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 Assembly 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>
43 <li><a href="#binaryops">Binary Operations</a>
45 <li><a href="#i_add" >'<tt>add</tt>' Instruction</a>
46 <li><a href="#i_sub" >'<tt>sub</tt>' Instruction</a>
47 <li><a href="#i_mul" >'<tt>mul</tt>' Instruction</a>
48 <li><a href="#i_div" >'<tt>div</tt>' Instruction</a>
49 <li><a href="#i_rem" >'<tt>rem</tt>' Instruction</a>
50 <li><a href="#i_setcc">'<tt>set<i>cc</i></tt>' Instructions</a>
52 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
54 <li><a href="#i_and">'<tt>and</tt>' Instruction</a>
55 <li><a href="#i_or" >'<tt>or</tt>' Instruction</a>
56 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a>
57 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a>
58 <li><a href="#i_shr">'<tt>shr</tt>' Instruction</a>
60 <li><a href="#memoryops">Memory Access Operations</a>
62 <li><a href="#i_malloc" >'<tt>malloc</tt>' Instruction</a>
63 <li><a href="#i_free" >'<tt>free</tt>' Instruction</a>
64 <li><a href="#i_alloca" >'<tt>alloca</tt>' Instruction</a>
65 <li><a href="#i_load" >'<tt>load</tt>' Instruction</a>
66 <li><a href="#i_store" >'<tt>store</tt>' Instruction</a>
67 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
69 <li><a href="#otherops">Other Operations</a>
71 <li><a href="#i_phi" >'<tt>phi</tt>' Instruction</a>
72 <li><a href="#i_cast">'<tt>cast .. to</tt>' Instruction</a>
73 <li><a href="#i_call" >'<tt>call</tt>' Instruction</a>
77 <li><a href="#related">Related Work</a>
80 <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>
86 <!-- *********************************************************************** -->
87 <p><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
88 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
89 <a name="abstract">Abstract
90 </b></font></td></tr></table><ul>
91 <!-- *********************************************************************** -->
94 This document is a reference manual for the LLVM assembly language. LLVM is
95 an SSA based representation that provides type safety, low level operations,
96 flexibility, and the capability of representing 'all' high level languages
97 cleanly. It is the common code representation used throughout all phases of
98 the LLVM compilation strategy.
104 <!-- *********************************************************************** -->
105 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
106 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
107 <a name="introduction">Introduction
108 </b></font></td></tr></table><ul>
109 <!-- *********************************************************************** -->
111 The LLVM code representation is designed to be used in three different forms: as
112 an in-memory compiler IR, as an on-disk bytecode representation, suitable for
113 fast loading by a dynamic compiler, and as a human readable assembly language
114 representation. This allows LLVM to provide a powerful intermediate
115 representation for efficient compiler transformations and analysis, while
116 providing a natural means to debug and visualize the transformations. The three
117 different forms of LLVM are all equivalent. This document describes the human
118 readable representation and notation.<p>
120 The LLVM representation aims to be a light weight and low level while being
121 expressive, typed, and extensible at the same time. It aims to be a "universal
122 IR" of sorts, by being at a low enough level that high level ideas may be
123 cleanly mapped to it (similar to how microprocessors are "universal IR's",
124 allowing many source languages to be mapped to them). By providing type
125 information, LLVM can be used as the target of optimizations: for example,
126 through pointer analysis, it can be proven that a C automatic variable is never
127 accessed outside of the current function... allowing it to be promoted to a
128 simple SSA value instead of a memory location.<p>
130 <!-- _______________________________________________________________________ -->
131 </ul><a name="wellformed"><h4><hr size=0>Well Formedness</h4><ul>
133 It is important to note that this document describes 'well formed' llvm assembly
134 language. There is a difference between what the parser accepts and what is
135 considered 'well formed'. For example, the following instruction is
136 syntactically okay, but not well formed:<p>
139 %x = <a href="#i_add">add</a> int 1, %x
142 ...because the definition of %x does not dominate all of its uses. The LLVM
143 infrastructure provides a verification pass that may be used to verify that an
144 LLVM module is well formed. This pass is automatically run by the parser after
145 parsing input assembly, and by the optimizer before it outputs bytecode. The
146 violations pointed out by the verifier pass indicate bugs in transformation
147 passes or input to the parser.<p>
149 <!-- Describe the typesetting conventions here. -->
152 <!-- *********************************************************************** -->
153 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
154 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
155 <a name="identifiers">Identifiers
156 </b></font></td></tr></table><ul>
157 <!-- *********************************************************************** -->
159 LLVM uses three different forms of identifiers, for different purposes:<p>
162 <li>Numeric constants are represented as you would expect: 12, -3 123.421, etc. Floating point constants have an optional hexidecimal notation.
163 <li>Named values are represented as a string of characters with a '%' prefix. For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
164 <li>Unnamed values are represented as an unsigned numeric value with a '%' prefix. For example, %12, %2, %44.
167 LLVM requires the values start with a '%' sign for two reasons: Compilers don't
168 need to worry about name clashes with reserved words, and the set of reserved
169 words may be expanded in the future without penalty. Additionally, unnamed
170 identifiers allow a compiler to quickly come up with a temporary variable
171 without having to avoid symbol table conflicts.<p>
173 Reserved words in LLVM are very similar to reserved words in other languages.
174 There are keywords for different opcodes ('<tt><a href="#i_add">add</a></tt>',
175 '<tt><a href="#i_cast">cast</a></tt>', '<tt><a href="#i_ret">ret</a></tt>',
176 etc...), for primitive type names ('<tt><a href="#t_void">void</a></tt>',
177 '<tt><a href="#t_uint">uint</a></tt>', etc...), and others. These reserved
178 words cannot conflict with variable names, because none of them start with a '%'
181 Here is an example of LLVM code to multiply the integer variable '<tt>%X</tt>'
186 %result = <a href="#i_mul">mul</a> uint %X, 8
189 After strength reduction:
191 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
196 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
197 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
198 %result = <a href="#i_add">add</a> uint %1, %1
201 This last way of multiplying <tt>%X</tt> by 8 illustrates several important lexical features of LLVM:<p>
204 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of line.
205 <li>Unnamed temporaries are created when the result of a computation is not
206 assigned to a named value.
