1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
2 <html><head><title>LLVM Assembly Language Reference Manual</title></head>
5 <table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
6 <tr><td> <font size=+5 color="#EEEEFF" face="Georgia,Palatino,Times,Roman"><b>LLVM Language Reference Manual</b></font></td>
10 <li><a href="#abstract">Abstract</a>
11 <li><a href="#introduction">Introduction</a>
12 <li><a href="#identifiers">Identifiers</a>
13 <li><a href="#typesystem">Type System</a>
15 <li><a href="#t_primitive">Primitive Types</a>
17 <li><a href="#t_classifications">Type Classifications</a>
19 <li><a href="#t_derived">Derived Types</a>
21 <li><a href="#t_array" >Array Type</a>
22 <li><a href="#t_function">Function Type</a>
23 <li><a href="#t_pointer">Pointer Type</a>
24 <li><a href="#t_struct" >Structure Type</a>
25 <!-- <li><a href="#t_packed" >Packed Type</a> -->
28 <li><a href="#highlevel">High Level Structure</a>
30 <li><a href="#modulestructure">Module Structure</a>
31 <li><a href="#globalvars">Global Variables</a>
32 <li><a href="#functionstructure">Function Structure</a>
34 <li><a href="#instref">Instruction Reference</a>
36 <li><a href="#terminators">Terminator Instructions</a>
38 <li><a href="#i_ret" >'<tt>ret</tt>' Instruction</a>
39 <li><a href="#i_br" >'<tt>br</tt>' Instruction</a>
40 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a>
41 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a>
42 <li><a href="#i_unwind" >'<tt>unwind</tt>' Instruction</a>
44 <li><a href="#binaryops">Binary Operations</a>
46 <li><a href="#i_add" >'<tt>add</tt>' Instruction</a>
47 <li><a href="#i_sub" >'<tt>sub</tt>' Instruction</a>
48 <li><a href="#i_mul" >'<tt>mul</tt>' Instruction</a>
49 <li><a href="#i_div" >'<tt>div</tt>' Instruction</a>
50 <li><a href="#i_rem" >'<tt>rem</tt>' Instruction</a>
51 <li><a href="#i_setcc">'<tt>set<i>cc</i></tt>' Instructions</a>
53 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
55 <li><a href="#i_and">'<tt>and</tt>' Instruction</a>
56 <li><a href="#i_or" >'<tt>or</tt>' Instruction</a>
57 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a>
58 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a>
59 <li><a href="#i_shr">'<tt>shr</tt>' Instruction</a>
61 <li><a href="#memoryops">Memory Access Operations</a>
63 <li><a href="#i_malloc" >'<tt>malloc</tt>' Instruction</a>
64 <li><a href="#i_free" >'<tt>free</tt>' Instruction</a>
65 <li><a href="#i_alloca" >'<tt>alloca</tt>' Instruction</a>
66 <li><a href="#i_load" >'<tt>load</tt>' Instruction</a>
67 <li><a href="#i_store" >'<tt>store</tt>' Instruction</a>
68 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
70 <li><a href="#otherops">Other Operations</a>
72 <li><a href="#i_phi" >'<tt>phi</tt>' Instruction</a>
73 <li><a href="#i_cast">'<tt>cast .. to</tt>' Instruction</a>
74 <li><a href="#i_call" >'<tt>call</tt>' Instruction</a>
75 <li><a href="#i_vanext">'<tt>vanext</tt>' Instruction</a>
76 <li><a href="#i_vaarg" >'<tt>vaarg</tt>' Instruction</a>
79 <li><a href="#intrinsics">Intrinsic Functions</a>
81 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
83 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
84 <li><a href="#i_va_end" >'<tt>llvm.va_end</tt>' Intrinsic</a>
85 <li><a href="#i_va_copy" >'<tt>llvm.va_copy</tt>' Intrinsic</a>
89 <p><b>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a> and <A href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></b><p>
95 <!-- *********************************************************************** -->
96 <p><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
97 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
98 <a name="abstract">Abstract
99 </b></font></td></tr></table><ul>
100 <!-- *********************************************************************** -->
103 This document is a reference manual for the LLVM assembly language. LLVM is
104 an SSA based representation that provides type safety, low-level operations,
105 flexibility, and the capability of representing 'all' high-level languages
106 cleanly. It is the common code representation used throughout all phases of
107 the LLVM compilation strategy.
113 <!-- *********************************************************************** -->
114 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
115 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
116 <a name="introduction">Introduction
117 </b></font></td></tr></table><ul>
118 <!-- *********************************************************************** -->
120 The LLVM code representation is designed to be used in three different forms: as
121 an in-memory compiler IR, as an on-disk bytecode representation (suitable for
122 fast loading by a Just-In-Time compiler), and as a human readable assembly
123 language representation. This allows LLVM to provide a powerful intermediate
124 representation for efficient compiler transformations and analysis, while
125 providing a natural means to debug and visualize the transformations. The three
126 different forms of LLVM are all equivalent. This document describes the human
127 readable representation and notation.<p>
129 The LLVM representation aims to be a light-weight and low-level while being
130 expressive, typed, and extensible at the same time. It aims to be a "universal
131 IR" of sorts, by being at a low enough level that high-level ideas may be
132 cleanly mapped to it (similar to how microprocessors are "universal IR's",
133 allowing many source languages to be mapped to them). By providing type
134 information, LLVM can be used as the target of optimizations: for example,
135 through pointer analysis, it can be proven that a C automatic variable is never
136 accessed outside of the current function... allowing it to be promoted to a
137 simple SSA value instead of a memory location.<p>
139 <!-- _______________________________________________________________________ -->
140 </ul><a name="wellformed"><h4><hr size=0>Well Formedness</h4><ul>
142 It is important to note that this document describes 'well formed' LLVM assembly
143 language. There is a difference between what the parser accepts and what is
144 considered 'well formed'. For example, the following instruction is
145 syntactically okay, but not well formed:<p>
148 %x = <a href="#i_add">add</a> int 1, %x
151 ...because the definition of <tt>%x</tt> does not dominate all of its uses. The
152 LLVM infrastructure provides a verification pass that may be used to verify that
153 an LLVM module is well formed. This pass is automatically run by the parser
154 after parsing input assembly, and by the optimizer before it outputs bytecode.
155 The violations pointed out by the verifier pass indicate bugs in transformation
156 passes or input to the parser.<p>
158 <!-- Describe the typesetting conventions here. -->
161 <!-- *********************************************************************** -->
162 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
163 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
164 <a name="identifiers">Identifiers
165 </b></font></td></tr></table><ul>
166 <!-- *********************************************************************** -->
168 LLVM uses three different forms of identifiers, for different purposes:<p>
171 <li>Numeric constants are represented as you would expect: 12, -3 123.421, etc.
172 Floating point constants have an optional hexidecimal notation.
174 <li>Named values are represented as a string of characters with a '%' prefix.
175 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
176 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers
177 which require other characters in their names can be surrounded with quotes. In
178 this way, anything except a <tt>"</tt> character can be used in a name.
180 <li>Unnamed values are represented as an unsigned numeric value with a '%'
181 prefix. For example, %12, %2, %44.
