointer</a> typed. They must also be identical types.</p>
3264 <p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3265 the condition code given as <tt>cond</tt>. The comparison performed always
3266 yields a <a href="#t_primitive">i1</a> result, as follows:
3268 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3269 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3271 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3272 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3273 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3274 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3275 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3276 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3277 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3278 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3279 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3280 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3281 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3282 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3283 <li><tt>sge</tt>: interprets the operands as signed values and yields
3284 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3285 <li><tt>slt</tt>: interprets the operands as signed values and yields
3286 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3287 <li><tt>sle</tt>: interprets the operands as signed values and yields
3288 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3290 <p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3291 values are treated as integers and then compared.</p>
3294 <pre> <result> = icmp eq i32 4, 5 <i>; yields: result=false</i>
3295 <result> = icmp ne float* %X, %X <i>; yields: result=false</i>
3296 <result> = icmp ult i16 4, 5 <i>; yields: result=true</i>
3297 <result> = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3298 <result> = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3299 <result> = icmp sge i16 4, 5 <i>; yields: result=false</i>
3303 <!-- _______________________________________________________________________ -->
3304 <div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3306 <div class="doc_text">
3308 <pre> <result> = fcmp <cond> <ty> <var1>, <var2>
3309 <i>; yields {i1}:result</i>
3312 <p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3313 of its floating point operands.</p>
3315 <p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3316 the condition code which indicates the kind of comparison to perform. It is not
3317 a value, just a keyword. The possibilities for the condition code are:
3319 <li><tt>false</tt>: no comparison, always returns false</li>
3320 <li><tt>oeq</tt>: ordered and equal</li>
3321 <li><tt>ogt</tt>: ordered and greater than </li>
3322 <li><tt>oge</tt>: ordered and greater than or equal</li>
3323 <li><tt>olt</tt>: ordered and less than </li>
3324 <li><tt>ole</tt>: ordered and less than or equal</li>
3325 <li><tt>one</tt>: ordered and not equal</li>
3326 <li><tt>ord</tt>: ordered (no nans)</li>
3327 <li><tt>ueq</tt>: unordered or equal</li>
3328 <li><tt>ugt</tt>: unordered or greater than </li>
3329 <li><tt>uge</tt>: unordered or greater than or equal</li>
3330 <li><tt>ult</tt>: unordered or less than </li>
3331 <li><tt>ule</tt>: unordered or less than or equal</li>
3332 <li><tt>une</tt>: unordered or not equal</li>
3333 <li><tt>uno</tt>: unordered (either nans)</li>
3334 <li><tt>true</tt>: no comparison, always returns true</li>
3336 <p>In the preceding, <i>ordered</i> means that neither operand is a QNAN while
3337 <i>unordered</i> means that either operand may be a QNAN.</p>
3338 <p>The <tt>val1</tt> and <tt>val2</tt> arguments must be
3339 <a href="#t_floating">floating point</a> typed. They must have identical
3341 <p>In the foregoing, <i>ordered</i> means that neither operand is a QNAN and
3342 <i>unordered</i> means that either operand is a QNAN.</p>
3344 <p>The '<tt>fcmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3345 the condition code given as <tt>cond</tt>. The comparison performed always
3346 yields a <a href="#t_primitive">i1</a> result, as follows:
3348 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
3349 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3350 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
3351 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3352 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
3353 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3354 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3355 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3356 <tt>var1</tt> is less than <tt>var2</tt>.</li>
3357 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3358 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3359 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
3360 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
3361 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3362 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
3363 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
3364 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
3365 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3366 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
3367 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3368 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
3369 <tt>var1</tt> is less than <tt>var2</tt>.</li>
3370 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
3371 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3372 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
3373 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
3374 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
3375 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3379 <pre> <result> = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3380 <result> = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3381 <result> = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3382 <result> = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3386 <!-- _______________________________________________________________________ -->
3387 <div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
3388 Instruction</a> </div>
3389 <div class="doc_text">
3391 <pre> <result> = phi <ty> [ <val0>, <label0>], ...<br></pre>
3393 <p>The '<tt>phi</tt>' instruction is used to implement the φ node in
3394 the SSA graph representing the function.</p>
3396 <p>The type of the incoming values are specified with the first type
3397 field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
3398 as arguments, with one pair for each predecessor basic block of the
3399 current block. Only values of <a href="#t_firstclass">first class</a>
3400 type may be used as the value arguments to the PHI node. Only labels
3401 may be used as the label arguments.</p>
3402 <p>There must be no non-phi instructions between the start of a basic
3403 block and the PHI instructions: i.e. PHI instructions must be first in
3406 <p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
3407 value specified by the parameter, depending on which basic block we
3408 came from in the last <a href="#terminators">terminator</a> instruction.</p>
3410 <pre>Loop: ; Infinite loop that counts from 0 on up...<br> %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]<br> %nextindvar = add i32 %indvar, 1<br> br label %Loop<br></pre>
3413 <!-- _______________________________________________________________________ -->
3414 <div class="doc_subsubsection">
3415 <a name="i_select">'<tt>select</tt>' Instruction</a>
3418 <div class="doc_text">
3423 <result> = select i1 <cond>, <ty> <val1>, <ty> <val2> <i>; yields ty</i>
3429 The '<tt>select</tt>' instruction is used to choose one value based on a
3430 condition, without branching.