207 <li>Unnamed temporaries are numbered sequentially
210 ...and it also show a convention that we follow in this document. When
211 demonstrating instructions, we will follow an instruction with a comment that
212 defines the type and name of value produced. Comments are shown in italic
215 The one unintuitive notation for constants is the optional hexidecimal form of
216 floating point constants. For example, the form '<tt>double
217 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
218 4.5e+15</tt>' which is also supported by the parser. The only time hexadecimal
219 floating point constants are useful (and the only time that they are generated
220 by the disassembler) is when an FP constant has to be emitted that is not
221 representable as a decimal floating point number exactly. For example, NaN's,
222 infinities, and other special cases are represented in their IEEE hexadecimal
223 format so that assembly and disassembly do not cause any bits to change in the
227 <!-- *********************************************************************** -->
228 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
229 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
230 <a name="typesystem">Type System
231 </b></font></td></tr></table><ul>
232 <!-- *********************************************************************** -->
234 The LLVM type system is one of the most important features of the intermediate
235 representation. Being typed enables a number of optimizations to be performed
236 on the IR directly, without having to do extra analyses on the side before the
237 transformation. A strong type system makes it easier to read the generated code
238 and enables novel analyses and transformations that are not feasible to perform
239 on normal three address code representations.<p>
241 <!-- The written form for the type system was heavily influenced by the
242 syntactic problems with types in the C language<sup><a
243 href="#rw_stroustrup">1</a></sup>.<p> -->
247 <!-- ======================================================================= -->
248 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
249 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
250 <a name="t_primitive">Primitive Types
251 </b></font></td></tr></table><ul>
253 The primitive types are the fundemental building blocks of the LLVM system. The
254 current set of primitive types are as follows:<p>
256 <table border=0 align=center><tr><td>
258 <table border=1 cellspacing=0 cellpadding=4 align=center>
259 <tr><td><tt>void</tt></td> <td>No value</td></tr>
260 <tr><td><tt>ubyte</tt></td> <td>Unsigned 8 bit value</td></tr>
261 <tr><td><tt>ushort</tt></td><td>Unsigned 16 bit value</td></tr>
262 <tr><td><tt>uint</tt></td> <td>Unsigned 32 bit value</td></tr>
263 <tr><td><tt>ulong</tt></td> <td>Unsigned 64 bit value</td></tr>
264 <tr><td><tt>float</tt></td> <td>32 bit floating point value</td></tr>
265 <tr><td><tt>label</tt></td> <td>Branch destination</td></tr>
270 <table border=1 cellspacing=0 cellpadding=4 align=center>
271 <tr><td><tt>bool</tt></td> <td>True or False value</td></tr>
272 <tr><td><tt>sbyte</tt></td> <td>Signed 8 bit value</td></tr>
273 <tr><td><tt>short</tt></td> <td>Signed 16 bit value</td></tr>
274 <tr><td><tt>int</tt></td> <td>Signed 32 bit value</td></tr>
275 <tr><td><tt>long</tt></td> <td>Signed 64 bit value</td></tr>
276 <tr><td><tt>double</tt></td><td>64 bit floating point value</td></tr>
279 </td></tr></table><p>
283 <!-- _______________________________________________________________________ -->
284 </ul><a name="t_classifications"><h4><hr size=0>Type Classifications</h4><ul>
286 These different primitive types fall into a few useful classifications:<p>
288 <table border=1 cellspacing=0 cellpadding=4 align=center>
289 <tr><td><a name="t_signed">signed</td> <td><tt>sbyte, short, int, long, float, double</tt></td></tr>
290 <tr><td><a name="t_unsigned">unsigned</td><td><tt>ubyte, ushort, uint, ulong</tt></td></tr>
291 <tr><td><a name="t_integral">integer</td><td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td></tr>
292 <tr><td><a name="t_integral">integral</td><td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td></tr>
293 <tr><td><a name="t_floating">floating point</td><td><tt>float, double</tt></td></tr>
294 <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>
301 <!-- ======================================================================= -->
302 </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>
303 <a name="t_derived">Derived Types
304 </b></font></td></tr></table><ul>
306 The real power in LLVM comes from the derived types in the system. This is what
307 allows a programmer to represent arrays, functions, pointers, and other useful
308 types. Note that these derived types may be recursive: For example, it is
309 possible to have a two dimensional array.<p>
313 <!-- _______________________________________________________________________ -->
314 </ul><a name="t_array"><h4><hr size=0>Array Type</h4><ul>
318 The array type is a very simple derived type that arranges elements sequentially
319 in memory. The array type requires a size (number of elements) and an
320 underlying data type.<p>
324 [<# elements> x <elementtype>]
327 The number of elements is a constant integer value, elementtype may be any type
332 <tt>[40 x int ]</tt>: Array of 40 integer values.<br>
333 <tt>[41 x int ]</tt>: Array of 41 integer values.<br>
334 <tt>[40 x uint]</tt>: Array of 40 unsigned integer values.<p>
337 Here are some examples of multidimensional arrays:<p>
339 <table border=0 cellpadding=0 cellspacing=0>
340 <tr><td><tt>[3 x [4 x int]]</tt></td><td>: 3x4 array integer values.</td></tr>
341 <tr><td><tt>[12 x [10 x float]]</tt></td><td>: 2x10 array of single precision floating point values.</td></tr>
342 <tr><td><tt>[2 x [3 x [4 x uint]]]</tt></td><td>: 2x3x4 array of unsigned integer values.</td></tr>
347 <!-- _______________________________________________________________________ -->
348 </ul><a name="t_function"><h4><hr size=0>Function Type</h4><ul>
352 The function type can be thought of as a function signature. It consists of a
353 return type and a list of formal parameter types. Function types are usually
354 used when to build virtual function tables (which are structures of pointers to
355 functions), for indirect function calls, and when defining a function.<p>
359 <returntype> (<parameter list>)
362 Where '<tt><parameter list></tt>' is a comma seperated list of type
363 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
364 which indicates that the function takes a variable number of arguments. Note
365 that there currently is no way to define a function in LLVM that takes a
366 variable number of arguments, but it is possible to <b>call</b> a function that
371 <table border=0 cellpadding=0 cellspacing=0>
373 <tr><td><tt>int (int)</tt></td><td>: function taking an <tt>int</tt>, returning
374 an <tt>int</tt></td></tr>
376 <tr><td><tt>float (int, int *) *</tt></td><td>: <a href="#t_pointer">Pointer</a>
377 to a function that takes an <tt>int</tt> and a <a href="#t_pointer">pointer</a>
378 to <tt>int</tt>, returning <tt>float</tt>.</td></tr>
380 <tr><td><tt>int (sbyte *, ...)</tt></td><td>: A vararg function that takes at
381 least one <a href="#t_pointer">pointer</a> to <tt>sbyte</tt> (signed char in C),
382 which returns an integer. This is the signature for <tt>printf</tt> in
390 <!-- _______________________________________________________________________ -->
391 </ul><a name="t_struct"><h4><hr size=0>Structure Type</h4><ul>