184 LLVM requires the values start with a '%' sign for two reasons: Compilers don't
185 need to worry about name clashes with reserved words, and the set of reserved
186 words may be expanded in the future without penalty. Additionally, unnamed
187 identifiers allow a compiler to quickly come up with a temporary variable
188 without having to avoid symbol table conflicts.<p>
190 Reserved words in LLVM are very similar to reserved words in other languages.
191 There are keywords for different opcodes ('<tt><a href="#i_add">add</a></tt>',
192 '<tt><a href="#i_cast">cast</a></tt>', '<tt><a href="#i_ret">ret</a></tt>',
193 etc...), for primitive type names ('<tt><a href="#t_void">void</a></tt>',
194 '<tt><a href="#t_uint">uint</a></tt>', etc...), and others. These reserved
195 words cannot conflict with variable names, because none of them start with a '%'
198 Here is an example of LLVM code to multiply the integer variable '<tt>%X</tt>'
203 %result = <a href="#i_mul">mul</a> uint %X, 8
206 After strength reduction:
208 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
213 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
214 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
215 %result = <a href="#i_add">add</a> uint %1, %1
218 This last way of multiplying <tt>%X</tt> by 8 illustrates several important lexical features of LLVM:<p>
221 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of line.
222 <li>Unnamed temporaries are created when the result of a computation is not
223 assigned to a named value.
224 <li>Unnamed temporaries are numbered sequentially
227 ...and it also show a convention that we follow in this document. When
228 demonstrating instructions, we will follow an instruction with a comment that
229 defines the type and name of value produced. Comments are shown in italic
232 The one non-intuitive notation for constants is the optional hexidecimal form of
233 floating point constants. For example, the form '<tt>double
234 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
235 4.5e+15</tt>' which is also supported by the parser. The only time hexadecimal
236 floating point constants are useful (and the only time that they are generated
237 by the disassembler) is when an FP constant has to be emitted that is not
238 representable as a decimal floating point number exactly. For example, NaN's,
239 infinities, and other special cases are represented in their IEEE hexadecimal
240 format so that assembly and disassembly do not cause any bits to change in the
244 <!-- *********************************************************************** -->
245 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
246 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
247 <a name="typesystem">Type System
248 </b></font></td></tr></table><ul>
249 <!-- *********************************************************************** -->
251 The LLVM type system is one of the most important features of the intermediate
252 representation. Being typed enables a number of optimizations to be performed
253 on the IR directly, without having to do extra analyses on the side before the
254 transformation. A strong type system makes it easier to read the generated code
255 and enables novel analyses and transformations that are not feasible to perform
256 on normal three address code representations.<p>
258 <!-- The written form for the type system was heavily influenced by the
259 syntactic problems with types in the C language<sup><a
260 href="#rw_stroustrup">1</a></sup>.<p> -->
264 <!-- ======================================================================= -->
265 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
266 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
267 <a name="t_primitive">Primitive Types
268 </b></font></td></tr></table><ul>
270 The primitive types are the fundemental building blocks of the LLVM system. The
271 current set of primitive types are as follows:<p>
273 <table border=0 align=center><tr><td>
275 <table border=1 cellspacing=0 cellpadding=4 align=center>
276 <tr><td><tt>void</tt></td> <td>No value</td></tr>
277 <tr><td><tt>ubyte</tt></td> <td>Unsigned 8 bit value</td></tr>
278 <tr><td><tt>ushort</tt></td><td>Unsigned 16 bit value</td></tr>
279 <tr><td><tt>uint</tt></td> <td>Unsigned 32 bit value</td></tr>
280 <tr><td><tt>ulong</tt></td> <td>Unsigned 64 bit value</td></tr>
281 <tr><td><tt>float</tt></td> <td>32 bit floating point value</td></tr>
282 <tr><td><tt>label</tt></td> <td>Branch destination</td></tr>
287 <table border=1 cellspacing=0 cellpadding=4 align=center>
288 <tr><td><tt>bool</tt></td> <td>True or False value</td></tr>
289 <tr><td><tt>sbyte</tt></td> <td>Signed 8 bit value</td></tr>
290 <tr><td><tt>short</tt></td> <td>Signed 16 bit value</td></tr>
291 <tr><td><tt>int</tt></td> <td>Signed 32 bit value</td></tr>
292 <tr><td><tt>long</tt></td> <td>Signed 64 bit value</td></tr>
293 <tr><td><tt>double</tt></td><td>64 bit floating point value</td></tr>
296 </td></tr></table><p>
300 <!-- _______________________________________________________________________ -->
301 </ul><a name="t_classifications"><h4><hr size=0>Type Classifications</h4><ul>
303 These different primitive types fall into a few useful classifications:<p>
305 <table border=1 cellspacing=0 cellpadding=4 align=center>
306 <tr><td><a name="t_signed">signed</td> <td><tt>sbyte, short, int, long, float, double</tt></td></tr>
307 <tr><td><a name="t_unsigned">unsigned</td><td><tt>ubyte, ushort, uint, ulong</tt></td></tr>
308 <tr><td><a name="t_integer">integer</td><td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td></tr>
309 <tr><td><a name="t_integral">integral</td><td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td></tr>
310 <tr><td><a name="t_floating">floating point</td><td><tt>float, double</tt></td></tr>
311 <tr><td><a name="t_firstclass">first class</td><td><tt>bool, ubyte, sbyte, ushort, short,<br> uint, int, ulong, long, float, double, <a href="#t_pointer">pointer</a></tt></td></tr>
314 The <a href="#t_firstclass">first class</a> types are perhaps the most
315 important. Values of these types are the only ones which can be produced by
316 instructions, passed as arguments, or used as operands to instructions. This
317 means that all structures and arrays must be manipulated either by pointer or by
321 <!-- ======================================================================= -->
322 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0><tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
323 <a name="t_derived">Derived Types
324 </b></font></td></tr></table><ul>
326 The real power in LLVM comes from the derived types in the system. This is what
327 allows a programmer to represent arrays, functions, pointers, and other useful
328 types. Note that these derived types may be recursive: For example, it is
329 possible to have a two dimensional array.<p>
333 <!-- _______________________________________________________________________ -->
334 </ul><a name="t_array"><h4><hr size=0>Array Type</h4><ul>
338 The array type is a very simple derived type that arranges elements sequentially
339 in memory. The array type requires a size (number of elements) and an
340 underlying data type.<p>
344 [<# elements> x <elementtype>]
347 The number of elements is a constant integer value, elementtype may be any type
352 <tt>[40 x int ]</tt>: Array of 40 integer values.<br>
353 <tt>[41 x int ]</tt>: Array of 41 integer values.<br>
354 <tt>[40 x uint]</tt>: Array of 40 unsigned integer values.<p>
357 Here are some examples of multidimensional arrays:<p>
359 <table border=0 cellpadding=0 cellspacing=0>
360 <tr><td><tt>[3 x [4 x int]]</tt></td><td>: 3x4 array integer values.</td></tr>
361 <tr><td><tt>[12 x [10 x float]]</tt></td><td>: 12x10 array of single precision floating point values.</td></tr>
362 <tr><td><tt>[2 x [3 x [4 x uint]]]</tt></td><td>: 2x3x4 array of unsigned integer values.</td></tr>
367 <!-- _______________________________________________________________________ -->
368 </ul><a name="t_function"><h4><hr size=0>Function Type</h4><ul>
372 The function type can be thought of as a function signature. It consists of a
373 return type and a list of formal parameter types. Function types are usually
374 used when to build virtual function tables (which are structures of pointers to
375 functions), for indirect function calls, and when defining a function.<p>
379 <returntype> (<parameter list>)
382 Where '<tt><parameter list></tt>' is a comma-separated list of type
383 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
384 which indicates that the function takes a variable number of arguments.
385 Variable argument functions can access their arguments with the <a
386 href="#int_varargs">variable argument handling intrinsic</a> functions.