3437 The '<tt>select</tt>' instruction requires a boolean value indicating the condition, and two values of the same <a href="#t_firstclass">first class</a> type.
3443 If the boolean condition evaluates to true, the instruction returns the first
3444 value argument; otherwise, it returns the second value argument.
3450 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
3455 <!-- _______________________________________________________________________ -->
3456 <div class="doc_subsubsection">
3457 <a name="i_call">'<tt>call</tt>' Instruction</a>
3460 <div class="doc_text">
3464 <result> = [tail] call [<a href="#callingconv">cconv</a>] <ty>* <fnptrval>(<param list>)
3469 <p>The '<tt>call</tt>' instruction represents a simple function call.</p>
3473 <p>This instruction requires several arguments:</p>
3477 <p>The optional "tail" marker indicates whether the callee function accesses
3478 any allocas or varargs in the caller. If the "tail" marker is present, the
3479 function call is eligible for tail call optimization. Note that calls may
3480 be marked "tail" even if they do not occur before a <a
3481 href="#i_ret"><tt>ret</tt></a> instruction.
3484 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
3485 convention</a> the call should use. If none is specified, the call defaults
3486 to using C calling conventions.
3489 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
3490 being invoked. The argument types must match the types implied by this
3491 signature. This type can be omitted if the function is not varargs and
3492 if the function type does not return a pointer to a function.</p>
3495 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
3496 be invoked. In most cases, this is a direct function invocation, but
3497 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
3498 to function value.</p>
3501 <p>'<tt>function args</tt>': argument list whose types match the
3502 function signature argument types. All arguments must be of
3503 <a href="#t_firstclass">first class</a> type. If the function signature
3504 indicates the function accepts a variable number of arguments, the extra
3505 arguments can be specified.</p>
3511 <p>The '<tt>call</tt>' instruction is used to cause control flow to
3512 transfer to a specified function, with its incoming arguments bound to
3513 the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
3514 instruction in the called function, control flow continues with the
3515 instruction after the function call, and the return value of the
3516 function is bound to the result argument. This is a simpler case of
3517 the <a href="#i_invoke">invoke</a> instruction.</p>
3522 %retval = call i32 %test(i32 %argc)
3523 call i32(i8 *, ...) *%printf(i8 * %msg, i32 12, i8 42);
3524 %X = tail call i32 %foo()
3525 %Y = tail call <a href="#callingconv">fastcc</a> i32 %foo()
3530 <!-- _______________________________________________________________________ -->
3531 <div class="doc_subsubsection">
3532 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
3535 <div class="doc_text">
3540 <resultval> = va_arg <va_list*> <arglist>, <argty>
3545 <p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
3546 the "variable argument" area of a function call. It is used to implement the
3547 <tt>va_arg</tt> macro in C.</p>
3551 <p>This instruction takes a <tt>va_list*</tt> value and the type of
3552 the argument. It returns a value of the specified argument type and
3553 increments the <tt>va_list</tt> to point to the next argument. Again, the
3554 actual type of <tt>va_list</tt> is target specific.</p>
3558 <p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
3559 type from the specified <tt>va_list</tt> and causes the
3560 <tt>va_list</tt> to point to the next argument. For more information,
3561 see the variable argument handling <a href="#int_varargs">Intrinsic
3564 <p>It is legal for this instruction to be called in a function which does not
