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="#moduleasm">Module-Level Inline Assembly</a></li>
30 <li><a href="#typesystem">Type System</a>
32 <li><a href="#t_primitive">Primitive Types</a>
34 <li><a href="#t_classifications">Type Classifications</a></li>
37 <li><a href="#t_derived">Derived Types</a>
39 <li><a href="#t_array">Array Type</a></li>
40 <li><a href="#t_function">Function Type</a></li>
41 <li><a href="#t_pointer">Pointer Type</a></li>
42 <li><a href="#t_struct">Structure Type</a></li>
43 <li><a href="#t_packed">Packed Type</a></li>
44 <li><a href="#t_opaque">Opaque Type</a></li>
49 <li><a href="#constants">Constants</a>
51 <li><a href="#simpleconstants">Simple Constants</a>
52 <li><a href="#aggregateconstants">Aggregate Constants</a>
53 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
54 <li><a href="#undefvalues">Undefined Values</a>
55 <li><a href="#constantexprs">Constant Expressions</a>
58 <li><a href="#othervalues">Other Values</a>
60 <li><a href="#inlineasm">Inline Assembler Expressions</a>
63 <li><a href="#instref">Instruction Reference</a>
65 <li><a href="#terminators">Terminator Instructions</a>
67 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
68 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
69 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
70 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
71 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
72 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
75 <li><a href="#binaryops">Binary Operations</a>
77 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
78 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
79 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
80 <li><a href="#i_div">'<tt>div</tt>' Instruction</a></li>
81 <li><a href="#i_rem">'<tt>rem</tt>' Instruction</a></li>
82 <li><a href="#i_setcc">'<tt>set<i>cc</i></tt>' Instructions</a></li>
85 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
87 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
88 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
89 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
90 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
91 <li><a href="#i_shr">'<tt>shr</tt>' Instruction</a></li>
94 <li><a href="#memoryops">Memory Access Operations</a>
96 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
97 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
98 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
99 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
100 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
101 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
104 <li><a href="#otherops">Other Operations</a>
106 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
107 <li><a href="#i_cast">'<tt>cast .. to</tt>' Instruction</a></li>
108 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
109 <li><a href="#i_vset">'<tt>vset</tt>' Instruction</a></li>
110 <li><a href="#i_vselect">'<tt>vselect</tt>' Instruction</a></li>
111 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
112 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
113 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
114 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
119 <li><a href="#intrinsics">Intrinsic Functions</a>
121 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
123 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
124 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
125 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
128 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
130 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
131 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
132 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
135 <li><a href="#int_codegen">Code Generator Intrinsics</a>
137 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
138 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
139 <li><a href="#i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
140 <li><a href="#i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
141 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
142 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
143 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
146 <li><a href="#int_libc">Standard C Library Intrinsics</a>
148 <li><a href="#i_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
149 <li><a href="#i_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
150 <li><a href="#i_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
151 <li><a href="#i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a></li>
152 <li><a href="#i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
156 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
158 <li><a href="#i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
159 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
160 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
161 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
164 <li><a href="#int_debugger">Debugger intrinsics</a></li>
169 <div class="doc_author">
170 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
171 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
174 <!-- *********************************************************************** -->
175 <div class="doc_section"> <a name="abstract">Abstract </a></div>
176 <!-- *********************************************************************** -->
178 <div class="doc_text">
179 <p>This document is a reference manual for the LLVM assembly language.
180 LLVM is an SSA based representation that provides type safety,
181 low-level operations, flexibility, and the capability of representing
182 'all' high-level languages cleanly. It is the common code
183 representation used throughout all phases of the LLVM compilation
187 <!-- *********************************************************************** -->
188 <div class="doc_section"> <a name="introduction">Introduction</a> </div>
189 <!-- *********************************************************************** -->
191 <div class="doc_text">
193 <p>The LLVM code representation is designed to be used in three
194 different forms: as an in-memory compiler IR, as an on-disk bytecode
195 representation (suitable for fast loading by a Just-In-Time compiler),
196 and as a human readable assembly language representation. This allows
197 LLVM to provide a powerful intermediate representation for efficient
198 compiler transformations and analysis, while providing a natural means
199 to debug and visualize the transformations. The three different forms
200 of LLVM are all equivalent. This document describes the human readable
201 representation and notation.</p>
203 <p>The LLVM representation aims to be light-weight and low-level
204 while being expressive, typed, and extensible at the same time. It
205 aims to be a "universal IR" of sorts, by being at a low enough level
206 that high-level ideas may be cleanly mapped to it (similar to how
207 microprocessors are "universal IR's", allowing many source languages to
208 be mapped to them). By providing type information, LLVM can be used as
209 the target of optimizations: for example, through pointer analysis, it
210 can be proven that a C automatic variable is never accessed outside of
211 the current function... allowing it to be promoted to a simple SSA
212 value instead of a memory location.</p>
216 <!-- _______________________________________________________________________ -->
217 <div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
219 <div class="doc_text">
221 <p>It is important to note that this document describes 'well formed'
222 LLVM assembly language. There is a difference between what the parser
223 accepts and what is considered 'well formed'. For example, the
224 following instruction is syntactically okay, but not well formed:</p>
227 %x = <a href="#i_add">add</a> int 1, %x
230 <p>...because the definition of <tt>%x</tt> does not dominate all of
231 its uses. The LLVM infrastructure provides a verification pass that may
232 be used to verify that an LLVM module is well formed. This pass is
233 automatically run by the parser after parsing input assembly and by
234 the optimizer before it outputs bytecode. The violations pointed out
235 by the verifier pass indicate bugs in transformation passes or input to
238 <!-- Describe the typesetting conventions here. --> </div>
240 <!-- *********************************************************************** -->
241 <div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
242 <!-- *********************************************************************** -->
244 <div class="doc_text">
246 <p>LLVM uses three different forms of identifiers, for different
250 <li>Named values are represented as a string of characters with a '%' prefix.
251 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
252 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
253 Identifiers which require other characters in their names can be surrounded
254 with quotes. In this way, anything except a <tt>"</tt> character can be used
257 <li>Unnamed values are represented as an unsigned numeric value with a '%'
258 prefix. For example, %12, %2, %44.</li>
260 <li>Constants, which are described in a <a href="#constants">section about
261 constants</a>, below.</li>
264 <p>LLVM requires that values start with a '%' sign for two reasons: Compilers
265 don't need to worry about name clashes with reserved words, and the set of
266 reserved words may be expanded in the future without penalty. Additionally,
267 unnamed identifiers allow a compiler to quickly come up with a temporary
268 variable without having to avoid symbol table conflicts.</p>
270 <p>Reserved words in LLVM are very similar to reserved words in other
271 languages. There are keywords for different opcodes ('<tt><a
272 href="#i_add">add</a></tt>', '<tt><a href="#i_cast">cast</a></tt>', '<tt><a
273 href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
274 href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
275 and others. These reserved words cannot conflict with variable names, because
276 none of them start with a '%' character.</p>
278 <p>Here is an example of LLVM code to multiply the integer variable
279 '<tt>%X</tt>' by 8:</p>
284 %result = <a href="#i_mul">mul</a> uint %X, 8
287 <p>After strength reduction:</p>
290 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
293 <p>And the hard way:</p>
296 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
297 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
298 %result = <a href="#i_add">add</a> uint %1, %1
301 <p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
302 important lexical features of LLVM:</p>
306 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
309 <li>Unnamed temporaries are created when the result of a computation is not
310 assigned to a named value.</li>
312 <li>Unnamed temporaries are numbered sequentially</li>
316 <p>...and it also shows a convention that we follow in this document. When
317 demonstrating instructions, we will follow an instruction with a comment that
318 defines the type and name of value produced. Comments are shown in italic
323 <!-- *********************************************************************** -->
324 <div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
325 <!-- *********************************************************************** -->
327 <!-- ======================================================================= -->
328 <div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
331 <div class="doc_text">
333 <p>LLVM programs are composed of "Module"s, each of which is a
334 translation unit of the input programs. Each module consists of
335 functions, global variables, and symbol table entries. Modules may be
336 combined together with the LLVM linker, which merges function (and
337 global variable) definitions, resolves forward declarations, and merges
338 symbol table entries. Here is an example of the "hello world" module:</p>
340 <pre><i>; Declare the string constant as a global constant...</i>
341 <a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
342 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
344 <i>; External declaration of the puts function</i>
345 <a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
347 <i>; Definition of main function</i>
348 int %main() { <i>; int()* </i>
349 <i>; Convert [13x sbyte]* to sbyte *...</i>
351 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
353 <i>; Call puts function to write out the string to stdout...</i>
355 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
357 href="#i_ret">ret</a> int 0<br>}<br></pre>
359 <p>This example is made up of a <a href="#globalvars">global variable</a>
360 named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
361 function, and a <a href="#functionstructure">function definition</a>
362 for "<tt>main</tt>".</p>
364 <p>In general, a module is made up of a list of global values,
365 where both functions and global variables are global values. Global values are
366 represented by a pointer to a memory location (in this case, a pointer to an
367 array of char, and a pointer to a function), and have one of the following <a
368 href="#linkage">linkage types</a>.</p>
372 <!-- ======================================================================= -->
373 <div class="doc_subsection">
374 <a name="linkage">Linkage Types</a>
377 <div class="doc_text">
380 All Global Variables and Functions have one of the following types of linkage:
385 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
387 <dd>Global values with internal linkage are only directly accessible by
388 objects in the current module. In particular, linking code into a module with
389 an internal global value may cause the internal to be renamed as necessary to
390 avoid collisions. Because the symbol is internal to the module, all
391 references can be updated. This corresponds to the notion of the
392 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
395 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
397 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
398 the twist that linking together two modules defining the same
399 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
400 is typically used to implement inline functions. Unreferenced
401 <tt>linkonce</tt> globals are allowed to be discarded.
404 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
406 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
407 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
408 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.
411 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
413 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
414 pointer to array type. When two global variables with appending linkage are
415 linked together, the two global arrays are appended together. This is the
416 LLVM, typesafe, equivalent of having the system linker append together
417 "sections" with identical names when .o files are linked.
420 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
422 <dd>If none of the above identifiers are used, the global is externally
423 visible, meaning that it participates in linkage and can be used to resolve
424 external symbol references.
428 <p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
429 variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
430 variable and was linked with this one, one of the two would be renamed,
431 preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
432 external (i.e., lacking any linkage declarations), they are accessible
433 outside of the current module. It is illegal for a function <i>declaration</i>
434 to have any linkage type other than "externally visible".</a></p>
438 <!-- ======================================================================= -->
439 <div class="doc_subsection">
440 <a name="callingconv">Calling Conventions</a>
443 <div class="doc_text">
445 <p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
446 and <a href="#i_invoke">invokes</a> can all have an optional calling convention
447 specified for the call. The calling convention of any pair of dynamic
448 caller/callee must match, or the behavior of the program is undefined. The
449 following calling conventions are supported by LLVM, and more may be added in
453 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
455 <dd>This calling convention (the default if no other calling convention is
456 specified) matches the target C calling conventions. This calling convention
457 supports varargs function calls and tolerates some mismatch in the declared
458 prototype and implemented declaration of the function (as does normal C).
461 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
463 <dd>This calling convention attempts to make calls as fast as possible
464 (e.g. by passing things in registers). This calling convention allows the
465 target to use whatever tricks it wants to produce fast code for the target,
466 without having to conform to an externally specified ABI. Implementations of
467 this convention should allow arbitrary tail call optimization to be supported.
