1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <title>LLVM Assembly Language Reference Manual</title>
6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
8 <meta name="description"
9 content="LLVM Assembly Language Reference Manual.">
10 <link rel="stylesheet" href="llvm.css" type="text/css">
15 <h1>LLVM Language Reference Manual</h1>
17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
20 <li><a href="#highlevel">High Level Structure</a>
22 <li><a href="#modulestructure">Module Structure</a></li>
23 <li><a href="#linkage">Linkage Types</a>
25 <li><a href="#linkage_private">'<tt>private</tt>' Linkage</a></li>
26 <li><a href="#linkage_linker_private">'<tt>linker_private</tt>' Linkage</a></li>
27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
28 <li><a href="#linkage_linker_private_weak_def_auto">'<tt>linker_private_weak_def_auto</tt>' Linkage</a></li>
29 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
30 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
31 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
32 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
33 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
34 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
35 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
36 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
37 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
38 <li><a href="#linkage_external">'<tt>externally visible</tt>' Linkage</a></li>
39 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
40 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
43 <li><a href="#callingconv">Calling Conventions</a></li>
44 <li><a href="#namedtypes">Named Types</a></li>
45 <li><a href="#globalvars">Global Variables</a></li>
46 <li><a href="#functionstructure">Functions</a></li>
47 <li><a href="#aliasstructure">Aliases</a></li>
48 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
49 <li><a href="#paramattrs">Parameter Attributes</a></li>
50 <li><a href="#fnattrs">Function Attributes</a></li>
51 <li><a href="#gc">Garbage Collector Names</a></li>
52 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
53 <li><a href="#datalayout">Data Layout</a></li>
54 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
55 <li><a href="#volatile">Volatile Memory Accesses</a></li>
56 <li><a href="#memmodel">Memory Model for Concurrent Operations</a></li>
57 <li><a href="#ordering">Atomic Memory Ordering Constraints</a></li>
60 <li><a href="#typesystem">Type System</a>
62 <li><a href="#t_classifications">Type Classifications</a></li>
63 <li><a href="#t_primitive">Primitive Types</a>
65 <li><a href="#t_integer">Integer Type</a></li>
66 <li><a href="#t_floating">Floating Point Types</a></li>
67 <li><a href="#t_x86mmx">X86mmx Type</a></li>
68 <li><a href="#t_void">Void Type</a></li>
69 <li><a href="#t_label">Label Type</a></li>
70 <li><a href="#t_metadata">Metadata Type</a></li>
73 <li><a href="#t_derived">Derived Types</a>
75 <li><a href="#t_aggregate">Aggregate Types</a>
77 <li><a href="#t_array">Array Type</a></li>
78 <li><a href="#t_struct">Structure Type</a></li>
79 <li><a href="#t_opaque">Opaque Structure Types</a></li>
80 <li><a href="#t_vector">Vector Type</a></li>
83 <li><a href="#t_function">Function Type</a></li>
84 <li><a href="#t_pointer">Pointer Type</a></li>
89 <li><a href="#constants">Constants</a>
91 <li><a href="#simpleconstants">Simple Constants</a></li>
92 <li><a href="#complexconstants">Complex Constants</a></li>
93 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
94 <li><a href="#undefvalues">Undefined Values</a></li>
95 <li><a href="#trapvalues">Trap Values</a></li>
96 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
97 <li><a href="#constantexprs">Constant Expressions</a></li>
100 <li><a href="#othervalues">Other Values</a>
102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
106 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
108 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
109 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
110 Global Variable</a></li>
111 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
112 Global Variable</a></li>
113 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
114 Global Variable</a></li>
117 <li><a href="#instref">Instruction Reference</a>
119 <li><a href="#terminators">Terminator Instructions</a>
121 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
122 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
123 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
124 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
125 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
126 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
127 <li><a href="#i_resume">'<tt>resume</tt>' Instruction</a></li>
128 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
131 <li><a href="#binaryops">Binary Operations</a>
133 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
134 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
135 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
136 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
137 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
138 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
139 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
140 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
141 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
142 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
143 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
144 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
147 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
149 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
150 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
151 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
152 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
153 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
154 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
157 <li><a href="#vectorops">Vector Operations</a>
159 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
160 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
161 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
164 <li><a href="#aggregateops">Aggregate Operations</a>
166 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
167 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
170 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
172 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
173 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
174 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
175 <li><a href="#i_fence">'<tt>fence</tt>' Instruction</a></li>
176 <li><a href="#i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a></li>
177 <li><a href="#i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a></li>
178 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
181 <li><a href="#convertops">Conversion Operations</a>
183 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
184 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
185 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
186 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
187 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
188 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
189 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
190 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
191 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
192 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
193 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
194 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
197 <li><a href="#otherops">Other Operations</a>
199 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
200 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
201 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
202 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
203 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
204 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
209 <li><a href="#intrinsics">Intrinsic Functions</a>
211 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