395 The structure type is used to represent a collection of data members together in
396 memory. The packing of the field types is defined to match the ABI of the
397 underlying processor. The elements of a structure may be any type that has a
400 Structures are accessed using '<tt><a href="#i_load">load</a></tt> and '<tt><a
401 href="#i_store">store</a></tt>' by getting a pointer to a field with the '<tt><a
402 href="#i_getelementptr">getelementptr</a></tt>' instruction.<p>
406 { <type list> }
411 <table border=0 cellpadding=0 cellspacing=0>
413 <tr><td><tt>{ int, int, int }</tt></td><td>: a triple of three <tt>int</tt>
416 <tr><td><tt>{ float, int (int) * }</tt></td><td>: A pair, where the first
417 element is a <tt>float</tt> and the second element is a <a
418 href="#t_pointer">pointer</a> to a <a href="t_function">function</a> that takes
419 an <tt>int</tt>, returning an <tt>int</tt>.</td></tr>
424 <!-- _______________________________________________________________________ -->
425 </ul><a name="t_pointer"><h4><hr size=0>Pointer Type</h4><ul>
429 As in many languages, the pointer type represents a pointer or reference to
430 another object, which must live in memory.<p>
439 <table border=0 cellpadding=0 cellspacing=0>
441 <tr><td><tt>[4x int]*</tt></td><td>: <a href="#t_pointer">pointer</a> to <a
442 href="#t_array">array</a> of four <tt>int</tt> values</td></tr>
444 <tr><td><tt>int (int *) *</tt></td><td>: A <a href="#t_pointer">pointer</a> to a
445 <a href="t_function">function</a> that takes an <tt>int</tt>, returning an
446 <tt>int</tt>.</td></tr>
452 <!-- _______________________________________________________________________ -->
454 </ul><a name="t_packed"><h4><hr size=0>Packed Type</h4><ul>
456 Mention/decide that packed types work with saturation or not. Maybe have a packed+saturated type in addition to just a packed type.<p>
458 Packed types should be 'nonsaturated' because standard data types are not saturated. Maybe have a saturated packed type?<p>
463 <!-- *********************************************************************** -->
464 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
465 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
466 <a name="highlevel">High Level Structure
467 </b></font></td></tr></table><ul>
468 <!-- *********************************************************************** -->
471 <!-- ======================================================================= -->
472 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
473 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
474 <a name="modulestructure">Module Structure
475 </b></font></td></tr></table><ul>
477 LLVM programs are composed of "Module"s, each of which is a translation unit of
478 the input programs. Each module consists of functions, global variables, and
479 symbol table entries. Modules may be combined together with the LLVM linker,
480 which merges function (and global variable) definitions, resolves forward
481 declarations, and merges symbol table entries. Here is an example of the "hello world" module:<p>
484 <i>; Declare the string constant as a global constant...</i>
485 <a href="#identifiers">%.LC0</a> = <a href="#linkage_decl">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>
487 <i>; Forward declaration of puts</i>
488 <a href="#functionstructure">declare</a> int "puts"(sbyte*) <i>; int(sbyte*)* </i>
490 <i>; Definition of main function</i>
491 int "main"() { <i>; int()* </i>
492 <i>; Convert [13x sbyte]* to sbyte *...</i>
493 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, uint 0, uint 0 <i>; sbyte*</i>
495 <i>; Call puts function to write out the string to stdout...</i>
496 <a href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
497 <a href="#i_ret">ret</a> int 0
501 This example is made up of a <a href="#globalvars">global variable</a> named
502 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function, and a
503 <a href="#functionstructure">function definition</a> for "<tt>main</tt>".<p>
505 <a name="linkage_decl">
506 In general, a module is made up of a list of global values, where both functions
507 and global variables are global values. Global values are represented by a
508 pointer to a memory location (in this case, a pointer to an array of char, and a
509 pointer to a function), and can be either "internal" or externally accessible
510 (which corresponds to the static keyword in C, when used at global scope).<p>
512 For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
513 another module defined a "<tt>.LC0</tt>" variable and was linked with this one,
514 one of the two would be renamed, preventing a collision. Since "<tt>main</tt>"
515 and "<tt>puts</tt>" are external (i.e., lacking "<tt>internal</tt>"
516 declarations), they are accessible outside of the current module. It is illegal
517 for a function declaration to be "<tt>internal</tt>".<p>
520 <!-- ======================================================================= -->
521 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
522 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
523 <a name="globalvars">Global Variables
524 </b></font></td></tr></table><ul>
526 Global variables define regions of memory allocated at compilation time instead
527 of run-time. Global variables may optionally be initialized. A variable may
528 be defined as a global "constant", which indicates that the contents of the
529 variable will never be modified (opening options for optimization). Constants
530 must always have an initial value.<p>
532 As SSA values, global variables define pointer values that are in scope
533 (i.e. they dominate) for all basic blocks in the program. Global variables
534 always define a pointer to their "content" type because they describe a region
535 of memory, and all memory objects in LLVM are accessed through pointers.<p>
539 <!-- ======================================================================= -->
540 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
541 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
542 <a name="functionstructure">Function Structure
543 </b></font></td></tr></table><ul>
545 LLVM functions definitions are composed of a (possibly empty) argument list, an
546 opening curly brace, a list of basic blocks, and a closing curly brace. LLVM
547 function declarations are defined with the "<tt>declare</tt>" keyword, a
548 function name and a function signature.<p>
550 A function definition contains a list of basic blocks, forming the CFG for the
551 function. Each basic block may optionally start with a label (giving the basic
552 block a symbol table entry), contains a list of instructions, and ends with a <a
553 href="#terminators">terminator</a> instruction (such as a branch or function
556 The first basic block in program is special in two ways: it is immediately
557 executed on entrance to the function, and it is not allowed to have predecessor
558 basic blocks (i.e. there can not be any branches to the entry block of a
562 <!-- *********************************************************************** -->
563 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
564 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
565 <a name="instref">Instruction Reference
566 </b></font></td></tr></table><ul>
567 <!-- *********************************************************************** -->
569 The LLVM instruction set consists of several different classifications of
570 instructions: <a href="#terminators">terminator instructions</a>, <a
571 href="#binaryops">binary instructions</a>, <a href="#memoryops">memory