468 This calling convention does not support varargs and requires the prototype of
469 all callees to exactly match the prototype of the function definition.
472 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
474 <dd>This calling convention attempts to make code in the caller as efficient
475 as possible under the assumption that the call is not commonly executed. As
476 such, these calls often preserve all registers so that the call does not break
477 any live ranges in the caller side. This calling convention does not support
478 varargs and requires the prototype of all callees to exactly match the
479 prototype of the function definition.
482 <dt><b>"<tt>cc <<em>n</em>></tt>" - Numbered convention</b>:</dt>
484 <dd>Any calling convention may be specified by number, allowing
485 target-specific calling conventions to be used. Target specific calling
486 conventions start at 64.
490 <p>More calling conventions can be added/defined on an as-needed basis, to
491 support pascal conventions or any other well-known target-independent
496 <!-- ======================================================================= -->
497 <div class="doc_subsection">
498 <a name="globalvars">Global Variables</a>
501 <div class="doc_text">
503 <p>Global variables define regions of memory allocated at compilation time
504 instead of run-time. Global variables may optionally be initialized, may have
505 an explicit section to be placed in, and may
506 have an optional explicit alignment specified. A
507 variable may be defined as a global "constant," which indicates that the
508 contents of the variable will <b>never</b> be modified (enabling better
509 optimization, allowing the global data to be placed in the read-only section of
510 an executable, etc). Note that variables that need runtime initialization
511 cannot be marked "constant" as there is a store to the variable.</p>
514 LLVM explicitly allows <em>declarations</em> of global variables to be marked
515 constant, even if the final definition of the global is not. This capability
516 can be used to enable slightly better optimization of the program, but requires
517 the language definition to guarantee that optimizations based on the
518 'constantness' are valid for the translation units that do not include the
522 <p>As SSA values, global variables define pointer values that are in
523 scope (i.e. they dominate) all basic blocks in the program. Global
524 variables always define a pointer to their "content" type because they
525 describe a region of memory, and all memory objects in LLVM are
526 accessed through pointers.</p>
528 <p>LLVM allows an explicit section to be specified for globals. If the target
529 supports it, it will emit globals to the section specified.</p>
531 <p>An explicit alignment may be specified for a global. If not present, or if
532 the alignment is set to zero, the alignment of the global is set by the target
533 to whatever it feels convenient. If an explicit alignment is specified, the
534 global is forced to have at least that much alignment. All alignments must be
540 <!-- ======================================================================= -->
541 <div class="doc_subsection">
542 <a name="functionstructure">Functions</a>
545 <div class="doc_text">
547 <p>LLVM function definitions consist of an optional <a href="#linkage">linkage
548 type</a>, an optional <a href="#callingconv">calling convention</a>, a return
549 type, a function name, a (possibly empty) argument list, an optional section,
550 an optional alignment, an opening curly brace,
551 a list of basic blocks, and a closing curly brace. LLVM function declarations
552 are defined with the "<tt>declare</tt>" keyword, an optional <a
553 href="#callingconv">calling convention</a>, a return type, a function name,
554 a possibly empty list of arguments, and an optional alignment.</p>
556 <p>A function definition contains a list of basic blocks, forming the CFG for
557 the function. Each basic block may optionally start with a label (giving the
558 basic block a symbol table entry), contains a list of instructions, and ends
559 with a <a href="#terminators">terminator</a> instruction (such as a branch or
560 function return).</p>
562 <p>The first basic block in a program is special in two ways: it is immediately
563 executed on entrance to the function, and it is not allowed to have predecessor
564 basic blocks (i.e. there can not be any branches to the entry block of a
565 function). Because the block can have no predecessors, it also cannot have any
566 <a href="#i_phi">PHI nodes</a>.</p>
568 <p>LLVM functions are identified by their name and type signature. Hence, two
569 functions with the same name but different parameter lists or return values are
570 considered different functions, and LLVM will resolve references to each
573 <p>LLVM allows an explicit section to be specified for functions. If the target
574 supports it, it will emit functions to the section specified.</p>
576 <p>An explicit alignment may be specified for a function. If not present, or if
577 the alignment is set to zero, the alignment of the function is set by the target
578 to whatever it feels convenient. If an explicit alignment is specified, the
579 function is forced to have at least that much alignment. All alignments must be
584 <!-- ======================================================================= -->
585 <div class="doc_subsection">
586 <a name="moduleasm">Module-Level Inline Assembly</a></li>
589 <div class="doc_text">
591 Modules may contain "module-level inline asm" blocks, which corresponds to the
592 GCC "file scope inline asm" blocks. These blocks are internally concatenated by
593 LLVM and treated as a single unit, but may be separated in the .ll file if
594 desired. The syntax is very simple:
597 <div class="doc_code"><pre>
598 module asm "inline asm code goes here"
599 module asm "more can go here"
602 <p>The strings can contain any character by escaping non-printable characters.
603 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
608 The inline asm code is simply printed to the machine code .s file when
609 assembly code is generated.
614 <!-- *********************************************************************** -->
615 <div class="doc_section"> <a name="typesystem">Type System</a> </div>
616 <!-- *********************************************************************** -->
618 <div class="doc_text">
620 <p>The LLVM type system is one of the most important features of the
621 intermediate representation. Being typed enables a number of
622 optimizations to be performed on the IR directly, without having to do
623 extra analyses on the side before the transformation. A strong type
624 system makes it easier to read the generated code and enables novel
625 analyses and transformations that are not feasible to perform on normal
626 three address code representations.</p>
630 <!-- ======================================================================= -->
631 <div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
632 <div class="doc_text">
633 <p>The primitive types are the fundamental building blocks of the LLVM
634 system. The current set of primitive types is as follows:</p>
636 <table class="layout">
641 <tr><th>Type</th><th>Description</th></tr>
642 <tr><td><tt>void</tt></td><td>No value</td></tr>
643 <tr><td><tt>ubyte</tt></td><td>Unsigned 8-bit value</td></tr>
644 <tr><td><tt>ushort</tt></td><td>Unsigned 16-bit value</td></tr>
645 <tr><td><tt>uint</tt></td><td>Unsigned 32-bit value</td></tr>
646 <tr><td><tt>ulong</tt></td><td>Unsigned 64-bit value</td></tr>
647 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
648 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
655 <tr><th>Type</th><th>Description</th></tr>
656 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
657 <tr><td><tt>sbyte</tt></td><td>Signed 8-bit value</td></tr>
658 <tr><td><tt>short</tt></td><td>Signed 16-bit value</td></tr>
659 <tr><td><tt>int</tt></td><td>Signed 32-bit value</td></tr>
660 <tr><td><tt>long</tt></td><td>Signed 64-bit value</td></tr>
661 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
669 <!-- _______________________________________________________________________ -->
670 <div class="doc_subsubsection"> <a name="t_classifications">Type
671 Classifications</a> </div>
672 <div class="doc_text">
673 <p>These different primitive types fall into a few useful
676 <table border="1" cellspacing="0" cellpadding="4">
678 <tr><th>Classification</th><th>Types</th></tr>
680 <td><a name="t_signed">signed</a></td>
681 <td><tt>sbyte, short, int, long, float, double</tt></td>
684 <td><a name="t_unsigned">unsigned</a></td>
685 <td><tt>ubyte, ushort, uint, ulong</tt></td>
688 <td><a name="t_integer">integer</a></td>
689 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
692 <td><a name="t_integral">integral</a></td>
693 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
697 <td><a name="t_floating">floating point</a></td>
698 <td><tt>float, double</tt></td>
701 <td><a name="t_firstclass">first class</a></td>
702 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
703 float, double, <a href="#t_pointer">pointer</a>,
704 <a href="#t_packed">packed</a></tt></td>
709 <p>The <a href="#t_firstclass">first class</a> types are perhaps the
710 most important. Values of these types are the only ones which can be
711 produced by instructions, passed as arguments, or used as operands to
712 instructions. This means that all structures and arrays must be
713 manipulated either by pointer or by component.</p>
716 <!-- ======================================================================= -->
717 <div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
719 <div class="doc_text">
721 <p>The real power in LLVM comes from the derived types in the system.
722 This is what allows a programmer to represent arrays, functions,
723 pointers, and other useful types. Note that these derived types may be
724 recursive: For example, it is possible to have a two dimensional array.</p>
728 <!-- _______________________________________________________________________ -->
729 <div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
731 <div class="doc_text">
735 <p>The array type is a very simple derived type that arranges elements
736 sequentially in memory. The array type requires a size (number of
737 elements) and an underlying data type.</p>
742 [<# elements> x <elementtype>]
745 <p>The number of elements is a constant integer value; elementtype may
746 be any type with a size.</p>
749 <table class="layout">
752 <tt>[40 x int ]</tt><br/>
753 <tt>[41 x int ]</tt><br/>
754 <tt>[40 x uint]</tt><br/>
757 Array of 40 integer values.<br/>
758 Array of 41 integer values.<br/>
759 Array of 40 unsigned integer values.<br/>
763 <p>Here are some examples of multidimensional arrays:</p>
764 <table class="layout">
767 <tt>[3 x [4 x int]]</tt><br/>
768 <tt>[12 x [10 x float]]</tt><br/>
769 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
772 3x4 array of integer values.<br/>
773 12x10 array of single precision floating point values.<br/>
774 2x3x4 array of unsigned integer values.<br/>
779 <p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
780 length array. Normally, accesses past the end of an array are undefined in
781 LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
782 As a special case, however, zero length arrays are recognized to be variable
783 length. This allows implementation of 'pascal style arrays' with the LLVM
784 type "{ int, [0 x float]}", for example.</p>
788 <!-- _______________________________________________________________________ -->
789 <div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
790 <div class="doc_text">
792 <p>The function type can be thought of as a function signature. It
793 consists of a return type and a list of formal parameter types.
794 Function types are usually used to build virtual function tables
795 (which are structures of pointers to functions), for indirect function
796 calls, and when defining a function.</p>
798 The return type of a function type cannot be an aggregate type.
801 <pre> <returntype> (<parameter list>)<br></pre>
802 <p>...where '<tt><parameter list></tt>' is a comma-separated list of type
803 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
804 which indicates that the function takes a variable number of arguments.
805 Variable argument functions can access their arguments with the <a
806 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
808 <table class="layout">
811 <tt>int (int)</tt> <br/>
812 <tt>float (int, int *) *</tt><br/>
813 <tt>int (sbyte *, ...)</tt><br/>
816 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
817 <a href="#t_pointer">Pointer</a> to a function that takes an
818 <tt>int</tt> and a <a href="#t_pointer">pointer</a> to <tt>int</tt>,
819 returning <tt>float</tt>.<br/>
820 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
821 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
822 the signature for <tt>printf</tt> in LLVM.<br/>
828 <!-- _______________________________________________________________________ -->
829 <div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
830 <div class="doc_text">
832 <p>The structure type is used to represent a collection of data members
833 together in memory. The packing of the field types is defined to match
834 the ABI of the underlying processor. The elements of a structure may
835 be any type that has a size.</p>
836 <p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
837 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
838 field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
841 <pre> { <type list> }<br></pre>
843 <table class="layout">
846 <tt>{ int, int, int }</tt><br/>
847 <tt>{ float, int (int) * }</tt><br/>
850 a triple of three <tt>int</tt> values<br/>
851 A pair, where the first element is a <tt>float</tt> and the second element
852 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
853 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
859 <!-- _______________________________________________________________________ -->
860 <div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
861 <div class="doc_text">
863 <p>As in many languages, the pointer type represents a pointer or
864 reference to another object, which must live in memory.</p>
866 <pre> <type> *<br></pre>
868 <table class="layout">
871 <tt>[4x int]*</tt><br/>
872 <tt>int (int *) *</tt><br/>
875 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
876 four <tt>int</tt> values<br/>
877 A <a href="#t_pointer">pointer</a> to a <a
878 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
885 <!-- _______________________________________________________________________ -->
886 <div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
887 <div class="doc_text">
891 <p>A packed type is a simple derived type that represents a vector
892 of elements. Packed types are used when multiple primitive data
893 are operated in parallel using a single instruction (SIMD).