3565 take a variable number of arguments, for example, the <tt>vfprintf</tt>
3568 <p><tt>va_arg</tt> is an LLVM instruction instead of an <a
3569 href="#intrinsics">intrinsic function</a> because it takes a type as an
3574 <p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
3578 <!-- *********************************************************************** -->
3579 <div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
3580 <!-- *********************************************************************** -->
3582 <div class="doc_text">
3584 <p>LLVM supports the notion of an "intrinsic function". These functions have
3585 well known names and semantics and are required to follow certain
3586 restrictions. Overall, these instructions represent an extension mechanism for
3587 the LLVM language that does not require changing all of the transformations in
3588 LLVM to add to the language (or the bytecode reader/writer, the parser,
3591 <p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
3592 prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
3593 this. Intrinsic functions must always be external functions: you cannot define
3594 the body of intrinsic functions. Intrinsic functions may only be used in call
3595 or invoke instructions: it is illegal to take the address of an intrinsic
3596 function. Additionally, because intrinsic functions are part of the LLVM
3597 language, it is required that they all be documented here if any are added.</p>
3600 <p>To learn how to add an intrinsic function, please see the <a
3601 href="ExtendingLLVM.html">Extending LLVM Guide</a>.
3606 <!-- ======================================================================= -->
3607 <div class="doc_subsection">
3608 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
3611 <div class="doc_text">
3613 <p>Variable argument support is defined in LLVM with the <a
3614 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
3615 intrinsic functions. These functions are related to the similarly
3616 named macros defined in the <tt><stdarg.h></tt> header file.</p>
3618 <p>All of these functions operate on arguments that use a
3619 target-specific value type "<tt>va_list</tt>". The LLVM assembly
3620 language reference manual does not define what this type is, so all
3621 transformations should be prepared to handle intrinsics with any type
3624 <p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
3625 instruction and the variable argument handling intrinsic functions are
3629 define i32 %test(i32 %X, ...) {
3630 ; Initialize variable argument processing
3632 %ap2 = bitcast i8** %ap to i8*
3633 call void %<a href="#i_va_start">llvm.va_start</a>(i8* %ap2)
3635 ; Read a single integer argument
3636 %tmp = va_arg i8 ** %ap, i32
3638 ; Demonstrate usage of llvm.va_copy and llvm.va_end
3640 %aq2 = bitcast i8** %aq to i8*
3641 call void %<a href="#i_va_copy">llvm.va_copy</a>(i8 *%aq2, i8* %ap2)
3642 call void %<a href="#i_va_end">llvm.va_end</a>(i8* %aq2)
3644 ; Stop processing of arguments.
3645 call void %<a href="#i_va_end">llvm.va_end</a>(i8* %ap2)
3651 <!-- _______________________________________________________________________ -->
3652 <div class="doc_subsubsection">
3653 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
3657 <div class="doc_text">
3659 <pre> declare void %llvm.va_start(i8* <arglist>)<br></pre>
3661 <P>The '<tt>llvm.va_start</tt>' intrinsic initializes
3662 <tt>*<arglist></tt> for subsequent use by <tt><a
3663 href="#i_va_arg">va_arg</a></tt>.</p>
3667 <P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
3671 <P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
3672 macro available in C. In a target-dependent way, it initializes the
3673 <tt>va_list</tt> element the argument points to, so that the next call to
3674 <tt>va_arg</tt> will produce the first variable argument passed to the function.