572 instructions</a>, and <a href="#otherops">other instructions</a>.<p>
575 <!-- ======================================================================= -->
576 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
577 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
578 <a name="terminators">Terminator Instructions
579 </b></font></td></tr></table><ul>
581 As mentioned <a href="#functionstructure">previously</a>, every basic block in a
582 program ends with a "Terminator" instruction, which indicates which block should
583 be executed after the current block is finished. These terminator instructions
584 typically yield a '<tt>void</tt>' value: they produce control flow, not values
585 (the one exception being the '<a href="#i_invoke"><tt>invoke</tt></a>'
588 There are four different terminator instructions: the '<a
589 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a
590 href="#i_br"><tt>br</tt></a>' instruction, the '<a
591 href="#i_switch"><tt>switch</tt></a>' instruction, and the '<a
592 href="#i_invoke"><tt>invoke</tt></a>' instruction.<p>
595 <!-- _______________________________________________________________________ -->
596 </ul><a name="i_ret"><h4><hr size=0>'<tt>ret</tt>' Instruction</h4><ul>
600 ret <type> <value> <i>; Return a value from a non-void function</i>
601 ret void <i>; Return from void function</i>
606 The '<tt>ret</tt>' instruction is used to return control flow (and a value) from
607 a function, back to the caller.<p>
609 There are two forms of the '<tt>ret</tt>' instructruction: one that returns a
610 value and then causes control flow, and one that just causes control flow to
615 The '<tt>ret</tt>' instruction may return any '<a href="#t_firstclass">first
616 class</a>' type. Notice that a function is not <a href="#wellformed">well
617 formed</a> if there exists a '<tt>ret</tt>' instruction inside of the function
618 that returns a value that does not match the return type of the function.<p>
622 When the '<tt>ret</tt>' instruction is executed, control flow returns back to
623 the calling function's context. If the instruction returns a value, that value
624 shall be propagated into the calling function's data space.<p>
628 ret int 5 <i>; Return an integer value of 5</i>
629 ret void <i>; Return from a void function</i>
633 <!-- _______________________________________________________________________ -->
634 </ul><a name="i_br"><h4><hr size=0>'<tt>br</tt>' Instruction</h4><ul>
638 br bool <cond>, label <iftrue>, label <iffalse>
639 br label <dest> <i>; Unconditional branch</i>
644 The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
645 different basic block in the current function. There are two forms of this
646 instruction, corresponding to a conditional branch and an unconditional
651 The conditional branch form of the '<tt>br</tt>' instruction takes a single
652 '<tt>bool</tt>' value and two '<tt>label</tt>' values. The unconditional form
653 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
658 Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
659 argument is evaluated. If the value is <tt>true</tt>, control flows to the
660 '<tt>iftrue</tt>' '<tt>label</tt>' argument. If "cond" is <tt>false</tt>,
661 control flows to the '<tt>iffalse</tt>' '<tt>label</tt>' argument.<p>
666 %cond = <a href="#i_setcc">seteq</a> int %a, %b
667 br bool %cond, label %IfEqual, label %IfUnequal
669 <a href="#i_ret">ret</a> int 1
671 <a href="#i_ret">ret</a> int 0
675 <!-- _______________________________________________________________________ -->
676 </ul><a name="i_switch"><h4><hr size=0>'<tt>switch</tt>' Instruction</h4><ul>
680 <i>; Definitions for lookup indirect branch</i>
681 %switchtype = type [<anysize> x { uint, label }]
683 <i>; Lookup indirect branch</i>
684 switch uint <value>, label <defaultdest>, %switchtype <switchtable>
686 <i>; Indexed indirect branch</i>
687 switch uint <idxvalue>, label <defaultdest>, [<anysize> x label] <desttable>
693 <b>NOTE:</b> The switch instruction may go away in the future. It is not very
694 well supported in LLVM anyway, so don't go to great lengths to support it. Talk
695 to <a href="mailto:sabre@nondot.org">Chris</a> for more info if this concerns
698 The '<tt>switch</tt>' instruction is used to transfer control flow to one of
699 several different places. It is a generalization of the '<tt>br</tt>'
700 instruction, allowing a branch to occur to one of many possible destinations.<p>
702 The '<tt>switch</tt>' statement supports two different styles of indirect
703 branching: lookup branching and indexed branching. Lookup branching is
704 generally useful if the values to switch on are spread far appart, where index
705 branching is useful if the values to switch on are generally dense.<p>
707 The two different forms of the '<tt>switch</tt>' statement are simple hints to
708 the underlying implementation. For example, the compiler may choose to
709 implement a small indirect branch table as a series of predicated comparisons:
710 if it is faster for the target architecture.<p>
714 The lookup form of the '<tt>switch</tt>' instruction uses three parameters: a
715 '<tt>uint</tt>' comparison value '<tt>value</tt>', a default '<tt>label</tt>'
716 destination, and an array of pairs of comparison value constants and
717 '<tt>label</tt>'s. The sized array must be a constant value.<p>
719 The indexed form of the '<tt>switch</tt>' instruction uses three parameters: an
720 '<tt>uint</tt>' index value, a default '<tt>label</tt>' and a sized array of
721 '<tt>label</tt>'s. The '<tt>dests</tt>' array must be a constant array.
725 The lookup style switch statement specifies a table of values and destinations.
726 When the '<tt>switch</tt>' instruction is executed, this table is searched for
727 the given value. If the value is found, the corresponding destination is
730 The index branch form simply looks up a label element directly in a table and
733 In either case, the compiler knows the static size of the array, because it is
734 provided as part of the constant values type.<p>
738 <i>; Emulate a conditional br instruction</i>
739 %Val = <a href="#i_cast">cast</a> bool %value to uint
740 switch uint %Val, label %truedest, [1 x label] [label %falsedest ]
742 <i>; Emulate an unconditional br instruction</i>
743 switch uint 0, label %dest, [ 0 x label] [ ]
745 <i>; Implement a jump table:</i>
746 switch uint %val, label %otherwise, [3 x label] [ label %onzero,
754 <!-- _______________________________________________________________________ -->
755 </ul><a name="i_invoke"><h4><hr size=0>'<tt>invoke</tt>' Instruction</h4><ul>
759 <result> = invoke <ptr to function ty> %<function ptr val>(<function args>)
760 to label <normal label> except label <exception label>
765 The '<tt>invoke</tt>' instruction is used to cause control flow to transfer to a
766 specified function, with the possibility of control flow transfer to either the
767 '<tt>normal label</tt>' label or the '<tt>exception label</tt>'. The '<tt><a
768 href="#i_call">call</a></tt>' instruction is closely related, but guarantees
769 that control flow either never returns from the called function, or that it
770 returns to the instruction following the '<tt><a href="#i_call">call</a></tt>'
775 This instruction requires several arguments:<p>
778 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
779 function value being invoked. In most cases, this is a direct function
780 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
781 an arbitrary pointer to function value.<p>
783 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
784 function to be invoked.