894 A packed type requires a size (number of
895 elements) and an underlying primitive data type. Vectors must have a power
896 of two length (1, 2, 4, 8, 16 ...). Packed types are
897 considered <a href="#t_firstclass">first class</a>.</p>
902 < <# elements> x <elementtype> >
905 <p>The number of elements is a constant integer value; elementtype may
906 be any integral or floating point type.</p>
910 <table class="layout">
913 <tt><4 x int></tt><br/>
914 <tt><8 x float></tt><br/>
915 <tt><2 x uint></tt><br/>
918 Packed vector of 4 integer values.<br/>
919 Packed vector of 8 floating-point values.<br/>
920 Packed vector of 2 unsigned integer values.<br/>
926 <!-- _______________________________________________________________________ -->
927 <div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
928 <div class="doc_text">
932 <p>Opaque types are used to represent unknown types in the system. This
933 corresponds (for example) to the C notion of a foward declared structure type.
934 In LLVM, opaque types can eventually be resolved to any type (not just a
945 <table class="layout">
958 <!-- *********************************************************************** -->
959 <div class="doc_section"> <a name="constants">Constants</a> </div>
960 <!-- *********************************************************************** -->
962 <div class="doc_text">
964 <p>LLVM has several different basic types of constants. This section describes
965 them all and their syntax.</p>
969 <!-- ======================================================================= -->
970 <div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
972 <div class="doc_text">
975 <dt><b>Boolean constants</b></dt>
977 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
978 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
981 <dt><b>Integer constants</b></dt>
983 <dd>Standard integers (such as '4') are constants of the <a
984 href="#t_integer">integer</a> type. Negative numbers may be used with signed
988 <dt><b>Floating point constants</b></dt>
990 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
991 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
992 notation (see below). Floating point constants must have a <a
993 href="#t_floating">floating point</a> type. </dd>
995 <dt><b>Null pointer constants</b></dt>
997 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
998 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1002 <p>The one non-intuitive notation for constants is the optional hexadecimal form
1003 of floating point constants. For example, the form '<tt>double
1004 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
1005 4.5e+15</tt>'. The only time hexadecimal floating point constants are required
1006 (and the only time that they are generated by the disassembler) is when a
1007 floating point constant must be emitted but it cannot be represented as a
1008 decimal floating point number. For example, NaN's, infinities, and other
1009 special values are represented in their IEEE hexadecimal format so that
1010 assembly and disassembly do not cause any bits to change in the constants.</p>
1014 <!-- ======================================================================= -->
1015 <div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1018 <div class="doc_text">
1019 <p>Aggregate constants arise from aggregation of simple constants
1020 and smaller aggregate constants.</p>
1023 <dt><b>Structure constants</b></dt>
1025 <dd>Structure constants are represented with notation similar to structure
1026 type definitions (a comma separated list of elements, surrounded by braces
1027 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0, int* %G }</tt>",
1028 where "<tt>%G</tt>" is declared as "<tt>%G = external global int</tt>". Structure constants
1029 must have <a href="#t_struct">structure type</a>, and the number and
1030 types of elements must match those specified by the type.
1033 <dt><b>Array constants</b></dt>
1035 <dd>Array constants are represented with notation similar to array type
1036 definitions (a comma separated list of elements, surrounded by square brackets
1037 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
1038 constants must have <a href="#t_array">array type</a>, and the number and
1039 types of elements must match those specified by the type.
1042 <dt><b>Packed constants</b></dt>
1044 <dd>Packed constants are represented with notation similar to packed type
1045 definitions (a comma separated list of elements, surrounded by
1046 less-than/greater-than's (<tt><></tt>)). For example: "<tt>< int 42,
1047 int 11, int 74, int 100 ></tt>". Packed constants must have <a
1048 href="#t_packed">packed type</a>, and the number and types of elements must
1049 match those specified by the type.
1052 <dt><b>Zero initialization</b></dt>
1054 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1055 value to zero of <em>any</em> type, including scalar and aggregate types.
1056 This is often used to avoid having to print large zero initializers (e.g. for
1057 large arrays) and is always exactly equivalent to using explicit zero
1064 <!-- ======================================================================= -->
1065 <div class="doc_subsection">
1066 <a name="globalconstants">Global Variable and Function Addresses</a>
1069 <div class="doc_text">
1071 <p>The addresses of <a href="#globalvars">global variables</a> and <a
1072 href="#functionstructure">functions</a> are always implicitly valid (link-time)
1073 constants. These constants are explicitly referenced when the <a
1074 href="#identifiers">identifier for the global</a> is used and always have <a
1075 href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1081 %Z = global [2 x int*] [ int* %X, int* %Y ]
1086 <!-- ======================================================================= -->
1087 <div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
1088 <div class="doc_text">
1089 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
1090 no specific value. Undefined values may be of any type and be used anywhere
1091 a constant is permitted.</p>
1093 <p>Undefined values indicate to the compiler that the program is well defined
1094 no matter what value is used, giving the compiler more freedom to optimize.
1098 <!-- ======================================================================= -->
1099 <div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1102 <div class="doc_text">
1104 <p>Constant expressions are used to allow expressions involving other constants
1105 to be used as constants. Constant expressions may be of any <a
1106 href="#t_firstclass">first class</a> type and may involve any LLVM operation
1107 that does not have side effects (e.g. load and call are not supported). The
1108 following is the syntax for constant expressions:</p>
1111 <dt><b><tt>cast ( CST to TYPE )</tt></b></dt>
1113 <dd>Cast a constant to another type.</dd>
1115 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1117 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1118 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1119 instruction, the index list may have zero or more indexes, which are required
1120 to make sense for the type of "CSTPTR".</dd>
1122 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1124 <dd>Perform the <a href="#i_select">select operation</a> on
1127 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1129 <dd>Perform the <a href="#i_extractelement">extractelement
1130 operation</a> on constants.
1132 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1134 <dd>Perform the <a href="#i_insertelement">insertelement
1135 operation</a> on constants.
1137 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1139 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1140 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
1141 binary</a> operations. The constraints on operands are the same as those for
1142 the corresponding instruction (e.g. no bitwise operations on floating point
1143 values are allowed).</dd>
1147 <!-- *********************************************************************** -->
1148 <div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1149 <!-- *********************************************************************** -->
1151 <!-- ======================================================================= -->
1152 <div class="doc_subsection">
1153 <a name="inlineasm">Inline Assembler Expressions</a>
1156 <div class="doc_text">
1159 LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1160 Module-Level Inline Assembly</a>) through the use of a special value. This
1161 value represents the inline assembler as a string (containing the instructions
1162 to emit), a list of operand constraints (stored as a string), and a flag that
1163 indicates whether or not the inline asm expression has side effects. An example
1164 inline assembler expression is:
1168 int(int) asm "bswap $0", "=r,r"
1172 Inline assembler expressions may <b>only</b> be used as the callee operand of
1173 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1177 %X = call int asm "<a href="#i_bswap">bswap</a> $0", "=r,r"(int %Y)
1181 Inline asms with side effects not visible in the constraint list must be marked
1182 as having side effects. This is done through the use of the
1183 '<tt>sideeffect</tt>' keyword, like so:
1187 call void asm sideeffect "eieio", ""()
1190 <p>TODO: The format of the asm and constraints string still need to be
1191 documented here. Constraints on what can be done (e.g. duplication, moving, etc
1192 need to be documented).
1197 <!-- *********************************************************************** -->
1198 <div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1199 <!-- *********************************************************************** -->
1201 <div class="doc_text">
1203 <p>The LLVM instruction set consists of several different
1204 classifications of instructions: <a href="#terminators">terminator
1205 instructions</a>, <a href="#binaryops">binary instructions</a>,
1206 <a href="#bitwiseops">bitwise binary instructions</a>, <a
1207 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1208 instructions</a>.</p>
1212 <!-- ======================================================================= -->
1213 <div class="doc_subsection"> <a name="terminators">Terminator
1214 Instructions</a> </div>
1216 <div class="doc_text">
1218 <p>As mentioned <a href="#functionstructure">previously</a>, every
1219 basic block in a program ends with a "Terminator" instruction, which
1220 indicates which block should be executed after the current block is
1221 finished. These terminator instructions typically yield a '<tt>void</tt>'
1222 value: they produce control flow, not values (the one exception being
1223 the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
1224 <p>There are six different terminator instructions: the '<a
1225 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1226 instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
1227 the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1228 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1229 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
1233 <!-- _______________________________________________________________________ -->
1234 <div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1235 Instruction</a> </div>
1236 <div class="doc_text">
1238 <pre> ret <type> <value> <i>; Return a value from a non-void function</i>
1239 ret void <i>; Return from void function</i>
1242 <p>The '<tt>ret</tt>' instruction is used to return control flow (and a
1243 value) from a function back to the caller.</p>
1244 <p>There are two forms of the '<tt>ret</tt>' instruction: one that
1245 returns a value and then causes control flow, and one that just causes
1246 control flow to occur.</p>
1248 <p>The '<tt>ret</tt>' instruction may return any '<a
1249 href="#t_firstclass">first class</a>' type. Notice that a function is
1250 not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1251 instruction inside of the function that returns a value that does not
1252 match the return type of the function.</p>
1254 <p>When the '<tt>ret</tt>' instruction is executed, control flow
1255 returns back to the calling function's context. If the caller is a "<a
1256 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
1257 the instruction after the call. If the caller was an "<a
1258 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1259 at the beginning of the "normal" destination block. If the instruction
1260 returns a value, that value shall set the call or invoke instruction's
1263 <pre> ret int 5 <i>; Return an integer value of 5</i>
1264 ret void <i>; Return from a void function</i>
1267 <!-- _______________________________________________________________________ -->
1268 <div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
1269 <div class="doc_text">
1271 <pre> br bool <cond>, label <iftrue>, label <iffalse><br> br label <dest> <i>; Unconditional branch</i>
1274 <p>The '<tt>br</tt>' instruction is used to cause control flow to
1275 transfer to a different basic block in the current function. There are
1276 two forms of this instruction, corresponding to a conditional branch
1277 and an unconditional branch.</p>
1279 <p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1280 single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1281 unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1282 value as a target.</p>
1284 <p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1285 argument is evaluated. If the value is <tt>true</tt>, control flows
1286 to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1287 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
1289 <pre>Test:<br> %cond = <a href="#i_setcc">seteq</a> int %a, %b<br> br bool %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
1290 href="#i_ret">ret</a> int 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> int 0<br></pre>
1292 <!-- _______________________________________________________________________ -->
1293 <div class="doc_subsubsection">
1294 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1297 <div class="doc_text">
1301 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
1306 <p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1307 several different places. It is a generalization of the '<tt>br</tt>'
1308 instruction, allowing a branch to occur to one of many possible
1314 <p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1315 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1316 an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1317 table is not allowed to contain duplicate constant entries.</p>
1321 <p>The <tt>switch</tt> instruction specifies a table of values and
1322 destinations. When the '<tt>switch</tt>' instruction is executed, this
1323 table is searched for the given value. If the value is found, control flow is
1324 transfered to the corresponding destination; otherwise, control flow is
1325 transfered to the default destination.</p>
1327 <h5>Implementation:</h5>
1329 <p>Depending on properties of the target machine and the particular
1330 <tt>switch</tt> instruction, this instruction may be code generated in different
1331 ways. For example, it could be generated as a series of chained conditional
1332 branches or with a lookup table.</p>
1337 <i>; Emulate a conditional br instruction</i>
1338 %Val = <a href="#i_cast">cast</a> bool %value to int
1339 switch int %Val, label %truedest [int 0, label %falsedest ]
1341 <i>; Emulate an unconditional br instruction</i>
1342 switch uint 0, label %dest [ ]
1344 <i>; Implement a jump table:</i>
1345 switch uint %val, label %otherwise [ uint 0, label %onzero
1346 uint 1, label %onone
1347 uint 2, label %ontwo ]
1351 <!-- _______________________________________________________________________ -->
1352 <div class="doc_subsubsection">
1353 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1356 <div class="doc_text">
1361 <result> = invoke [<a href="#callingconv">cconv</a>] <ptr to function ty> %<function ptr val>(<function args>)
1362 to label <normal label> except label <exception label>
1367 <p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1368 function, with the possibility of control flow transfer to either the
1369 '<tt>normal</tt>' label or the
1370 '<tt>exception</tt>' label. If the callee function returns with the
1371 "<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1372 "normal" label. If the callee (or any indirect callees) returns with the "<a
1373 href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1374 continued at the dynamically nearest "exception" label.</p>
1378 <p>This instruction requires several arguments:</p>
1382 The optional "cconv" marker indicates which <a href="callingconv">calling
1383 convention</a> the call should use. If none is specified, the call defaults
1384 to using C calling conventions.