3675 Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
3676 last argument of the function, the compiler can figure that out.</p>
3680 <!-- _______________________________________________________________________ -->
3681 <div class="doc_subsubsection">
3682 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
3685 <div class="doc_text">
3687 <pre> declare void %llvm.va_end(i8* <arglist>)<br></pre>
3690 <p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt><arglist></tt>
3691 which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
3692 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
3696 <p>The argument is a <tt>va_list</tt> to destroy.</p>
3700 <p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
3701 macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
3702 Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
3703 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
3704 with calls to <tt>llvm.va_end</tt>.</p>
3708 <!-- _______________________________________________________________________ -->
3709 <div class="doc_subsubsection">
3710 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
3713 <div class="doc_text">
3718 declare void %llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
3723 <p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
3724 the source argument list to the destination argument list.</p>
3728 <p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
3729 The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
3734 <p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
3735 available in C. In a target-dependent way, it copies the source
3736 <tt>va_list</tt> element into the destination list. This intrinsic is necessary
3737 because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
3738 arbitrarily complex and require memory allocation, for example.</p>
3742 <!-- ======================================================================= -->
3743 <div class="doc_subsection">
3744 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
3747 <div class="doc_text">
3750 LLVM support for <a href="GarbageCollection.html">Accurate Garbage
3751 Collection</a> requires the implementation and generation of these intrinsics.
3752 These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
3753 stack</a>, as well as garbage collector implementations that require <a
3754 href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
3755 Front-ends for type-safe garbage collected languages should generate these
3756 intrinsics to make use of the LLVM garbage collectors. For more details, see <a
3757 href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
3761 <!-- _______________________________________________________________________ -->
3762 <div class="doc_subsubsection">
3763 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
3766 <div class="doc_text">
3771 declare void %llvm.gcroot(<ty>** %ptrloc, <ty2>* %metadata)
3776 <p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
3777 the code generator, and allows some metadata to be associated with it.</p>
3781 <p>The first argument specifies the address of a stack object that contains the
3782 root pointer. The second pointer (which must be either a constant or a global
3783 value address) contains the meta-data to be associated with the root.</p>
3787 <p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
3788 location. At compile-time, the code generator generates information to allow
3789 the runtime to find the pointer at GC safe points.
3795 <!-- _______________________________________________________________________ -->
3796 <div class="doc_subsubsection">
3797 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
3800 <div class="doc_text">
3805 declare i8 * %llvm.gcread(i8 * %ObjPtr, i8 ** %Ptr)
3810 <p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
3811 locations, allowing garbage collector implementations that require read
3816 <p>The second argument is the address to read from, which should be an address
3817 allocated from the garbage collector. The first object is a pointer to the
3818 start of the referenced object, if needed by the language runtime (otherwise
3823 <p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
3824 instruction, but may be replaced with substantially more complex code by the
3825 garbage collector runtime, as needed.</p>
3830 <!-- _______________________________________________________________________ -->
3831 <div class="doc_subsubsection">
3832 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
3835 <div class="doc_text">
3840 declare void %llvm.gcwrite(i8 * %P1, i8 * %Obj, i8 ** %P2)
3845 <p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
3846 locations, allowing garbage collector implementations that require write
3847 barriers (such as generational or reference counting collectors).</p>
3851 <p>The first argument is the reference to store, the second is the start of the
3852 object to store it to, and the third is the address of the field of Obj to
3853 store to. If the runtime does not require a pointer to the object, Obj may be
3858 <p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
3859 instruction, but may be replaced with substantially more complex code by the
3860 garbage collector runtime, as needed.</p>
3866 <!-- ======================================================================= -->
3867 <div class="doc_subsection">
3868 <a name="int_codegen">Code Generator Intrinsics</a>
3871 <div class="doc_text">
3873 These intrinsics are provided by LLVM to expose special features that may only
3874 be implemented with code generator support.
3879 <!-- _______________________________________________________________________ -->
3880 <div class="doc_subsubsection">
3881 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
3884 <div class="doc_text">
3888 declare i8 *%llvm.returnaddress(i32 <level>)
3894 The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
3895 target-specific value indicating the return address of the current function
3896 or one of its callers.
3902 The argument to this intrinsic indicates which function to return the address
3903 for. Zero indicates the calling function, one indicates its caller, etc. The
3904 argument is <b>required</b> to be a constant integer value.
3910 The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
3911 the return address of the specified call frame, or zero if it cannot be
3912 identified. The value returned by this intrinsic is likely to be incorrect or 0
3913 for arguments other than zero, so it should only be used for debugging purposes.