786 <li>'<tt>function args</tt>': argument list whose types match the function
787 signature argument types. If the function signature indicates the function
788 accepts a variable number of arguments, the extra arguments can be specified.
790 <li>'<tt>normal label</tt>': the label reached when the called function executes
791 a '<tt><a href="#i_ret">ret</a></tt>' instruction.
793 <li>'<tt>exception label</tt>': the label reached when an exception is thrown.
798 This instruction is designed to operate as a standard '<tt><a
799 href="#i_call">call</a></tt>' instruction in most regards. The primary
800 difference is that it associates a label with the function invocation that may
801 be accessed via the runtime library provided by the execution environment. This
802 instruction is used in languages with destructors to ensure that proper cleanup
803 is performed in the case of either a <tt>longjmp</tt> or a thrown exception.
804 Additionally, this is important for implementation of '<tt>catch</tt>' clauses
805 in high-level languages that support them.<p>
807 <!-- For a more comprehensive explanation of how this instruction is used, look in the llvm/docs/2001-05-18-ExceptionHandling.txt document.<p> -->
811 %retval = invoke int %Test(int 15)
812 to label %Continue except label %TestCleanup <i>; {int}:retval set</i>
817 <!-- ======================================================================= -->
818 </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>
819 <a name="binaryops">Binary Operations
820 </b></font></td></tr></table><ul>
822 Binary operators are used to do most of the computation in a program. They
823 require two operands, execute an operation on them, and produce a single value.
824 The result value of a binary operator is not neccesarily the same type as its
827 There are several different binary operators:<p>
830 <!-- _______________________________________________________________________ -->
831 </ul><a name="i_add"><h4><hr size=0>'<tt>add</tt>' Instruction</h4><ul>
835 <result> = add <ty> <var1>, <var2> <i>; yields {ty}:result</i>
839 The '<tt>add</tt>' instruction returns the sum of its two operands.<p>
842 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>
846 The value produced is the integer or floating point sum of the two operands.<p>
850 <result> = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
854 <!-- _______________________________________________________________________ -->
855 </ul><a name="i_sub"><h4><hr size=0>'<tt>sub</tt>' Instruction</h4><ul>
859 <result> = sub <ty> <var1>, <var2> <i>; yields {ty}:result</i>
864 The '<tt>sub</tt>' instruction returns the difference of its two operands.<p>
866 Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
867 instruction present in most other intermediate representations.<p>
871 The two arguments to the '<tt>sub</tt>' instruction must be either <a
872 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
873 values. Both arguments must have identical types.<p>
877 The value produced is the integer or floating point difference of the two
882 <result> = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
883 <result> = sub int 0, %val <i>; yields {int}:result = -%var</i>
886 <!-- _______________________________________________________________________ -->
887 </ul><a name="i_mul"><h4><hr size=0>'<tt>mul</tt>' Instruction</h4><ul>
891 <result> = mul <ty> <var1>, <var2> <i>; yields {ty}:result</i>
895 The '<tt>mul</tt>' instruction returns the product of its two operands.<p>
898 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>
902 The value produced is the integer or floating point product of the two
905 There is no signed vs unsigned multiplication. The appropriate action is taken
906 based on the type of the operand. <p>
911 <result> = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
915 <!-- _______________________________________________________________________ -->
916 </ul><a name="i_div"><h4><hr size=0>'<tt>div</tt>' Instruction</h4><ul>
920 <result> = div <ty> <var1>, <var2> <i>; yields {ty}:result</i>
925 The '<tt>div</tt>' instruction returns the quotient of its two operands.<p>
929 The two arguments to the '<tt>div</tt>' instruction must be either <a
930 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
931 values. Both arguments must have identical types.<p>
935 The value produced is the integer or floating point quotient of the two
940 <result> = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
944 <!-- _______________________________________________________________________ -->
945 </ul><a name="i_rem"><h4><hr size=0>'<tt>rem</tt>' Instruction</h4><ul>
949 <result> = rem <ty> <var1>, <var2> <i>; yields {ty}:result</i>
953 The '<tt>rem</tt>' instruction returns the remainder from the division of its two operands.<p>
956 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>
960 This returns the <i>remainder</i> of a division (where the result has the same
961 sign as the divisor), not the <i>modulus</i> (where the result has the same sign
962 as the dividend) of a value. For more information about the difference, see: <a
963 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The Math
968 <result> = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
972 <!-- _______________________________________________________________________ -->
973 </ul><a name="i_setcc"><h4><hr size=0>'<tt>set<i>cc</i></tt>' Instructions</h4><ul>
977 <result> = seteq <ty> <var1>, <var2> <i>; yields {bool}:result</i>
978 <result> = setne <ty> <var1>, <var2> <i>; yields {bool}:result</i>
979 <result> = setlt <ty> <var1>, <var2> <i>; yields {bool}:result</i>
980 <result> = setgt <ty> <var1>, <var2> <i>; yields {bool}:result</i>
981 <result> = setle <ty> <var1>, <var2> <i>; yields {bool}:result</i>
982 <result> = setge <ty> <var1>, <var2> <i>; yields {bool}:result</i>
985 <h5>Overview:</h5> The '<tt>set<i>cc</i></tt>' family of instructions returns a
986 boolean value based on a comparison of their two operands.<p>
988 <h5>Arguments:</h5> The two arguments to the '<tt>set<i>cc</i></tt>'
989 instructions must be of <a href="#t_firstclass">first class</a> or <a
990 href="#t_pointer">pointer</a> type (it is not possible to compare