391 <table border=0 cellpadding=0 cellspacing=0>
393 <tr><td><tt>int (int)</tt></td><td>: function taking an <tt>int</tt>, returning
394 an <tt>int</tt></td></tr>
396 <tr><td><tt>float (int, int *) *</tt></td><td>: <a href="#t_pointer">Pointer</a>
397 to a function that takes an <tt>int</tt> and a <a href="#t_pointer">pointer</a>
398 to <tt>int</tt>, returning <tt>float</tt>.</td></tr>
400 <tr><td><tt>int (sbyte *, ...)</tt></td><td>: A vararg function that takes at
401 least one <a href="#t_pointer">pointer</a> to <tt>sbyte</tt> (signed char in C),
402 which returns an integer. This is the signature for <tt>printf</tt> in
410 <!-- _______________________________________________________________________ -->
411 </ul><a name="t_struct"><h4><hr size=0>Structure Type</h4><ul>
415 The structure type is used to represent a collection of data members together in
416 memory. The packing of the field types is defined to match the ABI of the
417 underlying processor. The elements of a structure may be any type that has a
420 Structures are accessed using '<tt><a href="#i_load">load</a></tt> and '<tt><a
421 href="#i_store">store</a></tt>' by getting a pointer to a field with the '<tt><a
422 href="#i_getelementptr">getelementptr</a></tt>' instruction.<p>
426 { <type list> }
431 <table border=0 cellpadding=0 cellspacing=0>
433 <tr><td><tt>{ int, int, int }</tt></td><td>: a triple of three <tt>int</tt>
436 <tr><td><tt>{ float, int (int) * }</tt></td><td>: A pair, where the first
437 element is a <tt>float</tt> and the second element is a <a
438 href="#t_pointer">pointer</a> to a <a href="t_function">function</a> that takes
439 an <tt>int</tt>, returning an <tt>int</tt>.</td></tr>
444 <!-- _______________________________________________________________________ -->
445 </ul><a name="t_pointer"><h4><hr size=0>Pointer Type</h4><ul>
449 As in many languages, the pointer type represents a pointer or reference to
450 another object, which must live in memory.<p>
459 <table border=0 cellpadding=0 cellspacing=0>
461 <tr><td><tt>[4x int]*</tt></td><td>: <a href="#t_pointer">pointer</a> to <a
462 href="#t_array">array</a> of four <tt>int</tt> values</td></tr>
464 <tr><td><tt>int (int *) *</tt></td><td>: A <a href="#t_pointer">pointer</a> to a
465 <a href="t_function">function</a> that takes an <tt>int</tt>, returning an
466 <tt>int</tt>.</td></tr>
472 <!-- _______________________________________________________________________ -->
474 </ul><a name="t_packed"><h4><hr size=0>Packed Type</h4><ul>
476 Mention/decide that packed types work with saturation or not. Maybe have a packed+saturated type in addition to just a packed type.<p>
478 Packed types should be 'nonsaturated' because standard data types are not saturated. Maybe have a saturated packed type?<p>
483 <!-- *********************************************************************** -->
484 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
485 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
486 <a name="highlevel">High Level Structure
487 </b></font></td></tr></table><ul>
488 <!-- *********************************************************************** -->
491 <!-- ======================================================================= -->
492 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
493 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
494 <a name="modulestructure">Module Structure
495 </b></font></td></tr></table><ul>
497 LLVM programs are composed of "Module"s, each of which is a translation unit of
498 the input programs. Each module consists of functions, global variables, and
499 symbol table entries. Modules may be combined together with the LLVM linker,
500 which merges function (and global variable) definitions, resolves forward
501 declarations, and merges symbol table entries. Here is an example of the "hello world" module:<p>
504 <i>; Declare the string constant as a global constant...</i>
505 <a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
507 <i>; External declaration of the puts function</i>
508 <a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
510 <i>; Definition of main function</i>
511 int %main() { <i>; int()* </i>
512 <i>; Convert [13x sbyte]* to sbyte *...</i>
513 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
515 <i>; Call puts function to write out the string to stdout...</i>
516 <a href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
517 <a href="#i_ret">ret</a> int 0
521 This example is made up of a <a href="#globalvars">global variable</a> named
522 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function, and a
523 <a href="#functionstructure">function definition</a> for "<tt>main</tt>".<p>
526 In general, a module is made up of a list of global values, where both functions
527 and global variables are global values. Global values are represented by a
528 pointer to a memory location (in this case, a pointer to an array of char, and a
529 pointer to a function), and have one of the following linkage types:<p>
532 <a name="linkage_internal">
533 <dt><tt><b>internal</b></tt>
535 <dd>Global values with internal linkage are only directly accessible by objects
536 in the current module. In particular, linking code into a module with an
537 internal global value may cause the internal to be renamed as necessary to avoid
538 collisions. Because the symbol is internal to the module, all references can be
539 updated. This corresponds to the notion of the '<tt>static</tt>' keyword in C,
540 or the idea of "anonymous namespaces" in C++.<p>
542 <a name="linkage_linkonce">
543 <dt><tt><b>linkonce</b></tt>:
545 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
546 the twist that linking together two modules defining the same <tt>linkonce</tt>
547 globals will cause one of the globals to be discarded. This is typically used
548 to implement inline functions. Unreferenced <tt>linkonce</tt> globals are
549 allowed to be discarded.<p>
551 <a name="linkage_weak">
552 <dt><tt><b>weak</b></tt>:
554 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
555 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
556 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.<p>
558 <a name="linkage_appending">
559 <dt><tt><b>appending</b></tt>:
561 <dd>"<tt>appending</tt>" linkage may only applied to global variables of pointer
562 to array type. When two global variables with appending linkage are linked
563 together, the two global arrays are appended together. This is the LLVM,
564 typesafe, equivalent of having the system linker append together "sections" with
565 identical names when .o files are linked.<p>
567 <a name="linkage_external">
568 <dt><tt><b>externally visible</b></tt>:
570 <dd>If none of the above identifiers are used, the global is externally visible,
571 meaning that it participates in linkage and can be used to resolve external
572 symbol references.<p>
577 For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
578 another module defined a "<tt>.LC0</tt>" variable and was linked with this one,
579 one of the two would be renamed, preventing a collision. Since "<tt>main</tt>"
580 and "<tt>puts</tt>" are external (i.e., lacking any linkage declarations), they
581 are accessible outside of the current module. It is illegal for a function
582 <i>declaration</i> to have any linkage type other than "externally visible".<p>
585 <!-- ======================================================================= -->
586 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
587 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
588 <a name="globalvars">Global Variables
589 </b></font></td></tr></table><ul>
591 Global variables define regions of memory allocated at compilation time instead
592 of run-time. Global variables may optionally be initialized. A variable may
593 be defined as a global "constant", which indicates that the contents of the
594 variable will never be modified (opening options for optimization). Constants
595 must always have an initial value.<p>
597 As SSA values, global variables define pointer values that are in scope
598 (i.e. they dominate) for all basic blocks in the program. Global variables
599 always define a pointer to their "content" type because they describe a region
600 of memory, and all memory objects in LLVM are accessed through pointers.<p>
604 <!-- ======================================================================= -->
605 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
606 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
607 <a name="functionstructure">Functions
608 </b></font></td></tr></table><ul>
610 LLVM functions definitions are composed of a (possibly empty) argument list, an
611 opening curly brace, a list of basic blocks, and a closing curly brace. LLVM
612 function declarations are defined with the "<tt>declare</tt>" keyword, a
613 function name and a function signature.<p>
615 A function definition contains a list of basic blocks, forming the CFG for the
616 function. Each basic block may optionally start with a label (giving the basic
617 block a symbol table entry), contains a list of instructions, and ends with a <a
618 href="#terminators">terminator</a> instruction (such as a branch or function
621 The first basic block in program is special in two ways: it is immediately
622 executed on entrance to the function, and it is not allowed to have predecessor