3917 Note that calling this intrinsic does not prevent function inlining or other
3918 aggressive transformations, so the value returned may not be that of the obvious
3919 source-language caller.
3924 <!-- _______________________________________________________________________ -->
3925 <div class="doc_subsubsection">
3926 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
3929 <div class="doc_text">
3933 declare i8 *%llvm.frameaddress(i32 <level>)
3939 The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
3940 target-specific frame pointer value for the specified stack frame.
3946 The argument to this intrinsic indicates which function to return the frame
3947 pointer for. Zero indicates the calling function, one indicates its caller,
3948 etc. The argument is <b>required</b> to be a constant integer value.
3954 The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
3955 the frame address of the specified call frame, or zero if it cannot be
3956 identified. The value returned by this intrinsic is likely to be incorrect or 0
3957 for arguments other than zero, so it should only be used for debugging purposes.
3961 Note that calling this intrinsic does not prevent function inlining or other
3962 aggressive transformations, so the value returned may not be that of the obvious
3963 source-language caller.
3967 <!-- _______________________________________________________________________ -->
3968 <div class="doc_subsubsection">
3969 <a name="i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
3972 <div class="doc_text">
3976 declare i8 *%llvm.stacksave()
3982 The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
3983 the function stack, for use with <a href="#i_stackrestore">
3984 <tt>llvm.stackrestore</tt></a>. This is useful for implementing language
3985 features like scoped automatic variable sized arrays in C99.
3991 This intrinsic returns a opaque pointer value that can be passed to <a
3992 href="#i_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
3993 <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
3994 <tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
3995 state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
3996 practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
3997 that were allocated after the <tt>llvm.stacksave</tt> was executed.
4002 <!-- _______________________________________________________________________ -->
4003 <div class="doc_subsubsection">
4004 <a name="i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
4007 <div class="doc_text">
4011 declare void %llvm.stackrestore(i8 * %ptr)
4017 The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4018 the function stack to the state it was in when the corresponding <a
4019 href="#llvm.stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
4020 useful for implementing language features like scoped automatic variable sized
4027 See the description for <a href="#i_stacksave"><tt>llvm.stacksave</tt></a>.
4033 <!-- _______________________________________________________________________ -->
4034 <div class="doc_subsubsection">
4035 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
4038 <div class="doc_text">
4042 declare void %llvm.prefetch(i8 * <address>,
4043 i32 <rw>, i32 <locality>)
4050 The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
4051 a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4053 effect on the behavior of the program but can change its performance
4060 <tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4061 determining if the fetch should be for a read (0) or write (1), and
4062 <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
4063 locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
4064 <tt>locality</tt> arguments must be constant integers.
4070 This intrinsic does not modify the behavior of the program. In particular,
4071 prefetches cannot trap and do not produce a value. On targets that support this
4072 intrinsic, the prefetch can provide hints to the processor cache for better
4078 <!-- _______________________________________________________________________ -->
4079 <div class="doc_subsubsection">
4080 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
4083 <div class="doc_text">
4087 declare void %llvm.pcmarker( i32 <id> )
4094 The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
4096 code to simulators and other tools. The method is target specific, but it is
4097 expected that the marker will use exported symbols to transmit the PC of the marker.
4098 The marker makes no guarantees that it will remain with any specific instruction
4099 after optimizations. It is possible that the presence of a marker will inhibit
4100 optimizations. The intended use is to be inserted after optimizations to allow
4101 correlations of simulation runs.
4107 <tt>id</tt> is a numerical id identifying the marker.
4113 This intrinsic does not modify the behavior of the program. Backends that do not
4114 support this intrinisic may ignore it.
4119 <!-- _______________________________________________________________________ -->
4120 <div class="doc_subsubsection">
4121 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
4124 <div class="doc_text">
4128 declare i64 %llvm.readcyclecounter( )
4135 The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
4136 counter register (or similar low latency, high accuracy clocks) on those targets
4137 that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
4138 As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
4139 should only be used for small timings.
4145 When directly supported, reading the cycle counter should not modify any memory.
4146 Implementations are allowed to either return a application specific value or a
4147 system wide value. On backends without support, this is lowered to a constant 0.