991 '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>' or '<tt>void</tt>'
992 values, etc...). Both arguments must have identical types.<p>
994 The '<tt>setlt</tt>', '<tt>setgt</tt>', '<tt>setle</tt>', and '<tt>setge</tt>'
995 instructions do not operate on '<tt>bool</tt>' typed arguments.<p>
999 The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1000 both operands are equal.<br>
1002 The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1003 both operands are unequal.<br>
1005 The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1006 the first operand is less than the second operand.<br>
1008 The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1009 the first operand is greater than the second operand.<br>
1011 The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1012 the first operand is less than or equal to the second operand.<br>
1014 The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>' value if
1015 the first operand is greater than or equal to the second operand.<p>
1019 <result> = seteq int 4, 5 <i>; yields {bool}:result = false</i>
1020 <result> = setne float 4, 5 <i>; yields {bool}:result = true</i>
1021 <result> = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1022 <result> = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1023 <result> = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1024 <result> = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1029 <!-- ======================================================================= -->
1030 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1031 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
1032 <a name="bitwiseops">Bitwise Binary Operations
1033 </b></font></td></tr></table><ul>
1035 Bitwise binary operators are used to do various forms of bit-twiddling in a
1036 program. They are generally very efficient instructions, and can commonly be
1037 strength reduced from other instructions. They require two operands, execute an
1038 operation on them, and produce a single value. The resulting value of the
1039 bitwise binary operators is always the same type as its first operand.<p>
1041 <!-- _______________________________________________________________________ -->
1042 </ul><a name="i_and"><h4><hr size=0>'<tt>and</tt>' Instruction</h4><ul>
1046 <result> = and <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1050 The '<tt>and</tt>' instruction returns the bitwise logical and of its two operands.<p>
1054 The two arguments to the '<tt>and</tt>' instruction must be <a
1055 href="#t_integral">integral</a> values. Both arguments must have identical
1061 The truth table used for the '<tt>and</tt>' instruction is:<p>
1063 <center><table border=1 cellspacing=0 cellpadding=4>
1064 <tr><td>In0</td> <td>In1</td> <td>Out</td></tr>
1065 <tr><td>0</td> <td>0</td> <td>0</td></tr>
1066 <tr><td>0</td> <td>1</td> <td>0</td></tr>
1067 <tr><td>1</td> <td>0</td> <td>0</td></tr>
1068 <tr><td>1</td> <td>1</td> <td>1</td></tr>
1069 </table></center><p>
1074 <result> = and int 4, %var <i>; yields {int}:result = 4 & %var</i>
1075 <result> = and int 15, 40 <i>; yields {int}:result = 8</i>
1076 <result> = and int 4, 8 <i>; yields {int}:result = 0</i>
1081 <!-- _______________________________________________________________________ -->
1082 </ul><a name="i_or"><h4><hr size=0>'<tt>or</tt>' Instruction</h4><ul>
1086 <result> = or <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1089 <h5>Overview:</h5> The '<tt>or</tt>' instruction returns the bitwise logical
1090 inclusive or of its two operands.<p>
1094 The two arguments to the '<tt>or</tt>' instruction must be <a
1095 href="#t_integral">integral</a> values. Both arguments must have identical
1101 The truth table used for the '<tt>or</tt>' instruction is:<p>
1103 <center><table border=1 cellspacing=0 cellpadding=4>
1104 <tr><td>In0</td> <td>In1</td> <td>Out</td></tr>
1105 <tr><td>0</td> <td>0</td> <td>0</td></tr>
1106 <tr><td>0</td> <td>1</td> <td>1</td></tr>
1107 <tr><td>1</td> <td>0</td> <td>1</td></tr>
1108 <tr><td>1</td> <td>1</td> <td>1</td></tr>
1109 </table></center><p>
1114 <result> = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
1115 <result> = or int 15, 40 <i>; yields {int}:result = 47</i>
1116 <result> = or int 4, 8 <i>; yields {int}:result = 12</i>
1120 <!-- _______________________________________________________________________ -->
1121 </ul><a name="i_xor"><h4><hr size=0>'<tt>xor</tt>' Instruction</h4><ul>
1125 <result> = xor <ty> <var1>, <var2> <i>; yields {ty}:result</i>
1130 The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of its
1135 The two arguments to the '<tt>xor</tt>' instruction must be <a
1136 href="#t_integral">integral</a> values. Both arguments must have identical
1142 The truth table used for the '<tt>xor</tt>' instruction is:<p>
1144 <center><table border=1 cellspacing=0 cellpadding=4>
1145 <tr><td>In0</td> <td>In1</td> <td>Out</td></tr>
1146 <tr><td>0</td> <td>0</td> <td>0</td></tr>
1147 <tr><td>0</td> <td>1</td> <td>1</td></tr>
1148 <tr><td>1</td> <td>0</td> <td>1</td></tr>
1149 <tr><td>1</td> <td>1</td> <td>0</td></tr>
1150 </table></center><p>
1155 <result> = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
1156 <result> = xor int 15, 40 <i>; yields {int}:result = 39</i>
1157 <result> = xor int 4, 8 <i>; yields {int}:result = 12</i>
1161 <!-- _______________________________________________________________________ -->
1162 </ul><a name="i_shl"><h4><hr size=0>'<tt>shl</tt>' Instruction</h4><ul>
1166 <result> = shl <ty> <var1>, ubyte <var2> <i>; yields {ty}:result</i>
1171 The '<tt>shl</tt>' instruction returns the first operand shifted to the left a
1172 specified number of bits.
1176 The first argument to the '<tt>shl</tt>' instruction must be an <a
1177 href="#t_integer">integer</a> type. The second argument must be an
1178 '<tt>ubyte</tt>' type.<p>
1182 The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.<p>
1187 <result> = shl int 4, ubyte %var <i>; yields {int}:result = 4 << %var</i>
1188 <result> = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1189 <result> = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1193 <!-- _______________________________________________________________________ -->