1386 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1387 function value being invoked. In most cases, this is a direct function
1388 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1389 an arbitrary pointer to function value.
1392 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1393 function to be invoked. </li>
1395 <li>'<tt>function args</tt>': argument list whose types match the function
1396 signature argument types. If the function signature indicates the function
1397 accepts a variable number of arguments, the extra arguments can be
1400 <li>'<tt>normal label</tt>': the label reached when the called function
1401 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1403 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1404 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1410 <p>This instruction is designed to operate as a standard '<tt><a
1411 href="#i_call">call</a></tt>' instruction in most regards. The primary
1412 difference is that it establishes an association with a label, which is used by
1413 the runtime library to unwind the stack.</p>
1415 <p>This instruction is used in languages with destructors to ensure that proper
1416 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1417 exception. Additionally, this is important for implementation of
1418 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
1422 %retval = invoke int %Test(int 15) to label %Continue
1423 except label %TestCleanup <i>; {int}:retval set</i>
1424 %retval = invoke <a href="#callingconv">coldcc</a> int %Test(int 15) to label %Continue
1425 except label %TestCleanup <i>; {int}:retval set</i>
1430 <!-- _______________________________________________________________________ -->
1432 <div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1433 Instruction</a> </div>
1435 <div class="doc_text">
1444 <p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1445 at the first callee in the dynamic call stack which used an <a
1446 href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1447 primarily used to implement exception handling.</p>
1451 <p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1452 immediately halt. The dynamic call stack is then searched for the first <a
1453 href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1454 execution continues at the "exceptional" destination block specified by the
1455 <tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1456 dynamic call chain, undefined behavior results.</p>
1459 <!-- _______________________________________________________________________ -->
1461 <div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1462 Instruction</a> </div>
1464 <div class="doc_text">
1473 <p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1474 instruction is used to inform the optimizer that a particular portion of the
1475 code is not reachable. This can be used to indicate that the code after a
1476 no-return function cannot be reached, and other facts.</p>
1480 <p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1485 <!-- ======================================================================= -->
1486 <div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
1487 <div class="doc_text">
1488 <p>Binary operators are used to do most of the computation in a
1489 program. They require two operands, execute an operation on them, and
1490 produce a single value. The operands might represent
1491 multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1492 The result value of a binary operator is not
1493 necessarily the same type as its operands.</p>
1494 <p>There are several different binary operators:</p>
1496 <!-- _______________________________________________________________________ -->
1497 <div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1498 Instruction</a> </div>
1499 <div class="doc_text">
1501 <pre> <result> = add <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1504 <p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
1506 <p>The two arguments to the '<tt>add</tt>' instruction must be either <a
1507 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1508 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1509 Both arguments must have identical types.</p>
1511 <p>The value produced is the integer or floating point sum of the two
1514 <pre> <result> = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
1517 <!-- _______________________________________________________________________ -->
1518 <div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1519 Instruction</a> </div>
1520 <div class="doc_text">
1522 <pre> <result> = sub <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1525 <p>The '<tt>sub</tt>' instruction returns the difference of its two
1527 <p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1528 instruction present in most other intermediate representations.</p>
1530 <p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
1531 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
1533 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1534 Both arguments must have identical types.</p>
1536 <p>The value produced is the integer or floating point difference of
1537 the two operands.</p>
1539 <pre> <result> = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
1540 <result> = sub int 0, %val <i>; yields {int}:result = -%var</i>
1543 <!-- _______________________________________________________________________ -->
1544 <div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1545 Instruction</a> </div>
1546 <div class="doc_text">
1548 <pre> <result> = mul <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1551 <p>The '<tt>mul</tt>' instruction returns the product of its two
1554 <p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
1555 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
1557 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1558 Both arguments must have identical types.</p>
1560 <p>The value produced is the integer or floating point product of the
1562 <p>There is no signed vs unsigned multiplication. The appropriate
1563 action is taken based on the type of the operand.</p>
1565 <pre> <result> = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
1568 <!-- _______________________________________________________________________ -->
1569 <div class="doc_subsubsection"> <a name="i_div">'<tt>div</tt>'
1570 Instruction</a> </div>
1571 <div class="doc_text">
1573 <pre> <result> = div <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1576 <p>The '<tt>div</tt>' instruction returns the quotient of its two
1579 <p>The two arguments to the '<tt>div</tt>' instruction must be either <a
1580 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
1582 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1583 Both arguments must have identical types.</p>
1585 <p>The value produced is the integer or floating point quotient of the
1588 <pre> <result> = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
1591 <!-- _______________________________________________________________________ -->
1592 <div class="doc_subsubsection"> <a name="i_rem">'<tt>rem</tt>'
1593 Instruction</a> </div>
1594 <div class="doc_text">
1596 <pre> <result> = rem <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1599 <p>The '<tt>rem</tt>' instruction returns the remainder from the
1600 division of its two operands.</p>
1602 <p>The two arguments to the '<tt>rem</tt>' instruction must be either <a
1603 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
1605 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1606 Both arguments must have identical types.</p>
1608 <p>This returns the <i>remainder</i> of a division (where the result
1609 has the same sign as the divisor), not the <i>modulus</i> (where the
1610 result has the same sign as the dividend) of a value. For more
1611 information about the difference, see <a
1612 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1615 <pre> <result> = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1619 <!-- _______________________________________________________________________ -->
1620 <div class="doc_subsubsection"> <a name="i_setcc">'<tt>set<i>cc</i></tt>'
1621 Instructions</a> </div>
1622 <div class="doc_text">
1624 <pre> <result> = seteq <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1625 <result> = setne <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1626 <result> = setlt <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1627 <result> = setgt <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1628 <result> = setle <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1629 <result> = setge <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1632 <p>The '<tt>set<i>cc</i></tt>' family of instructions returns a boolean
1633 value based on a comparison of their two operands.</p>
1635 <p>The two arguments to the '<tt>set<i>cc</i></tt>' instructions must
1636 be of <a href="#t_firstclass">first class</a> type (it is not possible
1637 to compare '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>'
1638 or '<tt>void</tt>' values, etc...). Both arguments must have identical
1641 <p>The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1642 value if both operands are equal.<br>
1643 The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1644 value if both operands are unequal.<br>
1645 The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1646 value if the first operand is less than the second operand.<br>
1647 The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1648 value if the first operand is greater than the second operand.<br>
1649 The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1650 value if the first operand is less than or equal to the second operand.<br>
1651 The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1652 value if the first operand is greater than or equal to the second
1655 <pre> <result> = seteq int 4, 5 <i>; yields {bool}:result = false</i>
1656 <result> = setne float 4, 5 <i>; yields {bool}:result = true</i>
1657 <result> = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1658 <result> = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1659 <result> = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1660 <result> = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1664 <!-- ======================================================================= -->
1665 <div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1666 Operations</a> </div>
1667 <div class="doc_text">
1668 <p>Bitwise binary operators are used to do various forms of
1669 bit-twiddling in a program. They are generally very efficient
1670 instructions and can commonly be strength reduced from other
1671 instructions. They require two operands, execute an operation on them,
1672 and produce a single value. The resulting value of the bitwise binary
1673 operators is always the same type as its first operand.</p>
1675 <!-- _______________________________________________________________________ -->
1676 <div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1677 Instruction</a> </div>
1678 <div class="doc_text">
1680 <pre> <result> = and <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1683 <p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1684 its two operands.</p>
1686 <p>The two arguments to the '<tt>and</tt>' instruction must be <a
1687 href="#t_integral">integral</a> values. Both arguments must have
1688 identical types.</p>
1690 <p>The truth table used for the '<tt>and</tt>' instruction is:</p>
1692 <div style="align: center">
1693 <table border="1" cellspacing="0" cellpadding="4">
1724 <pre> <result> = and int 4, %var <i>; yields {int}:result = 4 & %var</i>
1725 <result> = and int 15, 40 <i>; yields {int}:result = 8</i>
1726 <result> = and int 4, 8 <i>; yields {int}:result = 0</i>
1729 <!-- _______________________________________________________________________ -->
1730 <div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
1731 <div class="doc_text">
1733 <pre> <result> = or <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1736 <p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1737 or of its two operands.</p>
1739 <p>The two arguments to the '<tt>or</tt>' instruction must be <a
1740 href="#t_integral">integral</a> values. Both arguments must have
1741 identical types.</p>
1743 <p>The truth table used for the '<tt>or</tt>' instruction is:</p>
1745 <div style="align: center">
1746 <table border="1" cellspacing="0" cellpadding="4">
1777 <pre> <result> = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
1778 <result> = or int 15, 40 <i>; yields {int}:result = 47</i>
1779 <result> = or int 4, 8 <i>; yields {int}:result = 12</i>
1782 <!-- _______________________________________________________________________ -->
1783 <div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
1784 Instruction</a> </div>
1785 <div class="doc_text">
1787 <pre> <result> = xor <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1790 <p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
1791 or of its two operands. The <tt>xor</tt> is used to implement the
1792 "one's complement" operation, which is the "~" operator in C.</p>
1794 <p>The two arguments to the '<tt>xor</tt>' instruction must be <a
1795 href="#t_integral">integral</a> values. Both arguments must have
1796 identical types.</p>
1798 <p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
1800 <div style="align: center">
1801 <table border="1" cellspacing="0" cellpadding="4">
1833 <pre> <result> = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
1834 <result> = xor int 15, 40 <i>; yields {int}:result = 39</i>
1835 <result> = xor int 4, 8 <i>; yields {int}:result = 12</i>
1836 <result> = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
1839 <!-- _______________________________________________________________________ -->
1840 <div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
1841 Instruction</a> </div>
1842 <div class="doc_text">
1844 <pre> <result> = shl <ty> <var1>, ubyte <var2> <i>; yields {ty}:result</i>
1847 <p>The '<tt>shl</tt>' instruction returns the first operand shifted to
1848 the left a specified number of bits.</p>
1850 <p>The first argument to the '<tt>shl</tt>' instruction must be an <a
1851 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1854 <p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
1856 <pre> <result> = shl int 4, ubyte %var <i>; yields {int}:result = 4 << %var</i>
1857 <result> = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1858 <result> = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1861 <!-- _______________________________________________________________________ -->
1862 <div class="doc_subsubsection"> <a name="i_shr">'<tt>shr</tt>'
1863 Instruction</a> </div>
1864 <div class="doc_text">
1866 <pre> <result> = shr <ty> <var1>, ubyte <var2> <i>; yields {ty}:result</i>
1869 <p>The '<tt>shr</tt>' instruction returns the first operand shifted to
1870 the right a specified number of bits.</p>
1872 <p>The first argument to the '<tt>shr</tt>' instruction must be an <a
1873 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1876 <p>If the first argument is a <a href="#t_signed">signed</a> type, the
1877 most significant bit is duplicated in the newly free'd bit positions.