623 basic blocks (i.e. there can not be any branches to the entry block of a
624 function). Because the block can have no predecessors, it also cannot have any
625 <a href="#i_phi">PHI nodes</a>.<p>
628 <!-- *********************************************************************** -->
629 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
630 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
631 <a name="instref">Instruction Reference
632 </b></font></td></tr></table><ul>
633 <!-- *********************************************************************** -->
635 The LLVM instruction set consists of several different classifications of
636 instructions: <a href="#terminators">terminator instructions</a>, <a
637 href="#binaryops">binary instructions</a>, <a href="#memoryops">memory
638 instructions</a>, and <a href="#otherops">other instructions</a>.<p>
641 <!-- ======================================================================= -->
642 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
643 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
644 <a name="terminators">Terminator Instructions
645 </b></font></td></tr></table><ul>
647 As mentioned <a href="#functionstructure">previously</a>, every basic block in a
648 program ends with a "Terminator" instruction, which indicates which block should
649 be executed after the current block is finished. These terminator instructions
650 typically yield a '<tt>void</tt>' value: they produce control flow, not values
651 (the one exception being the '<a href="#i_invoke"><tt>invoke</tt></a>'
654 There are five different terminator instructions: the '<a
655 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a
656 href="#i_br"><tt>br</tt></a>' instruction, the '<a
657 href="#i_switch"><tt>switch</tt></a>' instruction, the '<a
658 href="#i_invoke"><tt>invoke</tt></a>' instruction, and the '<a
659 href="#i_unwind"><tt>unwind</tt></a>' instruction.<p>
662 <!-- _______________________________________________________________________ -->
663 </ul><a name="i_ret"><h4><hr size=0>'<tt>ret</tt>' Instruction</h4><ul>
667 ret <type> <value> <i>; Return a value from a non-void function</i>
668 ret void <i>; Return from void function</i>
673 The '<tt>ret</tt>' instruction is used to return control flow (and a value) from
674 a function, back to the caller.<p>
676 There are two forms of the '<tt>ret</tt>' instructruction: one that returns a
677 value and then causes control flow, and one that just causes control flow to
682 The '<tt>ret</tt>' instruction may return any '<a href="#t_firstclass">first
683 class</a>' type. Notice that a function is not <a href="#wellformed">well
684 formed</a> if there exists a '<tt>ret</tt>' instruction inside of the function
685 that returns a value that does not match the return type of the function.<p>
689 When the '<tt>ret</tt>' instruction is executed, control flow returns back to
690 the calling function's context. If the caller is a "<a
691 href="#i_call"><tt>call</tt></a> instruction, execution continues at the
692 instruction after the call. If the caller was an "<a
693 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at the
694 beginning "normal" of the destination block. If the instruction returns a
695 value, that value shall set the call or invoke instruction's return value.<p>
700 ret int 5 <i>; Return an integer value of 5</i>
701 ret void <i>; Return from a void function</i>
705 <!-- _______________________________________________________________________ -->
706 </ul><a name="i_br"><h4><hr size=0>'<tt>br</tt>' Instruction</h4><ul>
710 br bool <cond>, label <iftrue>, label <iffalse>
711 br label <dest> <i>; Unconditional branch</i>
716 The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
717 different basic block in the current function. There are two forms of this
718 instruction, corresponding to a conditional branch and an unconditional
723 The conditional branch form of the '<tt>br</tt>' instruction takes a single
724 '<tt>bool</tt>' value and two '<tt>label</tt>' values. The unconditional form
725 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
730 Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
731 argument is evaluated. If the value is <tt>true</tt>, control flows to the
732 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
733 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.<p>
738 %cond = <a href="#i_setcc">seteq</a> int %a, %b
739 br bool %cond, label %IfEqual, label %IfUnequal
741 <a href="#i_ret">ret</a> int 1
743 <a href="#i_ret">ret</a> int 0
747 <!-- _______________________________________________________________________ -->
748 </ul><a name="i_switch"><h4><hr size=0>'<tt>switch</tt>' Instruction</h4><ul>
752 switch uint <value>, label <defaultdest> [ int <val>, label &dest>, ... ]
758 The '<tt>switch</tt>' instruction is used to transfer control flow to one of
759 several different places. It is a generalization of the '<tt>br</tt>'
760 instruction, allowing a branch to occur to one of many possible destinations.<p>
764 The '<tt>switch</tt>' instruction uses three parameters: a '<tt>uint</tt>'
765 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
766 an array of pairs of comparison value constants and '<tt>label</tt>'s.<p>
770 The <tt>switch</tt> instruction specifies a table of values and destinations.
771 When the '<tt>switch</tt>' instruction is executed, this table is searched for
772 the given value. If the value is found, the corresponding destination is
773 branched to, otherwise the default value it transfered to.<p>
775 <h5>Implementation:</h5>
777 Depending on properties of the target machine and the particular <tt>switch</tt>
778 instruction, this instruction may be code generated as a series of chained
779 conditional branches, or with a lookup table.<p>
783 <i>; Emulate a conditional br instruction</i>
784 %Val = <a href="#i_cast">cast</a> bool %value to uint
785 switch uint %Val, label %truedest [int 0, label %falsedest ]
787 <i>; Emulate an unconditional br instruction</i>
788 switch uint 0, label %dest [ ]
790 <i>; Implement a jump table:</i>
791 switch uint %val, label %otherwise [ int 0, label %onzero,
793 int 2, label %ontwo ]
798 <!-- _______________________________________________________________________ -->
799 </ul><a name="i_invoke"><h4><hr size=0>'<tt>invoke</tt>' Instruction</h4><ul>
803 <result> = invoke <ptr to function ty> %<function ptr val>(<function args>)
804 to label <normal label> except label <exception label>
809 The '<tt>invoke</tt>' instruction causes control to transfer to a specified
810 function, with the possibility of control flow transfer to either the
811 '<tt>normal</tt>' <tt>label</tt> label or the '<tt>exception</tt>'
812 <tt>label</tt>. If the callee function returns with the "<tt><a
813 href="#i_ret">ret</a></tt>" instruction, control flow will return to the
814 "normal" label. If the callee (or any indirect callees) returns with the "<a
815 href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted, and
816 continued at the dynamically nearest "except" label.<p>
821 This instruction requires several arguments:<p>
824 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
825 function value being invoked. In most cases, this is a direct function
826 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
827 an arbitrary pointer to function value.
829 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
830 function to be invoked.
832 <li>'<tt>function args</tt>': argument list whose types match the function
833 signature argument types. If the function signature indicates the function
834 accepts a variable number of arguments, the extra arguments can be specified.
836 <li>'<tt>normal label</tt>': the label reached when the called function executes
837 a '<tt><a href="#i_ret">ret</a></tt>' instruction.
839 <li>'<tt>exception label</tt>': the label reached when a callee returns with the
840 <a href="#i_unwind"><tt>unwind</tt></a> instruction.
845 This instruction is designed to operate as a standard '<tt><a
846 href="#i_call">call</a></tt>' instruction in most regards. The primary
847 difference is that it establishes an association with a label, which is used by the runtime library to unwind the stack.<p>
849 This instruction is used in languages with destructors to ensure that proper
850 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
851 exception. Additionally, this is important for implementation of
852 '<tt>catch</tt>' clauses in high-level languages that support them.<p>
856 %retval = invoke int %Test(int 15)
858 except label %TestCleanup <i>; {int}:retval set</i>
861 <!-- _______________________________________________________________________ -->
862 </ul><a name="i_unwind"><h4><hr size=0>'<tt>unwind</tt>' Instruction</h4><ul>
871 The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow at
872 the first callee in the dynamic call stack which used an <a
873 href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
874 primarily used to implement exception handling.
878 The '<tt>unwind</tt>' intrinsic causes execution of the current function to
879 immediately halt. The dynamic call stack is then searched for the first <a
880 href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
881 execution continues at the "exceptional" destination block specified by the
882 <tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
883 dynamic call chain, undefined behavior results.