1194 </ul><a name="i_shr"><h4><hr size=0>'<tt>shr</tt>' Instruction</h4><ul>
1199 <result> = shr <ty> <var1>, ubyte <var2> <i>; yields {ty}:result</i>
1203 The '<tt>shr</tt>' instruction returns the first operand shifted to the right a specified number of bits.
1206 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>
1210 If the first argument is a <a href="#t_signed">signed</a> type, the most
1211 significant bit is duplicated in the newly free'd bit positions. If the first
1212 argument is unsigned, zero bits shall fill the empty positions.<p>
1216 <result> = shr int 4, ubyte %var <i>; yields {int}:result = 4 >> %var</i>
1217 <result> = shr int 4, ubyte 1 <i>; yields {int}:result = 2</i>
1218 <result> = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
1219 <result> = shr int 4, ubyte 3 <i>; yields {int}:result = 0</i>
1226 <!-- ======================================================================= -->
1227 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1228 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
1229 <a name="memoryops">Memory Access Operations
1230 </b></font></td></tr></table><ul>
1232 Accessing memory in SSA form is, well, sticky at best. This section describes how to read, write, allocate and free memory in LLVM.<p>
1235 <!-- _______________________________________________________________________ -->
1236 </ul><a name="i_malloc"><h4><hr size=0>'<tt>malloc</tt>' Instruction</h4><ul>
1240 <result> = malloc <type>, uint <NumElements> <i>; yields {type*}:result</i>
1241 <result> = malloc <type> <i>; yields {type*}:result</i>
1245 The '<tt>malloc</tt>' instruction allocates memory from the system heap and returns a pointer to it.<p>
1249 The the '<tt>malloc</tt>' instruction allocates
1250 <tt>sizeof(<type>)*NumElements</tt> bytes of memory from the operating
1251 system, and returns a pointer of the appropriate type to the program. The
1252 second form of the instruction is a shorter version of the first instruction
1253 that defaults to allocating one element.<p>
1255 '<tt>type</tt>' must be a sized type<p>
1258 Memory is allocated, a pointer is returned.<p>
1262 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
1264 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
1265 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
1266 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
1270 <!-- _______________________________________________________________________ -->
1271 </ul><a name="i_free"><h4><hr size=0>'<tt>free</tt>' Instruction</h4><ul>
1275 free <type> <value> <i>; yields {void}</i>
1280 The '<tt>free</tt>' instruction returns memory back to the unused memory heap, to be reallocated in the future.<p>
1285 '<tt>value</tt>' shall be a pointer value that points to a value that was
1286 allocated with the '<tt><a href="#i_malloc">malloc</a></tt>' instruction.<p>
1291 Access to the memory pointed to by the pointer is not longer defined after this instruction executes.<p>
1295 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
1296 free [4 x ubyte]* %array
1300 <!-- _______________________________________________________________________ -->
1301 </ul><a name="i_alloca"><h4><hr size=0>'<tt>alloca</tt>' Instruction</h4><ul>
1305 <result> = alloca <type>, uint <NumElements> <i>; yields {type*}:result</i>
1306 <result> = alloca <type> <i>; yields {type*}:result</i>
1311 The '<tt>alloca</tt>' instruction allocates memory on the current stack frame of
1312 the procedure that is live until the current function returns to its caller.<p>
1316 The the '<tt>alloca</tt>' instruction allocates
1317 <tt>sizeof(<type>)*NumElements</tt> bytes of memory on the runtime stack,
1318 returning a pointer of the appropriate type to the program. The second form of
1319 the instruction is a shorter version of the first that defaults to allocating
1322 '<tt>type</tt>' may be any sized type.<p>
1326 Memory is allocated, a pointer is returned. '<tt>alloca</tt>'d memory is
1327 automatically released when the function returns. The '<tt>alloca</tt>'
1328 instruction is commonly used to represent automatic variables that must have an
1329 address available, as well as spilled variables.<p>
1333 %ptr = alloca int <i>; yields {int*}:ptr</i>
1334 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
1338 <!-- _______________________________________________________________________ -->
1339 </ul><a name="i_load"><h4><hr size=0>'<tt>load</tt>' Instruction</h4><ul>
1343 <result> = load <ty>* <pointer>
1347 The '<tt>load</tt>' instruction is used to read from memory.<p>
1351 The argument to the '<tt>load</tt>' instruction specifies the memory address to load from. The pointer must point to a <a href="t_firstclass">first class</a> type.<p>
1355 The location of memory pointed to is loaded.
1359 %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1360 <a href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
1361 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1367 <!-- _______________________________________________________________________ -->
1368 </ul><a name="i_store"><h4><hr size=0>'<tt>store</tt>' Instruction</h4><ul>
1372 store <ty> <value>, <ty>* <pointer> <i>; yields {void}</i>
1376 The '<tt>store</tt>' instruction is used to write to memory.<p>
1380 There are two arguments to the '<tt>store</tt>' instruction: a value to store
1381 and an address to store it into. The type of the '<tt><pointer></tt>'
1382 operand must be a pointer to the type of the '<tt><value></tt>'
1385 <h5>Semantics:</h5> The contents of memory are updated to contain
1386 '<tt><value></tt>' at the location specified by the
1387 '<tt><pointer></tt>' operand.<p>
1391 %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1392 <a href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
1393 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1399 <!-- _______________________________________________________________________ -->
1400 </ul><a name="i_getelementptr"><h4><hr size=0>'<tt>getelementptr</tt>' Instruction</h4><ul>
1404 <result> = getelementptr <ty>* <ptrval>{, uint <aidx>|, ubyte <sidx>}*
1409 The '<tt>getelementptr</tt>' instruction is used to get the address of a
1410 subelement of an aggregate data structure.<p>
1414 This instruction takes a list of <tt>uint</tt> values and <tt>ubyte</tt>
1415 constants that indicate what form of addressing to perform. The actual types of
1416 the arguments provided depend on the type of the first pointer argument. The
1417 '<tt>getelementptr</tt>' instruction is used to index down through the type
1418 levels of a structure.<p>
1420 For example, lets consider a C code fragment and how it gets compiled to
1435 int *foo(struct ST *s) {
1436 return &s[1].Z.B[5][13];
1440 The LLVM code generated by the GCC frontend is:
1443 %RT = type { sbyte, [10 x [20 x int]], sbyte }
1444 %ST = type { int, double, %RT }
1446 int* "foo"(%ST* %s) {
1447 %reg = getelementptr %ST* %s, uint 1, ubyte 2, ubyte 1, uint 5, uint 13
1454 The index types specified for the '<tt>getelementptr</tt>' instruction depend on
1455 the pointer type that is being index into. <a href="t_pointer">Pointer</a> and
1456 <a href="t_array">array</a> types require '<tt>uint</tt>' values, and <a
1457 href="t_struct">structure</a> types require '<tt>ubyte</tt>'
1458 <b>constants</b>.<p>
1460 In the example above, the first index is indexing into the '<tt>%ST*</tt>' type,
1461 which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT }</tt>'
1462 type, a structure. The second index indexes into the third element of the
1463 structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]], sbyte
1464 }</tt>' type, another structure. The third index indexes into the second
1465 element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
1466 array. The two dimensions of the array are subscripted into, yielding an
1467 '<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction return a pointer
1468 to this element, thus yielding a '<tt>int*</tt>' type.<p>
1470 Note that it is perfectly legal to index partially through a structure,
1471 returning a pointer to an inner element. Because of this, the LLVM code for the
1472 given testcase is equivalent to:<p>
1475 int* "foo"(%ST* %s) {
1476 %t1 = getelementptr %ST* %s , uint 1 <i>; yields %ST*:%t1</i>
1477 %t2 = getelementptr %ST* %t1, uint 0, ubyte 2 <i>; yields %RT*:%t2</i>
1478 %t3 = getelementptr %RT* %t2, uint 0, ubyte 1 <i>; yields [10 x [20 x int]]*:%t3</i>
1479 %t4 = getelementptr [10 x [20 x int]]* %t3, uint 0, uint 5 <i>; yields [20 x int]*:%t4</i>
1480 %t5 = getelementptr [20 x int]* %t4, uint 0, uint 13 <i>; yields int*:%t5</i>
1489 <i>; yields [12 x ubyte]*:aptr</i>
1490 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, uint 0, ubyte 1
1495 <!-- ======================================================================= -->
1496 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1497 <tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
1498 <a name="otherops">Other Operations
1499 </b></font></td></tr></table><ul>
1501 The instructions in this catagory are the "miscellaneous" functions, that defy better classification.<p>
1504 <!-- _______________________________________________________________________ -->
1505 </ul><a name="i_phi"><h4><hr size=0>'<tt>phi</tt>' Instruction</h4><ul>
1509 <result> = phi <ty> [ <val0>, <label0>], ...