1878 If the first argument is unsigned, zero bits shall fill the empty
1881 <pre> <result> = shr int 4, ubyte %var <i>; yields {int}:result = 4 >> %var</i>
1882 <result> = shr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
1883 <result> = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
1884 <result> = shr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
1885 <result> = shr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
1889 <!-- ======================================================================= -->
1890 <div class="doc_subsection">
1891 <a name="memoryops">Memory Access Operations</a>
1894 <div class="doc_text">
1896 <p>A key design point of an SSA-based representation is how it
1897 represents memory. In LLVM, no memory locations are in SSA form, which
1898 makes things very simple. This section describes how to read, write,
1899 allocate, and free memory in LLVM.</p>
1903 <!-- _______________________________________________________________________ -->
1904 <div class="doc_subsubsection">
1905 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
1908 <div class="doc_text">
1913 <result> = malloc <type>[, uint <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
1918 <p>The '<tt>malloc</tt>' instruction allocates memory from the system
1919 heap and returns a pointer to it.</p>
1923 <p>The '<tt>malloc</tt>' instruction allocates
1924 <tt>sizeof(<type>)*NumElements</tt>
1925 bytes of memory from the operating system and returns a pointer of the
1926 appropriate type to the program. If "NumElements" is specified, it is the
1927 number of elements allocated. If an alignment is specified, the value result
1928 of the allocation is guaranteed to be aligned to at least that boundary. If
1929 not specified, or if zero, the target can choose to align the allocation on any
1930 convenient boundary.</p>
1932 <p>'<tt>type</tt>' must be a sized type.</p>
1936 <p>Memory is allocated using the system "<tt>malloc</tt>" function, and
1937 a pointer is returned.</p>
1942 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
1944 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
1945 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
1946 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
1947 %array3 = malloc int, uint 4, align 1024 <i>; yields {int*}:array3</i>
1948 %array4 = malloc int, align 1024 <i>; yields {int*}:array4</i>
1952 <!-- _______________________________________________________________________ -->
1953 <div class="doc_subsubsection">
1954 <a name="i_free">'<tt>free</tt>' Instruction</a>
1957 <div class="doc_text">
1962 free <type> <value> <i>; yields {void}</i>
1967 <p>The '<tt>free</tt>' instruction returns memory back to the unused
1968 memory heap to be reallocated in the future.</p>
1972 <p>'<tt>value</tt>' shall be a pointer value that points to a value
1973 that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
1978 <p>Access to the memory pointed to by the pointer is no longer defined
1979 after this instruction executes.</p>
1984 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
1985 free [4 x ubyte]* %array
1989 <!-- _______________________________________________________________________ -->
1990 <div class="doc_subsubsection">
1991 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
1994 <div class="doc_text">
1999 <result> = alloca <type>[, uint <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
2004 <p>The '<tt>alloca</tt>' instruction allocates memory on the current
2005 stack frame of the procedure that is live until the current function
2006 returns to its caller.</p>
2010 <p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(<type>)*NumElements</tt>
2011 bytes of memory on the runtime stack, returning a pointer of the
2012 appropriate type to the program. If "NumElements" is specified, it is the
2013 number of elements allocated. If an alignment is specified, the value result
2014 of the allocation is guaranteed to be aligned to at least that boundary. If
2015 not specified, or if zero, the target can choose to align the allocation on any
2016 convenient boundary.</p>
2018 <p>'<tt>type</tt>' may be any sized type.</p>
2022 <p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
2023 memory is automatically released when the function returns. The '<tt>alloca</tt>'
2024 instruction is commonly used to represent automatic variables that must
2025 have an address available. When the function returns (either with the <tt><a
2026 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
2027 instructions), the memory is reclaimed.</p>
2032 %ptr = alloca int <i>; yields {int*}:ptr</i>
2033 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
2034 %ptr = alloca int, uint 4, align 1024 <i>; yields {int*}:ptr</i>
2035 %ptr = alloca int, align 1024 <i>; yields {int*}:ptr</i>
2039 <!-- _______________________________________________________________________ -->
2040 <div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2041 Instruction</a> </div>
2042 <div class="doc_text">
2044 <pre> <result> = load <ty>* <pointer><br> <result> = volatile load <ty>* <pointer><br></pre>
2046 <p>The '<tt>load</tt>' instruction is used to read from memory.</p>
2048 <p>The argument to the '<tt>load</tt>' instruction specifies the memory
2049 address from which to load. The pointer must point to a <a
2050 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
2051 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
2052 the number or order of execution of this <tt>load</tt> with other
2053 volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2056 <p>The location of memory pointed to is loaded.</p>
2058 <pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2060 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
2061 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2064 <!-- _______________________________________________________________________ -->
2065 <div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2066 Instruction</a> </div>
2068 <pre> store <ty> <value>, <ty>* <pointer> <i>; yields {void}</i>
2069 volatile store <ty> <value>, <ty>* <pointer> <i>; yields {void}</i>
2072 <p>The '<tt>store</tt>' instruction is used to write to memory.</p>
2074 <p>There are two arguments to the '<tt>store</tt>' instruction: a value
2075 to store and an address in which to store it. The type of the '<tt><pointer></tt>'
2076 operand must be a pointer to the type of the '<tt><value></tt>'
2077 operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
2078 optimizer is not allowed to modify the number or order of execution of
2079 this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2080 href="#i_store">store</a></tt> instructions.</p>
2082 <p>The contents of memory are updated to contain '<tt><value></tt>'
2083 at the location specified by the '<tt><pointer></tt>' operand.</p>
2085 <pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2087 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
2088 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2090 <!-- _______________________________________________________________________ -->
2091 <div class="doc_subsubsection">
2092 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2095 <div class="doc_text">
2098 <result> = getelementptr <ty>* <ptrval>{, <ty> <idx>}*
2104 The '<tt>getelementptr</tt>' instruction is used to get the address of a
2105 subelement of an aggregate data structure.</p>
2109 <p>This instruction takes a list of integer constants that indicate what
2110 elements of the aggregate object to index to. The actual types of the arguments
2111 provided depend on the type of the first pointer argument. The
2112 '<tt>getelementptr</tt>' instruction is used to index down through the type
2113 levels of a structure or to a specific index in an array. When indexing into a
2114 structure, only <tt>uint</tt>
2115 integer constants are allowed. When indexing into an array or pointer,
2116 <tt>int</tt> and <tt>long</tt> indexes are allowed of any sign.</p>
2118 <p>For example, let's consider a C code fragment and how it gets
2119 compiled to LLVM:</p>
2133 int *foo(struct ST *s) {
2134 return &s[1].Z.B[5][13];
2138 <p>The LLVM code generated by the GCC frontend is:</p>
2141 %RT = type { sbyte, [10 x [20 x int]], sbyte }
2142 %ST = type { int, double, %RT }
2146 int* %foo(%ST* %s) {
2148 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
2155 <p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
2156 on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
2157 and <a href="#t_array">array</a> types require <tt>uint</tt>, <tt>int</tt>,
2158 <tt>ulong</tt>, or <tt>long</tt> values, and <a href="#t_struct">structure</a>
2159 types require <tt>uint</tt> <b>constants</b>.</p>
2161 <p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
2162 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
2163 }</tt>' type, a structure. The second index indexes into the third element of
2164 the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
2165 sbyte }</tt>' type, another structure. The third index indexes into the second
2166 element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
2167 array. The two dimensions of the array are subscripted into, yielding an
2168 '<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
2169 to this element, thus computing a value of '<tt>int*</tt>' type.</p>
2171 <p>Note that it is perfectly legal to index partially through a
2172 structure, returning a pointer to an inner element. Because of this,
2173 the LLVM code for the given testcase is equivalent to:</p>
2176 int* %foo(%ST* %s) {
2177 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
2178 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
2179 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
2180 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
2181 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
2186 <p>Note that it is undefined to access an array out of bounds: array and
2187 pointer indexes must always be within the defined bounds of the array type.
2188 The one exception for this rules is zero length arrays. These arrays are
2189 defined to be accessible as variable length arrays, which requires access
2190 beyond the zero'th element.</p>
2195 <i>; yields [12 x ubyte]*:aptr</i>
2196 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
2200 <!-- ======================================================================= -->
2201 <div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
2202 <div class="doc_text">
2203 <p>The instructions in this category are the "miscellaneous"
2204 instructions, which defy better classification.</p>
2206 <!-- _______________________________________________________________________ -->
2207 <div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
2208 Instruction</a> </div>
2209 <div class="doc_text">
2211 <pre> <result> = phi <ty> [ <val0>, <label0>], ...<br></pre>
2213 <p>The '<tt>phi</tt>' instruction is used to implement the φ node in
2214 the SSA graph representing the function.</p>
2216 <p>The type of the incoming values are specified with the first type
2217 field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
2218 as arguments, with one pair for each predecessor basic block of the
2219 current block. Only values of <a href="#t_firstclass">first class</a>
2220 type may be used as the value arguments to the PHI node. Only labels
2221 may be used as the label arguments.</p>
2222 <p>There must be no non-phi instructions between the start of a basic
2223 block and the PHI instructions: i.e. PHI instructions must be first in
2226 <p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
2227 value specified by the parameter, depending on which basic block we
2228 came from in the last <a href="#terminators">terminator</a> instruction.</p>
2230 <pre>Loop: ; Infinite loop that counts from 0 on up...<br> %indvar = phi uint [ 0, %LoopHeader ], [ %nextindvar, %Loop ]<br> %nextindvar = add uint %indvar, 1<br> br label %Loop<br></pre>
2233 <!-- _______________________________________________________________________ -->
2234 <div class="doc_subsubsection">
2235 <a name="i_cast">'<tt>cast .. to</tt>' Instruction</a>
2238 <div class="doc_text">
2243 <result> = cast <ty> <value> to <ty2> <i>; yields ty2</i>
2249 The '<tt>cast</tt>' instruction is used as the primitive means to convert
2250 integers to floating point, change data type sizes, and break type safety (by
2258 The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
2259 class value, and a type to cast it to, which must also be a <a
2260 href="#t_firstclass">first class</a> type.
2266 This instruction follows the C rules for explicit casts when determining how the
2267 data being cast must change to fit in its new container.