887 <!-- ======================================================================= -->
888 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0><tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
889 <a name="binaryops">Binary Operations
890 </b></font></td></tr></table><ul>
892 Binary operators are used to do most of the computation in a program. They
893 require two operands, execute an operation on them, and produce a single value.
894 The result value of a binary operator is not necessarily the same type as its
897 There are several different binary operators:<p>
900 <!-- _______________________________________________________________________ -->
901 </ul><a name="i_add"><h4><hr size=0>'<tt>add</tt>' Instruction</h4><ul>
905 <result> = add <ty> <var1>, <var2> <i>; yields {ty}:result</i>
909 The '<tt>add</tt>' instruction returns the sum of its two operands.<p>
912 The two arguments to the '<tt>add</tt>' instruction must be either <a href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values. Both arguments must have identical types.<p>
916 The value produced is the integer or floating point sum of the two operands.<p>
920 <result> = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
924 <!-- _______________________________________________________________________ -->
925 </ul><a name="i_sub"><h4><hr size=0>'<tt>sub</tt>' Instruction</h4><ul>
929 <result> = sub <ty> <var1>, <var2> <i>; yields {ty}:result</i>
934 The '<tt>sub</tt>' instruction returns the difference of its two operands.<p>
936 Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
937 instruction present in most other intermediate representations.<p>
941 The two arguments to the '<tt>sub</tt>' instruction must be either <a
942 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
943 values. Both arguments must have identical types.<p>
947 The value produced is the integer or floating point difference of the two
952 <result> = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
953 <result> = sub int 0, %val <i>; yields {int}:result = -%var</i>
956 <!-- _______________________________________________________________________ -->
957 </ul><a name="i_mul"><h4><hr size=0>'<tt>mul</tt>' Instruction</h4><ul>
961 <result> = mul <ty> <var1>, <var2> <i>; yields {ty}:result</i>
965 The '<tt>mul</tt>' instruction returns the product of its two operands.<p>
968 The two arguments to the '<tt>mul</tt>' instruction must be either <a href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values. Both arguments must have identical types.<p>
972 The value produced is the integer or floating point product of the two
975 There is no signed vs unsigned multiplication. The appropriate action is taken
976 based on the type of the operand. <p>
981 <result> = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
985 <!-- _______________________________________________________________________ -->
986 </ul><a name="i_div"><h4><hr size=0>'<tt>div</tt>' Instruction</h4><ul>
990 <result> = div <ty> <var1>, <var2> <i>; yields {ty}:result</i>
995 The '<tt>div</tt>' instruction returns the quotient of its two operands.<p>
999 The two arguments to the '<tt>div</tt>' instruction must be either <a
1000 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
1001 values. Both arguments must have identical types.<p>
1005 The value produced is the integer or floating point quotient of the two
1010 <result> = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
1014 <!-- _______________________________________________________________________ -->
1015 </ul><a name="i_rem"><h4><hr size=0>'<tt>rem</tt>' Instruction</h4><ul>
1019 <result> = rem <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1023 The '<tt>rem</tt>' instruction returns the remainder from the division of its two operands.<p>
1026 The two arguments to the '<tt>rem</tt>' instruction must be either <a href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values. Both arguments must have identical types.<p>
1030 This returns the <i>remainder</i> of a division (where the result has the same
1031 sign as the divisor), not the <i>modulus</i> (where the result has the same sign
1032 as the dividend) of a value. For more information about the difference, see: <a
1033 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The Math
1038 <result> = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1042 <!-- _______________________________________________________________________ -->
1043 </ul><a name="i_setcc"><h4><hr size=0>'<tt>set<i>cc</i></tt>' Instructions</h4><ul>
1047 <result> = seteq <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1048 <result> = setne <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1049 <result> = setlt <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1050 <result> = setgt <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1051 <result> = setle <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1052 <result> = setge <ty> <var1>, <var2> <i>; yields {bool}:result</i>
1055 <h5>Overview:</h5> The '<tt>set<i>cc</i></tt>' family of instructions returns a
1056 boolean value based on a comparison of their two operands.<p>
1058 <h5>Arguments:</h5> The two arguments to the '<tt>set<i>cc</i></tt>'
1059 instructions must be of <a href="#t_firstclass">first class</a> type (it is not
1060 possible to compare '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>'
1061 or '<tt>void</tt>' values, etc...). Both arguments must have identical
1066 The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1067 both operands are equal.<br>
1069 The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1070 both operands are unequal.<br>
1072 The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1073 the first operand is less than the second operand.<br>
1075 The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1076 the first operand is greater than the second operand.<br>
1078 The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1079 the first operand is less than or equal to the second operand.<br>
1081 The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1082 the first operand is greater than or equal to the second operand.<p>
1086 <result> = seteq int 4, 5 <i>; yields {bool}:result = false</i>
1087 <result> = setne float 4, 5 <i>; yields {bool}:result = true</i>
1088 <result> = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1089 <result> = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1090 <result> = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1091 <result> = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1096 <!-- ======================================================================= -->
1097 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1098 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
1099 <a name="bitwiseops">Bitwise Binary Operations
1100 </b></font></td></tr></table><ul>
1102 Bitwise binary operators are used to do various forms of bit-twiddling in a
1103 program. They are generally very efficient instructions, and can commonly be
1104 strength reduced from other instructions. They require two operands, execute an
1105 operation on them, and produce a single value. The resulting value of the
1106 bitwise binary operators is always the same type as its first operand.<p>
1108 <!-- _______________________________________________________________________ -->
1109 </ul><a name="i_and"><h4><hr size=0>'<tt>and</tt>' Instruction</h4><ul>
1113 <result> = and <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1117 The '<tt>and</tt>' instruction returns the bitwise logical and of its two operands.<p>
1121 The two arguments to the '<tt>and</tt>' instruction must be <a
1122 href="#t_integral">integral</a> values. Both arguments must have identical
1128 The truth table used for the '<tt>and</tt>' instruction is:<p>
1130 <center><table border=1 cellspacing=0 cellpadding=4>
1131 <tr><td>In0</td> <td>In1</td> <td>Out</td></tr>
1132 <tr><td>0</td> <td>0</td> <td>0</td></tr>
1133 <tr><td>0</td> <td>1</td> <td>0</td></tr>
1134 <tr><td>1</td> <td>0</td> <td>0</td></tr>
1135 <tr><td>1</td> <td>1</td> <td>1</td></tr>
1136 </table></center><p>
1141 <result> = and int 4, %var <i>; yields {int}:result = 4 & %var</i>
1142 <result> = and int 15, 40 <i>; yields {int}:result = 8</i>
1143 <result> = and int 4, 8 <i>; yields {int}:result = 0</i>
1148 <!-- _______________________________________________________________________ -->
1149 </ul><a name="i_or"><h4><hr size=0>'<tt>or</tt>' Instruction</h4><ul>
1153 <result> = or <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1156 <h5>Overview:</h5> The '<tt>or</tt>' instruction returns the bitwise logical
1157 inclusive or of its two operands.<p>
1161 The two arguments to the '<tt>or</tt>' instruction must be <a
1162 href="#t_integral">integral</a> values. Both arguments must have identical
1168 The truth table used for the '<tt>or</tt>' instruction is:<p>
1170 <center><table border=1 cellspacing=0 cellpadding=4>
1171 <tr><td>In0</td> <td>In1</td> <td>Out</td></tr>
1172 <tr><td>0</td> <td>0</td> <td>0</td></tr>
1173 <tr><td>0</td> <td>1</td> <td>1</td></tr>
1174 <tr><td>1</td> <td>0</td> <td>1</td></tr>
1175 <tr><td>1</td> <td>1</td> <td>1</td></tr>
1176 </table></center><p>
1181 <result> = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
1182 <result> = or int 15, 40 <i>; yields {int}:result = 47</i>
1183 <result> = or int 4, 8 <i>; yields {int}:result = 12</i>
1187 <!-- _______________________________________________________________________ -->
1188 </ul><a name="i_xor"><h4><hr size=0>'<tt>xor</tt>' Instruction</h4><ul>
1192 <result> = xor <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1197 The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of its
1198 two operands. The <tt>xor</tt> is used to implement the "one's complement"
1199 operation, which is the "~" operator in C.<p>
1203 The two arguments to the '<tt>xor</tt>' instruction must be <a
1204 href="#t_integral">integral</a> values. Both arguments must have identical
1210 The truth table used for the '<tt>xor</tt>' instruction is:<p>
1212 <center><table border=1 cellspacing=0 cellpadding=4>
1213 <tr><td>In0</td> <td>In1</td> <td>Out</td></tr>
1214 <tr><td>0</td> <td>0</td> <td>0</td></tr>
1215 <tr><td>0</td> <td>1</td> <td>1</td></tr>
1216 <tr><td>1</td> <td>0</td> <td>1</td></tr>
1217 <tr><td>1</td> <td>1</td> <td>0</td></tr>
1218 </table></center><p>
1223 <result> = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
1224 <result> = xor int 15, 40 <i>; yields {int}:result = 39</i>
1225 <result> = xor int 4, 8 <i>; yields {int}:result = 12</i>
1226 <result> = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
1230 <!-- _______________________________________________________________________ -->
1231 </ul><a name="i_shl"><h4><hr size=0>'<tt>shl</tt>' Instruction</h4><ul>
1235 <result> = shl <ty> <var1>, ubyte <var2> <i>; yields {ty}:result</i>
1240 The '<tt>shl</tt>' instruction returns the first operand shifted to the left a
1241 specified number of bits.