1514 The '<tt>phi</tt>' instruction is used to implement the φ node in the SSA
1515 graph representing the function.<p>
1519 The type of the incoming values are specified with the first type field. After
1520 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
1521 one pair for each predecessor basic block of the current block.<p>
1523 There must be no non-phi instructions between the start of a basic block and the
1524 PHI instructions: i.e. PHI instructions must be first in a basic block.<p>
1528 At runtime, the '<tt>phi</tt>' instruction logically takes on the value
1529 specified by the parameter, depending on which basic block we came from in the
1530 last <a href="#terminators">terminator</a> instruction.<p>
1535 Loop: ; Infinite loop that counts from 0 on up...
1536 %indvar = phi uint [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
1537 %nextindvar = add uint %indvar, 1
1542 <!-- _______________________________________________________________________ -->
1543 </ul><a name="i_cast"><h4><hr size=0>'<tt>cast .. to</tt>' Instruction</h4><ul>
1547 <result> = cast <ty> <value> to <ty2> <i>; yields ty2</i>
1552 The '<tt>cast</tt>' instruction is used as the primitive means to convert
1553 integers to floating point, change data type sizes, and break type safety (by
1554 casting pointers).<p>
1558 The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
1559 class value, and a type to cast it to, which must also be a first class type.<p>
1563 This instruction follows the C rules for explicit casts when determining how the
1564 data being cast must change to fit in its new container.<p>
1566 When casting to bool, any value that would be considered true in the context of
1567 a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
1568 all else are '<tt>false</tt>'.<p>
1570 When extending an integral value from a type of one signness to another (for
1571 example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
1572 <b>source</b> value is signed, and zero-extended if the source value is
1573 unsigned. <tt>bool</tt> values are always zero extended into either zero or
1578 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
1579 %Y = cast int 123 to bool <i>; yields bool:true</i>
1584 <!-- _______________________________________________________________________ -->
1585 </ul><a name="i_call"><h4><hr size=0>'<tt>call</tt>' Instruction</h4><ul>
1589 <result> = call <ty>* <fnptrval>(<param list>)
1594 The '<tt>call</tt>' instruction represents a simple function call.<p>
1598 This instruction requires several arguments:<p>
1601 <li>'<tt>ty</tt>': shall be the signature of the pointer to function value being
1602 invoked. The argument types must match the types implied by this signature.<p>
1604 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to be
1605 invoked. In most cases, this is a direct function invocation, but indirect
1606 <tt>call</tt>s are just as possible, calling an arbitrary pointer to function
1609 <li>'<tt>function args</tt>': argument list whose types match the function
1610 signature argument types. If the function signature indicates the function
1611 accepts a variable number of arguments, the extra arguments can be specified.
1616 The '<tt>call</tt>' instruction is used to cause control flow to transfer to a
1617 specified function, with its incoming arguments bound to the specified values.
1618 Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called function,
1619 control flow continues with the instruction after the function call, and the
1620 return value of the function is bound to the result argument. This is a simpler
1621 case of the <a href="#i_invoke">invoke</a> instruction.<p>
1625 %retval = call int %test(int %argc)
1626 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
1632 <!x- *********************************************************************** -x>
1633 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
1634 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
1635 <a name="related">Related Work
1636 </b></font></td></tr></table><ul>
1637 <!x- *********************************************************************** -x>
1640 Codesigned virtual machines.<p>
1643 <a name="rw_safetsa">
1645 <DD>Description here<p>
1648 <dt><a href="http://www.javasoft.com">Java</a>
1649 <DD>Desciption here<p>
1652 <dt><a href="http://www.microsoft.com/net">Microsoft .net</a>
1653 <DD>Desciption here<p>
1655 <a name="rw_gccrtl">
1656 <dt><a href="http://www.math.umn.edu/systems_guide/gcc-2.95.1/gcc_15.html">GNU RTL Intermediate Representation</a>
1657 <DD>Desciption here<p>
1660 <dt><a href="http://developer.intel.com/design/ia-64/index.htm">IA64 Architecture & Instruction Set</a>
1661 <DD>Desciption here<p>
1664 <dt><a href="http://www-cs-faculty.stanford.edu/~knuth/mmix-news.html">MMIX Instruction Set</a>
1665 <DD>Desciption here<p>
1667 <a name="rw_stroustrup">
1668 <dt><a href="http://www.research.att.com/~bs/devXinterview.html">"Interview With Bjarne Stroustrup"</a>
1669 <DD>This interview influenced the design and thought process behind LLVM in several ways, most notably the way that derived types are written in text format. See the question that starts with "you defined the C declarator syntax as an experiment that failed".<p>
1672 <!x- _______________________________________________________________________ -x>
1673 </ul><a name="rw_vectorization"><h3><hr size=0>Vectorized Architectures</h3><ul>
1676 <a name="rw_intel_simd">
1677 <dt>Intel MMX, MMX2, SSE, SSE2
1678 <DD>Description here<p>
1680 <a name="rw_amd_simd">
1681 <dt><a href="http://www.nondot.org/~sabre/os/H1ChipFeatures/3DNow!TechnologyManual.pdf">AMD 3Dnow!, 3Dnow! 2</a>
1682 <DD>Desciption here<p>
1684 <a name="rw_sun_simd">
1685 <dt><a href="http://www.nondot.org/~sabre/os/H1ChipFeatures/VISInstructionSetUsersManual.pdf">Sun VIS ISA</a>
1686 <DD>Desciption here<p>
1688 <a name="rw_powerpc_simd">
1690 <DD>Desciption here<p>
1699 <!-- *********************************************************************** -->
1701 <!-- *********************************************************************** -->
1706 <address><a href="mailto:sabre@nondot.org">Chris Lattner</a></address>
1707 <!-- Created: Tue Jan 23 15:19:28 CST 2001 -->
1708 <!-- hhmts start -->
1709 Last modified: Mon Nov 4 18:20:38 CST 2002