213 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
214 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
215 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
218 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
220 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
221 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
222 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
225 <li><a href="#int_codegen">Code Generator Intrinsics</a>
227 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
228 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
229 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
230 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
231 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
232 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
233 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
236 <li><a href="#int_libc">Standard C Library Intrinsics</a>
238 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
239 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
241 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
242 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
243 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
244 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
245 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
246 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
247 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
248 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
251 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
253 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
254 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
255 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
256 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
259 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
261 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
262 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
263 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
264 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
265 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
266 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
269 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
271 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
272 <li><a href="#int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a></li>
275 <li><a href="#int_debugger">Debugger intrinsics</a></li>
276 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
277 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
279 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
282 <li><a href="#int_atomics">Atomic intrinsics</a>
284 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
285 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
286 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
287 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
288 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
289 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
290 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
291 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
292 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
293 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
294 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
295 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
296 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
299 <li><a href="#int_memorymarkers">Memory Use Markers</a>
301 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
302 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
303 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
304 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
307 <li><a href="#int_general">General intrinsics</a>
309 <li><a href="#int_var_annotation">
310 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
311 <li><a href="#int_annotation">
312 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
313 <li><a href="#int_trap">
314 '<tt>llvm.trap</tt>' Intrinsic</a></li>
315 <li><a href="#int_stackprotector">
316 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
317 <li><a href="#int_objectsize">
318 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
325 <div class="doc_author">
326 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
327 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
330 <!-- *********************************************************************** -->
331 <h2><a name="abstract">Abstract</a></h2>
332 <!-- *********************************************************************** -->
336 <p>This document is a reference manual for the LLVM assembly language. LLVM is
337 a Static Single Assignment (SSA) based representation that provides type
338 safety, low-level operations, flexibility, and the capability of representing
339 'all' high-level languages cleanly. It is the common code representation
340 used throughout all phases of the LLVM compilation strategy.</p>
344 <!-- *********************************************************************** -->
345 <h2><a name="introduction">Introduction</a></h2>
346 <!-- *********************************************************************** -->
350 <p>The LLVM code representation is designed to be used in three different forms:
351 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
352 for fast loading by a Just-In-Time compiler), and as a human readable
353 assembly language representation. This allows LLVM to provide a powerful
354 intermediate representation for efficient compiler transformations and
355 analysis, while providing a natural means to debug and visualize the
356 transformations. The three different forms of LLVM are all equivalent. This
357 document describes the human readable representation and notation.</p>
359 <p>The LLVM representation aims to be light-weight and low-level while being
360 expressive, typed, and extensible at the same time. It aims to be a
361 "universal IR" of sorts, by being at a low enough level that high-level ideas
362 may be cleanly mapped to it (similar to how microprocessors are "universal
363 IR's", allowing many source languages to be mapped to them). By providing
364 type information, LLVM can be used as the target of optimizations: for
365 example, through pointer analysis, it can be proven that a C automatic
366 variable is never accessed outside of the current function, allowing it to
367 be promoted to a simple SSA value instead of a memory location.</p>
369 <!-- _______________________________________________________________________ -->
371 <a name="wellformed">Well-Formedness</a>
376 <p>It is important to note that this document describes 'well formed' LLVM
377 assembly language. There is a difference between what the parser accepts and
378 what is considered 'well formed'. For example, the following instruction is
379 syntactically okay, but not well formed:</p>
381 <pre class="doc_code">
382 %x = <a href="#i_add">add</a> i32 1, %x
385 <p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
386 LLVM infrastructure provides a verification pass that may be used to verify
387 that an LLVM module is well formed. This pass is automatically run by the
388 parser after parsing input assembly and by the optimizer before it outputs
389 bitcode. The violations pointed out by the verifier pass indicate bugs in
390 transformation passes or input to the parser.</p>
396 <!-- Describe the typesetting conventions here. -->
398 <!-- *********************************************************************** -->
399 <h2><a name="identifiers">Identifiers</a></h2>
400 <!-- *********************************************************************** -->
404 <p>LLVM identifiers come in two basic types: global and local. Global
405 identifiers (functions, global variables) begin with the <tt>'@'</tt>
406 character. Local identifiers (register names, types) begin with
407 the <tt>'%'</tt> character. Additionally, there are three different formats
408 for identifiers, for different purposes:</p>
411 <li>Named values are represented as a string of characters with their prefix.