1245 The first argument to the '<tt>shl</tt>' instruction must be an <a
1246 href="#t_integer">integer</a> type. The second argument must be an
1247 '<tt>ubyte</tt>' type.<p>
1251 The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.<p>
1256 <result> = shl int 4, ubyte %var <i>; yields {int}:result = 4 << %var</i>
1257 <result> = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1258 <result> = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1262 <!-- _______________________________________________________________________ -->
1263 </ul><a name="i_shr"><h4><hr size=0>'<tt>shr</tt>' Instruction</h4><ul>
1268 <result> = shr <ty> <var1>, ubyte <var2> <i>; yields {ty}:result</i>
1272 The '<tt>shr</tt>' instruction returns the first operand shifted to the right a specified number of bits.
1275 The first argument to the '<tt>shr</tt>' instruction must be an <a href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>' type.<p>
1279 If the first argument is a <a href="#t_signed">signed</a> type, the most
1280 significant bit is duplicated in the newly free'd bit positions. If the first
1281 argument is unsigned, zero bits shall fill the empty positions.<p>
1285 <result> = shr int 4, ubyte %var <i>; yields {int}:result = 4 >> %var</i>
1286 <result> = shr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
1287 <result> = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
1288 <result> = shr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
1289 <result> = shr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
1296 <!-- ======================================================================= -->
1297 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1298 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
1299 <a name="memoryops">Memory Access Operations
1300 </b></font></td></tr></table><ul>
1302 A key design point of an SSA-based representation is how it represents memory.
1303 In LLVM, no memory locations are in SSA form, which makes things very simple.
1304 This section describes how to read, write, allocate and free memory in LLVM.<p>
1307 <!-- _______________________________________________________________________ -->
1308 </ul><a name="i_malloc"><h4><hr size=0>'<tt>malloc</tt>' Instruction</h4><ul>
1312 <result> = malloc <type>, uint <NumElements> <i>; yields {type*}:result</i>
1313 <result> = malloc <type> <i>; yields {type*}:result</i>
1317 The '<tt>malloc</tt>' instruction allocates memory from the system heap and returns a pointer to it.<p>
1321 The the '<tt>malloc</tt>' instruction allocates
1322 <tt>sizeof(<type>)*NumElements</tt> bytes of memory from the operating
1323 system, and returns a pointer of the appropriate type to the program. The
1324 second form of the instruction is a shorter version of the first instruction
1325 that defaults to allocating one element.<p>
1327 '<tt>type</tt>' must be a sized type.<p>
1331 Memory is allocated using the system "<tt>malloc</tt>" function, and a pointer
1336 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
1338 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
1339 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
1340 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
1344 <!-- _______________________________________________________________________ -->
1345 </ul><a name="i_free"><h4><hr size=0>'<tt>free</tt>' Instruction</h4><ul>
1349 free <type> <value> <i>; yields {void}</i>
1354 The '<tt>free</tt>' instruction returns memory back to the unused memory heap, to be reallocated in the future.<p>
1359 '<tt>value</tt>' shall be a pointer value that points to a value that was
1360 allocated with the '<tt><a href="#i_malloc">malloc</a></tt>' instruction.<p>
1365 Access to the memory pointed to by the pointer is not longer defined after this instruction executes.<p>
1369 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
1370 free [4 x ubyte]* %array
1374 <!-- _______________________________________________________________________ -->
1375 </ul><a name="i_alloca"><h4><hr size=0>'<tt>alloca</tt>' Instruction</h4><ul>
1379 <result> = alloca <type>, uint <NumElements> <i>; yields {type*}:result</i>
1380 <result> = alloca <type> <i>; yields {type*}:result</i>
1385 The '<tt>alloca</tt>' instruction allocates memory on the current stack frame of
1386 the procedure that is live until the current function returns to its caller.<p>
1390 The the '<tt>alloca</tt>' instruction allocates
1391 <tt>sizeof(<type>)*NumElements</tt> bytes of memory on the runtime stack,
1392 returning a pointer of the appropriate type to the program. The second form of
1393 the instruction is a shorter version of the first that defaults to allocating
1396 '<tt>type</tt>' may be any sized type.<p>
1400 Memory is allocated, a pointer is returned. '<tt>alloca</tt>'d memory is
1401 automatically released when the function returns. The '<tt>alloca</tt>'
1402 instruction is commonly used to represent automatic variables that must have an
1403 address available. When the function returns (either with the <tt><a
1404 href="#i_ret">ret</a></tt> or <tt><a href="#i_invoke">invoke</a></tt>
1405 instructions), the memory is reclaimed.<p>
1409 %ptr = alloca int <i>; yields {int*}:ptr</i>
1410 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
1414 <!-- _______________________________________________________________________ -->
1415 </ul><a name="i_load"><h4><hr size=0>'<tt>load</tt>' Instruction</h4><ul>
1419 <result> = load <ty>* <pointer>
1420 <result> = volatile load <ty>* <pointer>
1424 The '<tt>load</tt>' instruction is used to read from memory.<p>
1428 The argument to the '<tt>load</tt>' instruction specifies the memory address to
1429 load from. The pointer must point to a <a href="t_firstclass">first class</a>
1430 type. If the <tt>load</tt> is marked as <tt>volatile</tt> then the optimizer is
1431 not allowed to modify the number or order of execution of this <tt>load</tt>
1432 with other volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
1437 The location of memory pointed to is loaded.
1441 %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1442 <a href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
1443 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1449 <!-- _______________________________________________________________________ -->
1450 </ul><a name="i_store"><h4><hr size=0>'<tt>store</tt>' Instruction</h4><ul>
1454 store <ty> <value>, <ty>* <pointer> <i>; yields {void}</i>
1455 volatile store <ty> <value>, <ty>* <pointer> <i>; yields {void}</i>
1459 The '<tt>store</tt>' instruction is used to write to memory.<p>
1463 There are two arguments to the '<tt>store</tt>' instruction: a value to store
1464 and an address to store it into. The type of the '<tt><pointer></tt>'
1465 operand must be a pointer to the type of the '<tt><value></tt>' operand.