4152 <!-- ======================================================================= -->
4153 <div class="doc_subsection">
4154 <a name="int_libc">Standard C Library Intrinsics</a>
4157 <div class="doc_text">
4159 LLVM provides intrinsics for a few important standard C library functions.
4160 These intrinsics allow source-language front-ends to pass information about the
4161 alignment of the pointer arguments to the code generator, providing opportunity
4162 for more efficient code generation.
4167 <!-- _______________________________________________________________________ -->
4168 <div class="doc_subsubsection">
4169 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
4172 <div class="doc_text">
4176 declare void %llvm.memcpy.i32(i8 * <dest>, i8 * <src>,
4177 i32 <len>, i32 <align>)
4178 declare void %llvm.memcpy.i64(i8 * <dest>, i8 * <src>,
4179 i64 <len>, i32 <align>)
4185 The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
4186 location to the destination location.
4190 Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
4191 intrinsics do not return a value, and takes an extra alignment argument.
4197 The first argument is a pointer to the destination, the second is a pointer to
4198 the source. The third argument is an integer argument
4199 specifying the number of bytes to copy, and the fourth argument is the alignment
4200 of the source and destination locations.
4204 If the call to this intrinisic has an alignment value that is not 0 or 1, then
4205 the caller guarantees that both the source and destination pointers are aligned
4212 The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
4213 location to the destination location, which are not allowed to overlap. It
4214 copies "len" bytes of memory over. If the argument is known to be aligned to
4215 some boundary, this can be specified as the fourth argument, otherwise it should
4221 <!-- _______________________________________________________________________ -->
4222 <div class="doc_subsubsection">
4223 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
4226 <div class="doc_text">
4230 declare void %llvm.memmove.i32(i8 * <dest>, i8 * <src>,
4231 i32 <len>, i32 <align>)
4232 declare void %llvm.memmove.i64(i8 * <dest>, i8 * <src>,
4233 i64 <len>, i32 <align>)
4239 The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
4240 location to the destination location. It is similar to the
4241 '<tt>llvm.memcmp</tt>' intrinsic but allows the two memory locations to overlap.
4245 Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
4246 intrinsics do not return a value, and takes an extra alignment argument.
4252 The first argument is a pointer to the destination, the second is a pointer to
4253 the source. The third argument is an integer argument
4254 specifying the number of bytes to copy, and the fourth argument is the alignment
4255 of the source and destination locations.
4259 If the call to this intrinisic has an alignment value that is not 0 or 1, then
4260 the caller guarantees that the source and destination pointers are aligned to
4267 The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
4268 location to the destination location, which may overlap. It
4269 copies "len" bytes of memory over. If the argument is known to be aligned to
4270 some boundary, this can be specified as the fourth argument, otherwise it should
4276 <!-- _______________________________________________________________________ -->
4277 <div class="doc_subsubsection">
4278 <a name="i_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
4281 <div class="doc_text">
4285 declare void %llvm.memset.i32(i8 * <dest>, i8 <val>,
4286 i32 <len>, i32 <align>)
4287 declare void %llvm.memset.i64(i8 * <dest>, i8 <val>,
4288 i64 <len>, i32 <align>)
4294 The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
4299 Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
4300 does not return a value, and takes an extra alignment argument.
4306 The first argument is a pointer to the destination to fill, the second is the
4307 byte value to fill it with, the third argument is an integer
4308 argument specifying the number of bytes to fill, and the fourth argument is the
4309 known alignment of destination location.
4313 If the call to this intrinisic has an alignment value that is not 0 or 1, then
4314 the caller guarantees that the destination pointer is aligned to that boundary.
4320 The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
4322 destination location. If the argument is known to be aligned to some boundary,
4323 this can be specified as the fourth argument, otherwise it should be set to 0 or
4329 <!-- _______________________________________________________________________ -->
4330 <div class="doc_subsubsection">
4331 <a name="i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
4334 <div class="doc_text">
4338 declare float %llvm.sqrt.f32(float %Val)
4339 declare double %llvm.sqrt.f64(double %Val)
4345 The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
4346 returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
4347 <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
4348 negative numbers (which allows for better optimization).
4354 The argument and return value are floating point numbers of the same type.
4360 This function returns the sqrt of the specified operand if it is a positive
4361 floating point number.