412 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
413 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
414 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
415 other characters in their names can be surrounded with quotes. Special
416 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
417 ASCII code for the character in hexadecimal. In this way, any character
418 can be used in a name value, even quotes themselves.</li>
420 <li>Unnamed values are represented as an unsigned numeric value with their
421 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
423 <li>Constants, which are described in a <a href="#constants">section about
424 constants</a>, below.</li>
427 <p>LLVM requires that values start with a prefix for two reasons: Compilers
428 don't need to worry about name clashes with reserved words, and the set of
429 reserved words may be expanded in the future without penalty. Additionally,
430 unnamed identifiers allow a compiler to quickly come up with a temporary
431 variable without having to avoid symbol table conflicts.</p>
433 <p>Reserved words in LLVM are very similar to reserved words in other
434 languages. There are keywords for different opcodes
435 ('<tt><a href="#i_add">add</a></tt>',
436 '<tt><a href="#i_bitcast">bitcast</a></tt>',
437 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
438 ('<tt><a href="#t_void">void</a></tt>',
439 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
440 reserved words cannot conflict with variable names, because none of them
441 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
443 <p>Here is an example of LLVM code to multiply the integer variable
444 '<tt>%X</tt>' by 8:</p>
448 <pre class="doc_code">
449 %result = <a href="#i_mul">mul</a> i32 %X, 8
452 <p>After strength reduction:</p>
454 <pre class="doc_code">
455 %result = <a href="#i_shl">shl</a> i32 %X, i8 3
458 <p>And the hard way:</p>
460 <pre class="doc_code">
461 %0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
462 %1 = <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
463 %result = <a href="#i_add">add</a> i32 %1, %1
466 <p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
467 lexical features of LLVM:</p>
470 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
473 <li>Unnamed temporaries are created when the result of a computation is not
474 assigned to a named value.</li>
476 <li>Unnamed temporaries are numbered sequentially</li>
479 <p>It also shows a convention that we follow in this document. When
480 demonstrating instructions, we will follow an instruction with a comment that
481 defines the type and name of value produced. Comments are shown in italic
486 <!-- *********************************************************************** -->
487 <h2><a name="highlevel">High Level Structure</a></h2>
488 <!-- *********************************************************************** -->
490 <!-- ======================================================================= -->
492 <a name="modulestructure">Module Structure</a>
497 <p>LLVM programs are composed of "Module"s, each of which is a translation unit
498 of the input programs. Each module consists of functions, global variables,
499 and symbol table entries. Modules may be combined together with the LLVM
500 linker, which merges function (and global variable) definitions, resolves
501 forward declarations, and merges symbol table entries. Here is an example of
502 the "hello world" module:</p>
504 <pre class="doc_code">
505 <i>; Declare the string constant as a global constant.</i>
506 <a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
508 <i>; External declaration of the puts function</i>
509 <a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>
511 <i>; Definition of main function</i>
512 define i32 @main() { <i>; i32()* </i>
513 <i>; Convert [13 x i8]* to i8 *...</i>
514 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8*</i>
516 <i>; Call puts function to write out the string to stdout.</i>
517 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>
518 <a href="#i_ret">ret</a> i32 0
521 <i>; Named metadata</i>
522 !1 = metadata !{i32 41}
526 <p>This example is made up of a <a href="#globalvars">global variable</a> named
527 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
528 a <a href="#functionstructure">function definition</a> for
529 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
532 <p>In general, a module is made up of a list of global values, where both
533 functions and global variables are global values. Global values are
534 represented by a pointer to a memory location (in this case, a pointer to an
535 array of char, and a pointer to a function), and have one of the
536 following <a href="#linkage">linkage types</a>.</p>
540 <!-- ======================================================================= -->
542 <a name="linkage">Linkage Types</a>
547 <p>All Global Variables and Functions have one of the following types of
551 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
552 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
553 by objects in the current module. In particular, linking code into a
554 module with an private global value may cause the private to be renamed as
555 necessary to avoid collisions. Because the symbol is private to the
556 module, all references can be updated. This doesn't show up in any symbol
557 table in the object file.</dd>
559 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
560 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
561 assembler and evaluated by the linker. Unlike normal strong symbols, they
562 are removed by the linker from the final linked image (executable or
563 dynamic library).</dd>
565 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
566 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
567 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
568 linker. The symbols are removed by the linker from the final linked image
569 (executable or dynamic library).</dd>
571 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
572 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
573 of the object is not taken. For instance, functions that had an inline
574 definition, but the compiler decided not to inline it. Note,
575 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
576 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
577 visibility. The symbols are removed by the linker from the final linked
578 image (executable or dynamic library).</dd>
580 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
581 <dd>Similar to private, but the value shows as a local symbol
582 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
583 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
585 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
586 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
587 into the object file corresponding to the LLVM module. They exist to
588 allow inlining and other optimizations to take place given knowledge of
589 the definition of the global, which is known to be somewhere outside the
590 module. Globals with <tt>available_externally</tt> linkage are allowed to
591 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
592 This linkage type is only allowed on definitions, not declarations.</dd>
594 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
595 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
596 the same name when linkage occurs. This can be used to implement
597 some forms of inline functions, templates, or other code which must be
598 generated in each translation unit that uses it, but where the body may
599 be overridden with a more definitive definition later. Unreferenced
600 <tt>linkonce</tt> globals are allowed to be discarded. Note that
601 <tt>linkonce</tt> linkage does not actually allow the optimizer to
602 inline the body of this function into callers because it doesn't know if
603 this definition of the function is the definitive definition within the
604 program or whether it will be overridden by a stronger definition.