2271 When casting to bool, any value that would be considered true in the context of
2272 a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
2273 all else are '<tt>false</tt>'.
2277 When extending an integral value from a type of one signness to another (for
2278 example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
2279 <b>source</b> value is signed, and zero-extended if the source value is
2280 unsigned. <tt>bool</tt> values are always zero extended into either zero or
2287 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
2288 %Y = cast int 123 to bool <i>; yields bool:true</i>
2292 <!-- _______________________________________________________________________ -->
2293 <div class="doc_subsubsection">
2294 <a name="i_select">'<tt>select</tt>' Instruction</a>
2297 <div class="doc_text">
2302 <result> = select bool <cond>, <ty> <val1>, <ty> <val2> <i>; yields ty</i>
2308 The '<tt>select</tt>' instruction is used to choose one value based on a
2309 condition, without branching.
2316 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.
2322 If the boolean condition evaluates to true, the instruction returns the first
2323 value argument; otherwise, it returns the second value argument.
2329 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
2333 <!-- _______________________________________________________________________ -->
2334 <div class="doc_subsubsection"> <a name="i_vset">'<tt>vset</tt>'
2335 Instruction</a> </div>
2336 <div class="doc_text">
2338 <pre><result> = vset <op>, <n x <ty>> <var1>, <var2> <i>; yields <n x bool></i>
2343 <p>The '<tt>vset</tt>' instruction returns a vector of boolean
2344 values representing, at each position, the result of the comparison
2345 between the values at that position in the two operands.</p>
2349 <p>The arguments to a '<tt>vset</tt>' instruction are a comparison
2350 operation and two value arguments. The value arguments must be of <a
2351 href="#t_packed">packed</a> type, and they must have identical types.
2352 For value arguments of integral element type, the operation argument
2353 must be one of <tt>eq</tt>, <tt>ne</tt>, <tt>lt</tt>, <tt>gt</tt>,
2354 <tt>le</tt>, <tt>ge</tt>, <tt>ult</tt>, <tt>ugt</tt>, <tt>ule</tt>,
2355 <tt>uge</tt>, <tt>true</tt>, and <tt>false</tt>. For value arguments
2356 of floating point element type, the operation argument must be one of
2357 <tt>eq</tt>, <tt>ne</tt>, <tt>lt</tt>, <tt>gt</tt>, <tt>le</tt>,
2358 <tt>ge</tt>, <tt>oeq</tt>, <tt>one</tt>, <tt>olt</tt>, <tt>ogt</tt>,
2359 <tt>ole</tt>, <tt>oge</tt>, <tt>ueq</tt>, <tt>une</tt>, <tt>ult</tt>,
2360 <tt>ugt</tt>, <tt>ule</tt>, <tt>uge</tt>, <tt>o</tt>, <tt>u</tt>,
2361 <tt>true</tt>, and <tt>false</tt>. The result is a packed
2362 <tt>bool</tt> value with the same length as each operand.</p>
2366 <p>The following table shows the semantics of '<tt>vset</tt>' for
2367 integral value arguments. For each position of the result, the
2368 comparison is done on the corresponding positions of the two value
2369 arguments. Note that the signedness of the comparison depends on the
2370 comparison opcode and <i>not</i> on the signedness of the value
2371 operands. E.g., <tt>vset lt <4 x unsigned> %x, %y</tt> does an
2372 elementwise <i>signed</i> comparison of <tt>%x</tt> and
2375 <table border="1" cellspacing="0" cellpadding="4">
2377 <tr><th>Operation</th><th>Result is true iff</th><th>Comparison is</th></tr>
2378 <tr><td><tt>eq</tt></td><td>var1 == var2</td><td>--</td></tr>
2379 <tr><td><tt>ne</tt></td><td>var1 != var2</td><td>--</td></tr>
2380 <tr><td><tt>lt</tt></td><td>var1 < var2</td><td>signed</td></tr>
2381 <tr><td><tt>gt</tt></td><td>var1 > var2</td><td>signed</td></tr>
2382 <tr><td><tt>le</tt></td><td>var1 <= var2</td><td>signed</td></tr>
2383 <tr><td><tt>ge</tt></td><td>var1 >= var2</td><td>signed</td></tr>
2384 <tr><td><tt>ult</tt></td><td>var1 < var2</td><td>unsigned</td></tr>
2385 <tr><td><tt>ugt</tt></td><td>var1 > var2</td><td>unsigned</td></tr>
2386 <tr><td><tt>ule</tt></td><td>var1 <= var2</td><td>unsigned</td></tr>
2387 <tr><td><tt>uge</tt></td><td>var1 >= var2</td><td>unsigned</td></tr>
2388 <tr><td><tt>true</tt></td><td>always</td><td>--</td></tr>
2389 <tr><td><tt>false</tt></td><td>never</td><td>--</td></tr>
2393 <p>The following table shows the semantics of '<tt>vset</tt>' for
2394 floating point types. If either operand is a floating point Not a
2395 Number (NaN) value, the operation is unordered, and the value in the
2396 first column below is produced at that position. Otherwise, the
2397 operation is ordered, and the value in the second column is
2400 <table border="1" cellspacing="0" cellpadding="4">
2402 <tr><th>Operation</th><th>If unordered<th>Otherwise true iff</th></tr>
2403 <tr><td><tt>eq</tt></td><td>undefined</td><td>var1 == var2</td></tr>
2404 <tr><td><tt>ne</tt></td><td>undefined</td><td>var1 != var2</td></tr>
2405 <tr><td><tt>lt</tt></td><td>undefined</td><td>var1 < var2</td></tr>
2406 <tr><td><tt>gt</tt></td><td>undefined</td><td>var1 > var2</td></tr>
2407 <tr><td><tt>le</tt></td><td>undefined</td><td>var1 <= var2</td></tr>
2408 <tr><td><tt>ge</tt></td><td>undefined</td><td>var1 >= var2</td></tr>
2409 <tr><td><tt>oeq</tt></td><td>false</td><td>var1 == var2</td></tr>
2410 <tr><td><tt>one</tt></td><td>false</td><td>var1 != var2</td></tr>
2411 <tr><td><tt>olt</tt></td><td>false</td><td>var1 < var2</td></tr>
2412 <tr><td><tt>ogt</tt></td><td>false</td><td>var1 > var2</td></tr>
2413 <tr><td><tt>ole</tt></td><td>false</td><td>var1 <= var2</td></tr>
2414 <tr><td><tt>oge</tt></td><td>false</td><td>var1 >= var2</td></tr>
2415 <tr><td><tt>ueq</tt></td><td>true</td><td>var1 == var2</td></tr>
2416 <tr><td><tt>une</tt></td><td>true</td><td>var1 != var2</td></tr>
2417 <tr><td><tt>ult</tt></td><td>true</td><td>var1 < var2</td></tr>
2418 <tr><td><tt>ugt</tt></td><td>true</td><td>var1 > var2</td></tr>
2419 <tr><td><tt>ule</tt></td><td>true</td><td>var1 <= var2</td></tr>
2420 <tr><td><tt>uge</tt></td><td>true</td><td>var1 >= var2</td></tr>
2421 <tr><td><tt>o</tt></td><td>false</td><td>always</td></tr>
2422 <tr><td><tt>u</tt></td><td>true</td><td>never</td></tr>
2423 <tr><td><tt>true</tt></td><td>true</td><td>always</td></tr>
2424 <tr><td><tt>false</tt></td><td>false</td><td>never</td></tr>
2429 <pre> <result> = vset eq <2 x int> <int 0, int 1>, <int 1, int 0> <i>; yields {<2 x bool>}:result = false, false</i>
2430 <result> = vset ne <2 x int> <int 0, int 1>, <int 1, int 0> <i>; yields {<2 x bool>}:result = true, true</i>
2431 <result> = vset lt <2 x int> <int 0, int 1>, <int 1, int 0> <i>; yields {<2 x bool>}:result = true, false</i>
2432 <result> = vset gt <2 x int> <int 0, int 1>, <int 1, int 0> <i>; yields {<2 x bool>}:result = false, true</i>
2433 <result> = vset le <2 x int> <int 0, int 1>, <int 1, int 0> <i>; yields {<2 x bool>}:result = true, false</i>
2434 <result> = vset ge <2 x int> <int 0, int 1>, <int 1, int 0> <i>; yields {<2 x bool>}:result = false, true</i>
2438 <!-- _______________________________________________________________________ -->
2439 <div class="doc_subsubsection">
2440 <a name="i_vselect">'<tt>vselect</tt>' Instruction</a>
2443 <div class="doc_text">
2448 <result> = vselect <n x bool> <cond>, <n x <ty>> <val1>, <n x <ty>> <val2> <i>; yields <n x <ty>></i>
2454 The '<tt>vselect</tt>' instruction chooses one value at each position
2455 of a vector based on a condition.
2462 The '<tt>vselect</tt>' instruction requires a <a
2463 href="#t_packed">packed</a> <tt>bool</tt> value indicating the
2464 condition at each vector position, and two values of the same packed
2465 type. All three operands must have the same length. The type of the
2466 result is the same as the type of the two value operands.</p>
2471 At each position where the <tt>bool</tt> vector is true, that position
2472 of the result gets its value from the first value argument; otherwise,
2473 it gets its value from the second value argument.
2479 %X = vselect bool <2 x bool> <bool true, bool false>, <2 x ubyte> <ubyte 17, ubyte 17>,
2480 <2 x ubyte> <ubyte 42, ubyte 42> <i>; yields <2 x ubyte>:17, 42</i>
2484 <!-- _______________________________________________________________________ -->
2485 <div class="doc_subsubsection">
2486 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2489 <div class="doc_text">
2494 <result> = extractelement <n x <ty>> <val>, uint <idx> <i>; yields <ty></i>
2500 The '<tt>extractelement</tt>' instruction extracts a single scalar
2501 element from a packed vector at a specified index.
2508 The first operand of an '<tt>extractelement</tt>' instruction is a
2509 value of <a href="#t_packed">packed</a> type. The second operand is
2510 an index indicating the position from which to extract the element.
2511 The index may be a variable.</p>
2516 The result is a scalar of the same type as the element type of
2517 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2518 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2519 results are undefined.
2525 %result = extractelement <4 x int> %vec, uint 0 <i>; yields int</i>
2530 <!-- _______________________________________________________________________ -->
2531 <div class="doc_subsubsection">
2532 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2535 <div class="doc_text">
2540 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, uint <idx> <i>; yields <n x <ty>></i>
2546 The '<tt>insertelement</tt>' instruction inserts a scalar
2547 element into a packed vector at a specified index.
2554 The first operand of an '<tt>insertelement</tt>' instruction is a
2555 value of <a href="#t_packed">packed</a> type. The second operand is a
2556 scalar value whose type must equal the element type of the first
2557 operand. The third operand is an index indicating the position at
2558 which to insert the value. The index may be a variable.</p>
2563 The result is a packed vector of the same type as <tt>val</tt>. Its
2564 element values are those of <tt>val</tt> except at position
2565 <tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2566 exceeds the length of <tt>val</tt>, the results are undefined.