1466 If the <tt>store</tt> is marked as <tt>volatile</tt> then the optimizer is not
1467 allowed to modify the number or order of execution of this <tt>store</tt> with
1468 other volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
1471 <h5>Semantics:</h5> The contents of memory are updated to contain
1472 '<tt><value></tt>' at the location specified by the
1473 '<tt><pointer></tt>' operand.<p>
1477 %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1478 <a href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
1479 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1485 <!-- _______________________________________________________________________ -->
1486 </ul><a name="i_getelementptr"><h4><hr size=0>'<tt>getelementptr</tt>' Instruction</h4><ul>
1490 <result> = getelementptr <ty>* <ptrval>{, long <aidx>|, ubyte <sidx>}*
1495 The '<tt>getelementptr</tt>' instruction is used to get the address of a
1496 subelement of an aggregate data structure.<p>
1500 This instruction takes a list of <tt>long</tt> values and <tt>ubyte</tt>
1501 constants that indicate what form of addressing to perform. The actual types of
1502 the arguments provided depend on the type of the first pointer argument. The
1503 '<tt>getelementptr</tt>' instruction is used to index down through the type
1504 levels of a structure.<p>
1506 For example, lets consider a C code fragment and how it gets compiled to
1521 int *foo(struct ST *s) {
1522 return &s[1].Z.B[5][13];
1526 The LLVM code generated by the GCC frontend is:
1529 %RT = type { sbyte, [10 x [20 x int]], sbyte }
1530 %ST = type { int, double, %RT }
1532 int* "foo"(%ST* %s) {
1533 %reg = getelementptr %ST* %s, long 1, ubyte 2, ubyte 1, long 5, long 13
1540 The index types specified for the '<tt>getelementptr</tt>' instruction depend on
1541 the pointer type that is being index into. <a href="t_pointer">Pointer</a> and
1542 <a href="t_array">array</a> types require '<tt>long</tt>' values, and <a
1543 href="t_struct">structure</a> types require '<tt>ubyte</tt>'
1544 <b>constants</b>.<p>
1546 In the example above, the first index is indexing into the '<tt>%ST*</tt>' type,
1547 which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT }</tt>'
1548 type, a structure. The second index indexes into the third element of the
1549 structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]], sbyte
1550 }</tt>' type, another structure. The third index indexes into the second
1551 element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
1552 array. The two dimensions of the array are subscripted into, yielding an
1553 '<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction return a pointer
1554 to this element, thus yielding a '<tt>int*</tt>' type.<p>
1556 Note that it is perfectly legal to index partially through a structure,
1557 returning a pointer to an inner element. Because of this, the LLVM code for the
1558 given testcase is equivalent to:<p>
1561 int* "foo"(%ST* %s) {
1562 %t1 = getelementptr %ST* %s , long 1 <i>; yields %ST*:%t1</i>
1563 %t2 = getelementptr %ST* %t1, long 0, ubyte 2 <i>; yields %RT*:%t2</i>
1564 %t3 = getelementptr %RT* %t2, long 0, ubyte 1 <i>; yields [10 x [20 x int]]*:%t3</i>
1565 %t4 = getelementptr [10 x [20 x int]]* %t3, long 0, long 5 <i>; yields [20 x int]*:%t4</i>
1566 %t5 = getelementptr [20 x int]* %t4, long 0, long 13 <i>; yields int*:%t5</i>
1575 <i>; yields [12 x ubyte]*:aptr</i>
1576 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, ubyte 1
1581 <!-- ======================================================================= -->
1582 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1583 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
1584 <a name="otherops">Other Operations
1585 </b></font></td></tr></table><ul>
1587 The instructions in this catagory are the "miscellaneous" instructions, which defy better classification.<p>
1590 <!-- _______________________________________________________________________ -->
1591 </ul><a name="i_phi"><h4><hr size=0>'<tt>phi</tt>' Instruction</h4><ul>
1595 <result> = phi <ty> [ <val0>, <label0>], ...
1600 The '<tt>phi</tt>' instruction is used to implement the φ node in the SSA
1601 graph representing the function.<p>
1605 The type of the incoming values are specified with the first type field. After
1606 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
1607 one pair for each predecessor basic block of the current block. Only values of
1608 <a href="#t_firstclass">first class</a> type may be used as the value arguments
1609 to the PHI node. Only labels may be used as the label arguments.<p>
1611 There must be no non-phi instructions between the start of a basic block and the
1612 PHI instructions: i.e. PHI instructions must be first in a basic block.<p>
1616 At runtime, the '<tt>phi</tt>' instruction logically takes on the value
1617 specified by the parameter, depending on which basic block we came from in the
1618 last <a href="#terminators">terminator</a> instruction.<p>
1623 Loop: ; Infinite loop that counts from 0 on up...
1624 %indvar = phi uint [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
1625 %nextindvar = add uint %indvar, 1
1630 <!-- _______________________________________________________________________ -->
1631 </ul><a name="i_cast"><h4><hr size=0>'<tt>cast .. to</tt>' Instruction</h4><ul>
1635 <result> = cast <ty> <value> to <ty2> <i>; yields ty2</i>
1640 The '<tt>cast</tt>' instruction is used as the primitive means to convert
1641 integers to floating point, change data type sizes, and break type safety (by
1642 casting pointers).<p>
1646 The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
1647 class value, and a type to cast it to, which must also be a <a
1648 href="#t_firstclass">first class</a> type.<p>
1652 This instruction follows the C rules for explicit casts when determining how the
1653 data being cast must change to fit in its new container.<p>
1655 When casting to bool, any value that would be considered true in the context of
1656 a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
1657 all else are '<tt>false</tt>'.<p>
1659 When extending an integral value from a type of one signness to another (for
1660 example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
1661 <b>source</b> value is signed, and zero-extended if the source value is
1662 unsigned. <tt>bool</tt> values are always zero extended into either zero or
1667 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
1668 %Y = cast int 123 to bool <i>; yields bool:true</i>
1673 <!-- _______________________________________________________________________ -->
1674 </ul><a name="i_call"><h4><hr size=0>'<tt>call</tt>' Instruction</h4><ul>
1678 <result> = call <ty>* <fnptrval>(<param list>)
1683 The '<tt>call</tt>' instruction represents a simple function call.<p>
1687 This instruction requires several arguments:<p>
1690 <li>'<tt>ty</tt>': shall be the signature of the pointer to function value being
1691 invoked. The argument types must match the types implied by this signature.<p>
1693 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to be
1694 invoked. In most cases, this is a direct function invocation, but indirect
1695 <tt>call</tt>s are just as possible, calling an arbitrary pointer to function
1698 <li>'<tt>function args</tt>': argument list whose types match the function
1699 signature argument types. If the function signature indicates the function
1700 accepts a variable number of arguments, the extra arguments can be specified.
1705 The '<tt>call</tt>' instruction is used to cause control flow to transfer to a
1706 specified function, with its incoming arguments bound to the specified values.
1707 Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called function,
1708 control flow continues with the instruction after the function call, and the
1709 return value of the function is bound to the result argument. This is a simpler
1710 case of the <a href="#i_invoke">invoke</a> instruction.<p>
1714 %retval = call int %test(int %argc)
1715 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
1719 <!-- _______________________________________________________________________ -->
1720 </ul><a name="i_vanext"><h4><hr size=0>'<tt>vanext</tt>' Instruction</h4><ul>
1724 <resultarglist> = vanext <va_list> <arglist>, <argty>
1729 The '<tt>vanext</tt>' instruction is used to access arguments passed through
1730 the "variable argument" area of a function call. It is used to implement the
1731 <tt>va_arg</tt> macro in C.<p>
1735 This instruction takes a <tt>valist</tt> value and the type of the argument. It
1736 returns another <tt>valist</tt>.