4365 <!-- _______________________________________________________________________ -->
4366 <div class="doc_subsubsection">
4367 <a name="i_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
4370 <div class="doc_text">
4374 declare float %llvm.powi.f32(float %Val, i32 %power)
4375 declare double %llvm.powi.f64(double %Val, i32 %power)
4381 The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
4382 specified (positive or negative) power. The order of evaluation of
4383 multiplications is not defined.
4389 The second argument is an integer power, and the first is a value to raise to
4396 This function returns the first value raised to the second power with an
4397 unspecified sequence of rounding operations.</p>
4401 <!-- ======================================================================= -->
4402 <div class="doc_subsection">
4403 <a name="int_manip">Bit Manipulation Intrinsics</a>
4406 <div class="doc_text">
4408 LLVM provides intrinsics for a few important bit manipulation operations.
4409 These allow efficient code generation for some algorithms.
4414 <!-- _______________________________________________________________________ -->
4415 <div class="doc_subsubsection">
4416 <a name="i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
4419 <div class="doc_text">
4423 declare i16 %llvm.bswap.i16(i16 <id>)
4424 declare i32 %llvm.bswap.i32(i32 <id>)
4425 declare i64 %llvm.bswap.i64(i64 <id>)
4431 The '<tt>llvm.bwsap</tt>' family of intrinsics is used to byteswap a 16, 32 or
4432 64 bit quantity. These are useful for performing operations on data that is not
4433 in the target's native byte order.
4439 The <tt>llvm.bswap.16</tt> intrinsic returns an i16 value that has the high
4440 and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
4441 intrinsic returns an i32 value that has the four bytes of the input i32
4442 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the returned
4443 i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i64</tt>
4444 intrinsic extends this concept to 64 bits.
4449 <!-- _______________________________________________________________________ -->
4450 <div class="doc_subsubsection">
4451 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
4454 <div class="doc_text">
4458 declare i8 %llvm.ctpop.i8 (i8 <src>)
4459 declare i16 %llvm.ctpop.i16(i16 <src>)
4460 declare i32 %llvm.ctpop.i32(i32 <src>)
4461 declare i64 %llvm.ctpop.i64(i64 <src>)
4467 The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
4474 The only argument is the value to be counted. The argument may be of any
4475 integer type. The return type must match the argument type.
4481 The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
4485 <!-- _______________________________________________________________________ -->
4486 <div class="doc_subsubsection">
4487 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
4490 <div class="doc_text">
4494 declare i8 %llvm.ctlz.i8 (i8 <src>)
4495 declare i16 %llvm.ctlz.i16(i16 <src>)
4496 declare i32 %llvm.ctlz.i32(i32 <src>)
4497 declare i64 %llvm.ctlz.i64(i64 <src>)
4503 The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
4504 leading zeros in a variable.
4510 The only argument is the value to be counted. The argument may be of any
4511 integer type. The return type must match the argument type.
4517 The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
4518 in a variable. If the src == 0 then the result is the size in bits of the type
4519 of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
4525 <!-- _______________________________________________________________________ -->
4526 <div class="doc_subsubsection">
4527 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
4530 <div class="doc_text">
4534 declare i8 %llvm.cttz.i8 (i8 <src>)
4535 declare i16 %llvm.cttz.i16(i16 <src>)
4536 declare i32 %llvm.cttz.i32(i32 <src>)
4537 declare i64 %llvm.cttz.i64(i64 <src>)
4543 The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
4550 The only argument is the value to be counted. The argument may be of any
4551 integer type. The return type must match the argument type.
4557 The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
4558 in a variable. If the src == 0 then the result is the size in bits of the type
4559 of src. For example, <tt>llvm.cttz(2) = 1</tt>.
4563 <!-- ======================================================================= -->
4564 <div class="doc_subsection">
4565 <a name="int_debugger">Debugger Intrinsics</a>
4568 <div class="doc_text">
4570 The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
4571 are described in the <a
4572 href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
4573 Debugging</a> document.
4578 <!-- *********************************************************************** -->
4581 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
4582 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
4583 <a href="http://validator.w3.org/check/referer"><img
4584 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
4586 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
4587 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
4588 Last modified: $Date$