2572 %result = insertelement <4 x int> %vec, int 1, uint 0 <i>; yields <4 x int></i>
2577 <!-- _______________________________________________________________________ -->
2578 <div class="doc_subsubsection">
2579 <a name="i_call">'<tt>call</tt>' Instruction</a>
2582 <div class="doc_text">
2586 <result> = [tail] call [<a href="#callingconv">cconv</a>] <ty>* <fnptrval>(<param list>)
2591 <p>The '<tt>call</tt>' instruction represents a simple function call.</p>
2595 <p>This instruction requires several arguments:</p>
2599 <p>The optional "tail" marker indicates whether the callee function accesses
2600 any allocas or varargs in the caller. If the "tail" marker is present, the
2601 function call is eligible for tail call optimization. Note that calls may
2602 be marked "tail" even if they do not occur before a <a
2603 href="#i_ret"><tt>ret</tt></a> instruction.
2606 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
2607 convention</a> the call should use. If none is specified, the call defaults
2608 to using C calling conventions.
2611 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
2612 being invoked. The argument types must match the types implied by this
2613 signature. This type can be omitted if the function is not varargs and
2614 if the function type does not return a pointer to a function.</p>
2617 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
2618 be invoked. In most cases, this is a direct function invocation, but
2619 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
2620 to function value.</p>
2623 <p>'<tt>function args</tt>': argument list whose types match the
2624 function signature argument types. All arguments must be of
2625 <a href="#t_firstclass">first class</a> type. If the function signature
2626 indicates the function accepts a variable number of arguments, the extra
2627 arguments can be specified.</p>
2633 <p>The '<tt>call</tt>' instruction is used to cause control flow to
2634 transfer to a specified function, with its incoming arguments bound to
2635 the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
2636 instruction in the called function, control flow continues with the
2637 instruction after the function call, and the return value of the
2638 function is bound to the result argument. This is a simpler case of
2639 the <a href="#i_invoke">invoke</a> instruction.</p>
2644 %retval = call int %test(int %argc)
2645 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
2646 %X = tail call int %foo()
2647 %Y = tail call <a href="#callingconv">fastcc</a> int %foo()
2652 <!-- _______________________________________________________________________ -->
2653 <div class="doc_subsubsection">
2654 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
2657 <div class="doc_text">
2662 <resultval> = va_arg <va_list*> <arglist>, <argty>
2667 <p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
2668 the "variable argument" area of a function call. It is used to implement the
2669 <tt>va_arg</tt> macro in C.</p>
2673 <p>This instruction takes a <tt>va_list*</tt> value and the type of
2674 the argument. It returns a value of the specified argument type and
2675 increments the <tt>va_list</tt> to point to the next argument. Again, the
2676 actual type of <tt>va_list</tt> is target specific.</p>
2680 <p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
2681 type from the specified <tt>va_list</tt> and causes the
2682 <tt>va_list</tt> to point to the next argument. For more information,
2683 see the variable argument handling <a href="#int_varargs">Intrinsic
2686 <p>It is legal for this instruction to be called in a function which does not
2687 take a variable number of arguments, for example, the <tt>vfprintf</tt>
2690 <p><tt>va_arg</tt> is an LLVM instruction instead of an <a
2691 href="#intrinsics">intrinsic function</a> because it takes a type as an
2696 <p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
2700 <!-- *********************************************************************** -->
2701 <div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
2702 <!-- *********************************************************************** -->
2704 <div class="doc_text">
2706 <p>LLVM supports the notion of an "intrinsic function". These functions have
2707 well known names and semantics and are required to follow certain
2708 restrictions. Overall, these instructions represent an extension mechanism for
2709 the LLVM language that does not require changing all of the transformations in
2710 LLVM to add to the language (or the bytecode reader/writer, the parser,
2713 <p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
2714 prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
2715 this. Intrinsic functions must always be external functions: you cannot define
2716 the body of intrinsic functions. Intrinsic functions may only be used in call
2717 or invoke instructions: it is illegal to take the address of an intrinsic
2718 function. Additionally, because intrinsic functions are part of the LLVM
2719 language, it is required that they all be documented here if any are added.</p>
2722 <p>To learn how to add an intrinsic function, please see the <a
2723 href="ExtendingLLVM.html">Extending LLVM Guide</a>.
2728 <!-- ======================================================================= -->
2729 <div class="doc_subsection">
2730 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
2733 <div class="doc_text">
2735 <p>Variable argument support is defined in LLVM with the <a
2736 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
2737 intrinsic functions. These functions are related to the similarly
2738 named macros defined in the <tt><stdarg.h></tt> header file.</p>
2740 <p>All of these functions operate on arguments that use a
2741 target-specific value type "<tt>va_list</tt>". The LLVM assembly
2742 language reference manual does not define what this type is, so all
2743 transformations should be prepared to handle intrinsics with any type
2746 <p>This example shows how the <a href="#i_vanext"><tt>vanext</tt></a>
2747 instruction and the variable argument handling intrinsic functions are
2751 int %test(int %X, ...) {
2752 ; Initialize variable argument processing
2754 call void %<a href="#i_va_start">llvm.va_start</a>(sbyte** %ap)
2756 ; Read a single integer argument
2757 %tmp = va_arg sbyte** %ap, int
2759 ; Demonstrate usage of llvm.va_copy and llvm.va_end
2761 call void %<a href="#i_va_copy">llvm.va_copy</a>(sbyte** %aq, sbyte** %ap)
2762 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %aq)
2764 ; Stop processing of arguments.
2765 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %ap)
2771 <!-- _______________________________________________________________________ -->
2772 <div class="doc_subsubsection">
2773 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
2777 <div class="doc_text">
2779 <pre> declare void %llvm.va_start(<va_list>* <arglist>)<br></pre>
2781 <P>The '<tt>llvm.va_start</tt>' intrinsic initializes
2782 <tt>*<arglist></tt> for subsequent use by <tt><a
2783 href="#i_va_arg">va_arg</a></tt>.</p>
2787 <P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
2791 <P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
2792 macro available in C. In a target-dependent way, it initializes the
2793 <tt>va_list</tt> element the argument points to, so that the next call to
2794 <tt>va_arg</tt> will produce the first variable argument passed to the function.
2795 Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
2796 last argument of the function, the compiler can figure that out.</p>
2800 <!-- _______________________________________________________________________ -->
2801 <div class="doc_subsubsection">
2802 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
2805 <div class="doc_text">
2807 <pre> declare void %llvm.va_end(<va_list*> <arglist>)<br></pre>
2809 <p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt><arglist></tt>
2810 which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
2811 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
2813 <p>The argument is a <tt>va_list</tt> to destroy.</p>
2815 <p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
2816 macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
2817 Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
2818 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
2819 with calls to <tt>llvm.va_end</tt>.</p>
2822 <!-- _______________________________________________________________________ -->
2823 <div class="doc_subsubsection">
2824 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
2827 <div class="doc_text">
2832 declare void %llvm.va_copy(<va_list>* <destarglist>,
2833 <va_list>* <srcarglist>)
2838 <p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
2839 the source argument list to the destination argument list.</p>
2843 <p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
2844 The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
2849 <p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
2850 available in C. In a target-dependent way, it copies the source
2851 <tt>va_list</tt> element into the destination list. This intrinsic is necessary
2852 because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
2853 arbitrarily complex and require memory allocation, for example.</p>
2857 <!-- ======================================================================= -->
2858 <div class="doc_subsection">
2859 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
2862 <div class="doc_text">
2865 LLVM support for <a href="GarbageCollection.html">Accurate Garbage
2866 Collection</a> requires the implementation and generation of these intrinsics.
2867 These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
2868 stack</a>, as well as garbage collector implementations that require <a
2869 href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
2870 Front-ends for type-safe garbage collected languages should generate these
2871 intrinsics to make use of the LLVM garbage collectors. For more details, see <a
2872 href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
2876 <!-- _______________________________________________________________________ -->
2877 <div class="doc_subsubsection">
2878 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
2881 <div class="doc_text">
2886 declare void %llvm.gcroot(<ty>** %ptrloc, <ty2>* %metadata)
2891 <p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
2892 the code generator, and allows some metadata to be associated with it.</p>
2896 <p>The first argument specifies the address of a stack object that contains the
2897 root pointer. The second pointer (which must be either a constant or a global
2898 value address) contains the meta-data to be associated with the root.</p>
2902 <p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
2903 location. At compile-time, the code generator generates information to allow
2904 the runtime to find the pointer at GC safe points.
2910 <!-- _______________________________________________________________________ -->
2911 <div class="doc_subsubsection">
2912 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
2915 <div class="doc_text">
2920 declare sbyte* %llvm.gcread(sbyte* %ObjPtr, sbyte** %Ptr)
2925 <p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
2926 locations, allowing garbage collector implementations that require read
2931 <p>The second argument is the address to read from, which should be an address
2932 allocated from the garbage collector. The first object is a pointer to the
2933 start of the referenced object, if needed by the language runtime (otherwise
2938 <p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
2939 instruction, but may be replaced with substantially more complex code by the
2940 garbage collector runtime, as needed.</p>
2945 <!-- _______________________________________________________________________ -->
2946 <div class="doc_subsubsection">
2947 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
2950 <div class="doc_text">
2955 declare void %llvm.gcwrite(sbyte* %P1, sbyte* %Obj, sbyte** %P2)
2960 <p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
2961 locations, allowing garbage collector implementations that require write
2962 barriers (such as generational or reference counting collectors).</p>
2966 <p>The first argument is the reference to store, the second is the start of the
2967 object to store it to, and the third is the address of the field of Obj to
2968 store to. If the runtime does not require a pointer to the object, Obj may be
2973 <p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
2974 instruction, but may be replaced with substantially more complex code by the
2975 garbage collector runtime, as needed.</p>
2981 <!-- ======================================================================= -->
2982 <div class="doc_subsection">
2983 <a name="int_codegen">Code Generator Intrinsics</a>
2986 <div class="doc_text">
2988 These intrinsics are provided by LLVM to expose special features that may only
2989 be implemented with code generator support.
2994 <!-- _______________________________________________________________________ -->
2995 <div class="doc_subsubsection">
2996 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
2999 <div class="doc_text">
3003 declare sbyte *%llvm.returnaddress(uint <level>)
3009 The '<tt>llvm.returnaddress</tt>' intrinsic returns a target-specific value
3010 indicating the return address of the current function or one of its callers.
3016 The argument to this intrinsic indicates which function to return the address
3017 for. Zero indicates the calling function, one indicates its caller, etc. The
3018 argument is <b>required</b> to be a constant integer value.
3024 The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
3025 the return address of the specified call frame, or zero if it cannot be
3026 identified. The value returned by this intrinsic is likely to be incorrect or 0
3027 for arguments other than zero, so it should only be used for debugging purposes.
3031 Note that calling this intrinsic does not prevent function inlining or other
3032 aggressive transformations, so the value returned may not be that of the obvious
3033 source-language caller.
3038 <!-- _______________________________________________________________________ -->
3039 <div class="doc_subsubsection">
3040 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
3043 <div class="doc_text">
3047 declare sbyte *%llvm.frameaddress(uint <level>)
3053 The '<tt>llvm.frameaddress</tt>' intrinsic returns the target-specific frame
3054 pointer value for the specified stack frame.
3060 The argument to this intrinsic indicates which function to return the frame
3061 pointer for. Zero indicates the calling function, one indicates its caller,
3062 etc. The argument is <b>required</b> to be a constant integer value.
3068 The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
3069 the frame address of the specified call frame, or zero if it cannot be
3070 identified. The value returned by this intrinsic is likely to be incorrect or 0
3071 for arguments other than zero, so it should only be used for debugging purposes.
3075 Note that calling this intrinsic does not prevent function inlining or other
3076 aggressive transformations, so the value returned may not be that of the obvious
3077 source-language caller.