1740 The '<tt>vanext</tt>' instruction advances the specified <tt>valist</tt> past
1741 an argument of the specified type. In conjunction with the <a
1742 href="#i_vaarg"><tt>vaarg</tt></a> instruction, it is used to implement the
1743 <tt>va_arg</tt> macro available in C. For more information, see the variable
1744 argument handling <a href="#int_varargs">Intrinsic Functions</a>.<p>
1746 It is legal for this instruction to be called in a function which does not take
1747 a variable number of arguments, for example, the <tt>vfprintf</tt> function.<p>
1749 <tt>vanext</tt> is an LLVM instruction instead of an <a
1750 href="#intrinsics">intrinsic function</a> because it takes an type as an
1755 See the <a href="#int_varargs">variable argument processing</a> section.<p>
1759 <!-- _______________________________________________________________________ -->
1760 </ul><a name="i_vaarg"><h4><hr size=0>'<tt>vaarg</tt>' Instruction</h4><ul>
1764 <resultval> = vaarg <va_list> <arglist>, <argty>
1769 The '<tt>vaarg</tt>' instruction is used to access arguments passed through
1770 the "variable argument" area of a function call. It is used to implement the
1771 <tt>va_arg</tt> macro in C.<p>
1775 This instruction takes a <tt>valist</tt> value and the type of the argument. It
1776 returns a value of the specified argument type.
1780 The '<tt>vaarg</tt>' instruction loads an argument of the specified type from
1781 the specified <tt>va_list</tt>. In conjunction with the <a
1782 href="#i_vanext"><tt>vanext</tt></a> instruction, it is used to implement the
1783 <tt>va_arg</tt> macro available in C. For more information, see the variable
1784 argument handling <a href="#int_varargs">Intrinsic Functions</a>.<p>
1786 It is legal for this instruction to be called in a function which does not take
1787 a variable number of arguments, for example, the <tt>vfprintf</tt> function.<p>
1789 <tt>vaarg</tt> is an LLVM instruction instead of an <a
1790 href="#intrinsics">intrinsic function</a> because it takes an type as an
1795 See the <a href="#int_varargs">variable argument processing</a> section.<p>
1801 <!-- *********************************************************************** -->
1802 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
1803 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
1804 <a name="intrinsics">Intrinsic Functions
1805 </b></font></td></tr></table><ul>
1806 <!-- *********************************************************************** -->
1808 LLVM supports the notion of an "intrinsic function". These functions have well
1809 known names and semantics, and are required to follow certain restrictions.
1810 Overall, these instructions represent an extension mechanism for the LLVM
1811 language that does not require changing all of the transformations in LLVM to
1812 add to the language (or the bytecode reader/writer, the parser, etc...).<p>
1814 Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix, this
1815 prefix is reserved in LLVM for intrinsic names, thus functions may not be named
1816 this. Intrinsic functions must always be external functions: you cannot define
1817 the body of intrinsic functions. Intrinsic functions may only be used in call
1818 or invoke instructions: it is illegal to take the address of an intrinsic
1819 function. Additionally, because intrinsic functions are part of the LLVM
1820 language, it is required that they all be documented here if any are added.<p>
1822 Unless an intrinsic function is target-specific, there must be a lowering pass
1823 to eliminate the intrinsic or all backends must support the intrinsic
1827 <!-- ======================================================================= -->
1828 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1829 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
1830 <a name="int_varargs">Variable Argument Handling Intrinsics
1831 </b></font></td></tr></table><ul>
1833 Variable argument support is defined in LLVM with the <a
1834 href="#i_vanext"><tt>vanext</tt></a> instruction and these three intrinsic
1835 functions. These functions are related to the similarly named macros defined in
1836 the <tt><stdarg.h></tt> header file.<p>
1838 All of these functions operate on arguments that use a target-specific value
1839 type "<tt>va_list</tt>". The LLVM assembly language reference manual does not
1840 define what this type is, so all transformations should be prepared to handle
1841 intrinsics with any type used.<p>
1843 This example shows how the <a href="#i_vanext"><tt>vanext</tt></a> instruction
1844 and the variable argument handling intrinsic functions are used.<p>
1847 int %test(int %X, ...) {
1848 ; Initialize variable argument processing
1849 %ap = call sbyte*()* %<a href="#i_va_start">llvm.va_start</a>()
1851 ; Read a single integer argument
1852 %tmp = vaarg sbyte* %ap, int
1854 ; Advance to the next argument
1855 %ap2 = vanext sbyte* %ap, int
1857 ; Demonstrate usage of llvm.va_copy and llvm.va_end
1858 %aq = call sbyte* (sbyte*)* %<a href="#i_va_copy">llvm.va_copy</a>(sbyte* %ap2)
1859 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte* %aq)
1861 ; Stop processing of arguments.
1862 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte* %ap2)
1867 <!-- _______________________________________________________________________ -->
1868 </ul><a name="i_va_start"><h4><hr size=0>'<tt>llvm.va_start</tt>' Intrinsic</h4><ul>
1872 call va_list ()* %llvm.va_start()
1877 The '<tt>llvm.va_start</tt>' intrinsic returns a new <tt><arglist></tt>
1878 for subsequent use by the variable argument intrinsics.<p>
1882 The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
1883 macro available in C. In a target-dependent way, it initializes and returns a
1884 <tt>va_list</tt> element, so that the next <tt>vaarg</tt> will produce the first
1885 variable argument passed to the function. Unlike the C <tt>va_start</tt> macro,
1886 this intrinsic does not need to know the last argument of the function, the
1887 compiler can figure that out.<p>
1889 Note that this intrinsic function is only legal to be called from within the
1890 body of a variable argument function.<p>
1893 <!-- _______________________________________________________________________ -->
1894 </ul><a name="i_va_end"><h4><hr size=0>'<tt>llvm.va_end</tt>' Intrinsic</h4><ul>
1898 call void (va_list)* %llvm.va_end(va_list <arglist>)
1903 The '<tt>llvm.va_end</tt>' intrinsic destroys <tt><arglist></tt> which has
1904 been initialized previously with <tt><a
1905 href="#i_va_start">llvm.va_start</a></tt> or <tt><a
1906 href="#i_va_copy">llvm.va_copy</a></tt>.<p>
1910 The argument is a <tt>va_list</tt> to destroy.<p>
1914 The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt> macro
1915 available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
1916 Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
1917 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly with calls
1918 to <tt>llvm.va_end</tt>.<p>
1922 <!-- _______________________________________________________________________ -->
1923 </ul><a name="i_va_copy"><h4><hr size=0>'<tt>llvm.va_copy</tt>' Intrinsic</h4><ul>
1927 call va_list (va_list)* %llvm.va_copy(va_list <destarglist>)
1932 The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
1933 the source argument list to the destination argument list.<p>
1937 The argument is the <tt>va_list</tt> to copy.
1941 The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
1942 available in C. In a target-dependent way, it copies the source
1943 <tt>va_list</tt> element into the returned list. This intrinsic is necessary
1944 because the <tt><a href="i_va_start">llvm.va_start</a></tt> intrinsic may be
1945 arbitrarily complex and require memory allocation, for example.<p>
1948 <!-- *********************************************************************** -->
1950 <!-- *********************************************************************** -->
1955 <address><a href="mailto:sabre@nondot.org">Chris Lattner</a></address>
1956 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a>
1958 <!-- Created: Tue Jan 23 15:19:28 CST 2001 -->
1959 <!-- hhmts start -->
1960 Last modified: Wed Oct 29 19:30:46 CST 2003