3081 <!-- _______________________________________________________________________ -->
3082 <div class="doc_subsubsection">
3083 <a name="i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
3086 <div class="doc_text">
3090 declare sbyte *%llvm.stacksave()
3096 The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
3097 the function stack, for use with <a href="#i_stackrestore">
3098 <tt>llvm.stackrestore</tt></a>. This is useful for implementing language
3099 features like scoped automatic variable sized arrays in C99.
3105 This intrinsic returns a opaque pointer value that can be passed to <a
3106 href="#i_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
3107 <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
3108 <tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
3109 state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
3110 practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
3111 that were allocated after the <tt>llvm.stacksave</tt> was executed.
3116 <!-- _______________________________________________________________________ -->
3117 <div class="doc_subsubsection">
3118 <a name="i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
3121 <div class="doc_text">
3125 declare void %llvm.stackrestore(sbyte* %ptr)
3131 The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
3132 the function stack to the state it was in when the corresponding <a
3133 href="#llvm.stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
3134 useful for implementing language features like scoped automatic variable sized
3141 See the description for <a href="#i_stacksave"><tt>llvm.stacksave</tt></a>.
3147 <!-- _______________________________________________________________________ -->
3148 <div class="doc_subsubsection">
3149 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
3152 <div class="doc_text">
3156 declare void %llvm.prefetch(sbyte * <address>,
3157 uint <rw>, uint <locality>)
3164 The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
3165 a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
3167 effect on the behavior of the program but can change its performance
3174 <tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
3175 determining if the fetch should be for a read (0) or write (1), and
3176 <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
3177 locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
3178 <tt>locality</tt> arguments must be constant integers.
3184 This intrinsic does not modify the behavior of the program. In particular,
3185 prefetches cannot trap and do not produce a value. On targets that support this
3186 intrinsic, the prefetch can provide hints to the processor cache for better
3192 <!-- _______________________________________________________________________ -->
3193 <div class="doc_subsubsection">
3194 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
3197 <div class="doc_text">
3201 declare void %llvm.pcmarker( uint <id> )
3208 The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
3210 code to simulators and other tools. The method is target specific, but it is
3211 expected that the marker will use exported symbols to transmit the PC of the marker.
3212 The marker makes no guarantees that it will remain with any specific instruction
3213 after optimizations. It is possible that the presence of a marker will inhibit
3214 optimizations. The intended use is to be inserted after optmizations to allow
3215 correlations of simulation runs.
3221 <tt>id</tt> is a numerical id identifying the marker.
3227 This intrinsic does not modify the behavior of the program. Backends that do not
3228 support this intrinisic may ignore it.
3233 <!-- _______________________________________________________________________ -->
3234 <div class="doc_subsubsection">
3235 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
3238 <div class="doc_text">
3242 declare ulong %llvm.readcyclecounter( )
3249 The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
3250 counter register (or similar low latency, high accuracy clocks) on those targets
3251 that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
3252 As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
3253 should only be used for small timings.
3259 When directly supported, reading the cycle counter should not modify any memory.
3260 Implementations are allowed to either return a application specific value or a
3261 system wide value. On backends without support, this is lowered to a constant 0.
3266 <!-- ======================================================================= -->
3267 <div class="doc_subsection">
3268 <a name="int_libc">Standard C Library Intrinsics</a>
3271 <div class="doc_text">
3273 LLVM provides intrinsics for a few important standard C library functions.
3274 These intrinsics allow source-language front-ends to pass information about the
3275 alignment of the pointer arguments to the code generator, providing opportunity
3276 for more efficient code generation.
3281 <!-- _______________________________________________________________________ -->
3282 <div class="doc_subsubsection">
3283 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
3286 <div class="doc_text">
3290 declare void %llvm.memcpy.i32(sbyte* <dest>, sbyte* <src>,
3291 uint <len>, uint <align>)
3292 declare void %llvm.memcpy.i64(sbyte* <dest>, sbyte* <src>,
3293 ulong <len>, uint <align>)
3299 The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
3300 location to the destination location.
3304 Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
3305 intrinsics do not return a value, and takes an extra alignment argument.
3311 The first argument is a pointer to the destination, the second is a pointer to
3312 the source. The third argument is an integer argument
3313 specifying the number of bytes to copy, and the fourth argument is the alignment
3314 of the source and destination locations.
3318 If the call to this intrinisic has an alignment value that is not 0 or 1, then
3319 the caller guarantees that both the source and destination pointers are aligned
3326 The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
3327 location to the destination location, which are not allowed to overlap. It
3328 copies "len" bytes of memory over. If the argument is known to be aligned to
3329 some boundary, this can be specified as the fourth argument, otherwise it should
3335 <!-- _______________________________________________________________________ -->
3336 <div class="doc_subsubsection">
3337 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
3340 <div class="doc_text">
3344 declare void %llvm.memmove.i32(sbyte* <dest>, sbyte* <src>,
3345 uint <len>, uint <align>)
3346 declare void %llvm.memmove.i64(sbyte* <dest>, sbyte* <src>,
3347 ulong <len>, uint <align>)
3353 The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
3354 location to the destination location. It is similar to the
3355 '<tt>llvm.memcmp</tt>' intrinsic but allows the two memory locations to overlap.
3359 Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
3360 intrinsics do not return a value, and takes an extra alignment argument.
3366 The first argument is a pointer to the destination, the second is a pointer to
3367 the source. The third argument is an integer argument
3368 specifying the number of bytes to copy, and the fourth argument is the alignment
3369 of the source and destination locations.
3373 If the call to this intrinisic has an alignment value that is not 0 or 1, then
3374 the caller guarantees that the source and destination pointers are aligned to
3381 The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
3382 location to the destination location, which may overlap. It
3383 copies "len" bytes of memory over. If the argument is known to be aligned to
3384 some boundary, this can be specified as the fourth argument, otherwise it should
3390 <!-- _______________________________________________________________________ -->
3391 <div class="doc_subsubsection">
3392 <a name="i_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
3395 <div class="doc_text">
3399 declare void %llvm.memset.i32(sbyte* <dest>, ubyte <val>,
3400 uint <len>, uint <align>)
3401 declare void %llvm.memset.i64(sbyte* <dest>, ubyte <val>,
3402 ulong <len>, uint <align>)
3408 The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
3413 Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
3414 does not return a value, and takes an extra alignment argument.
3420 The first argument is a pointer to the destination to fill, the second is the
3421 byte value to fill it with, the third argument is an integer
3422 argument specifying the number of bytes to fill, and the fourth argument is the
3423 known alignment of destination location.
3427 If the call to this intrinisic has an alignment value that is not 0 or 1, then
3428 the caller guarantees that the destination pointer is aligned to that boundary.
3434 The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
3436 destination location. If the argument is known to be aligned to some boundary,
3437 this can be specified as the fourth argument, otherwise it should be set to 0 or
3443 <!-- _______________________________________________________________________ -->
3444 <div class="doc_subsubsection">
3445 <a name="i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a>
3448 <div class="doc_text">
3452 declare bool %llvm.isunordered.f32(float Val1, float Val2)
3453 declare bool %llvm.isunordered.f64(double Val1, double Val2)
3459 The '<tt>llvm.isunordered</tt>' intrinsics return true if either or both of the
3460 specified floating point values is a NAN.
3466 The arguments are floating point numbers of the same type.
3472 If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
3478 <!-- _______________________________________________________________________ -->
3479 <div class="doc_subsubsection">
3480 <a name="i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
3483 <div class="doc_text">
3487 declare double %llvm.sqrt.f32(float Val)
3488 declare double %llvm.sqrt.f64(double Val)
3494 The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
3495 returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
3496 <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
3497 negative numbers (which allows for better optimization).
3503 The argument and return value are floating point numbers of the same type.
3509 This function returns the sqrt of the specified operand if it is a positive
3510 floating point number.
3514 <!-- ======================================================================= -->
3515 <div class="doc_subsection">
3516 <a name="int_manip">Bit Manipulation Intrinsics</a>
3519 <div class="doc_text">
3521 LLVM provides intrinsics for a few important bit manipulation operations.
3522 These allow efficient code generation for some algorithms.
3527 <!-- _______________________________________________________________________ -->
3528 <div class="doc_subsubsection">
3529 <a name="i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
3532 <div class="doc_text">
3536 declare ushort %llvm.bswap.i16(ushort <id>)
3537 declare uint %llvm.bswap.i32(uint <id>)
3538 declare ulong %llvm.bswap.i64(ulong <id>)
3544 The '<tt>llvm.bwsap</tt>' family of intrinsics is used to byteswap a 16, 32 or
3545 64 bit quantity. These are useful for performing operations on data that is not
3546 in the target's native byte order.
3552 The <tt>llvm.bswap.16</tt> intrinsic returns a ushort value that has the high and low
3553 byte of the input ushort swapped. Similarly, the <tt>llvm.bswap.i32</tt> intrinsic
3554 returns a uint value that has the four bytes of the input uint swapped, so that
3555 if the input bytes are numbered 0, 1, 2, 3 then the returned uint will have its
3556 bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i64</tt> intrinsic extends this concept
3562 <!-- _______________________________________________________________________ -->
3563 <div class="doc_subsubsection">
3564 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
3567 <div class="doc_text">
3571 declare ubyte %llvm.ctpop.i8 (ubyte <src>)
3572 declare ushort %llvm.ctpop.i16(ushort <src>)
3573 declare uint %llvm.ctpop.i32(uint <src>)
3574 declare ulong %llvm.ctpop.i64(ulong <src>)
3580 The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
3587 The only argument is the value to be counted. The argument may be of any
3588 unsigned integer type. The return type must match the argument type.
3594 The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
3598 <!-- _______________________________________________________________________ -->
3599 <div class="doc_subsubsection">
3600 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
3603 <div class="doc_text">
3607 declare ubyte %llvm.ctlz.i8 (ubyte <src>)
3608 declare ushort %llvm.ctlz.i16(ushort <src>)
3609 declare uint %llvm.ctlz.i32(uint <src>)
3610 declare ulong %llvm.ctlz.i64(ulong <src>)
3616 The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
3617 leading zeros in a variable.
3623 The only argument is the value to be counted. The argument may be of any
3624 unsigned integer type. The return type must match the argument type.
3630 The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
3631 in a variable. If the src == 0 then the result is the size in bits of the type
3632 of src. For example, <tt>llvm.cttz(int 2) = 30</tt>.
3638 <!-- _______________________________________________________________________ -->
3639 <div class="doc_subsubsection">
3640 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
3643 <div class="doc_text">
3647 declare ubyte %llvm.cttz.i8 (ubyte <src>)
3648 declare ushort %llvm.cttz.i16(ushort <src>)
3649 declare uint %llvm.cttz.i32(uint <src>)
3650 declare ulong %llvm.cttz.i64(ulong <src>)
3656 The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
3663 The only argument is the value to be counted. The argument may be of any
3664 unsigned integer type. The return type must match the argument type.
3670 The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
3671 in a variable. If the src == 0 then the result is the size in bits of the type
3672 of src. For example, <tt>llvm.cttz(2) = 1</tt>.
3676 <!-- ======================================================================= -->
3677 <div class="doc_subsection">
3678 <a name="int_debugger">Debugger Intrinsics</a>
3681 <div class="doc_text">
3683 The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
3684 are described in the <a
3685 href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
3686 Debugging</a> document.
3691 <!-- *********************************************************************** -->
3694 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3695 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3696 <a href="http://validator.w3.org/check/referer"><img
3697 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3699 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
3700 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
3701 Last modified: $Date$