605 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
608 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
609 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
610 <tt>linkonce</tt> linkage, except that unreferenced globals with
611 <tt>weak</tt> linkage may not be discarded. This is used for globals that
612 are declared "weak" in C source code.</dd>
614 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
615 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
616 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
618 Symbols with "<tt>common</tt>" linkage are merged in the same way as
619 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
620 <tt>common</tt> symbols may not have an explicit section,
621 must have a zero initializer, and may not be marked '<a
622 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
623 have common linkage.</dd>
626 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
627 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
628 pointer to array type. When two global variables with appending linkage
629 are linked together, the two global arrays are appended together. This is
630 the LLVM, typesafe, equivalent of having the system linker append together
631 "sections" with identical names when .o files are linked.</dd>
633 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
634 <dd>The semantics of this linkage follow the ELF object file model: the symbol
635 is weak until linked, if not linked, the symbol becomes null instead of
636 being an undefined reference.</dd>
638 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
639 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
640 <dd>Some languages allow differing globals to be merged, such as two functions
641 with different semantics. Other languages, such as <tt>C++</tt>, ensure
642 that only equivalent globals are ever merged (the "one definition rule"
643 — "ODR"). Such languages can use the <tt>linkonce_odr</tt>
644 and <tt>weak_odr</tt> linkage types to indicate that the global will only
645 be merged with equivalent globals. These linkage types are otherwise the
646 same as their non-<tt>odr</tt> versions.</dd>
648 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
649 <dd>If none of the above identifiers are used, the global is externally
650 visible, meaning that it participates in linkage and can be used to
651 resolve external symbol references.</dd>
654 <p>The next two types of linkage are targeted for Microsoft Windows platform
655 only. They are designed to support importing (exporting) symbols from (to)
656 DLLs (Dynamic Link Libraries).</p>
659 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
660 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
661 or variable via a global pointer to a pointer that is set up by the DLL
662 exporting the symbol. On Microsoft Windows targets, the pointer name is
663 formed by combining <code>__imp_</code> and the function or variable
666 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
667 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
668 pointer to a pointer in a DLL, so that it can be referenced with the
669 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
670 name is formed by combining <code>__imp_</code> and the function or
674 <p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
675 another module defined a "<tt>.LC0</tt>" variable and was linked with this
676 one, one of the two would be renamed, preventing a collision. Since
677 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
678 declarations), they are accessible outside of the current module.</p>
680 <p>It is illegal for a function <i>declaration</i> to have any linkage type
681 other than "externally visible", <tt>dllimport</tt>
682 or <tt>extern_weak</tt>.</p>
684 <p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
685 or <tt>weak_odr</tt> linkages.</p>
689 <!-- ======================================================================= -->
691 <a name="callingconv">Calling Conventions</a>
696 <p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
697 and <a href="#i_invoke">invokes</a> can all have an optional calling
698 convention specified for the call. The calling convention of any pair of
699 dynamic caller/callee must match, or the behavior of the program is
700 undefined. The following calling conventions are supported by LLVM, and more
701 may be added in the future:</p>
704 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
705 <dd>This calling convention (the default if no other calling convention is
706 specified) matches the target C calling conventions. This calling
707 convention supports varargs function calls and tolerates some mismatch in
708 the declared prototype and implemented declaration of the function (as
711 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
712 <dd>This calling convention attempts to make calls as fast as possible
713 (e.g. by passing things in registers). This calling convention allows the
714 target to use whatever tricks it wants to produce fast code for the
715 target, without having to conform to an externally specified ABI
716 (Application Binary Interface).
717 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
718 when this or the GHC convention is used.</a> This calling convention
719 does not support varargs and requires the prototype of all callees to
720 exactly match the prototype of the function definition.</dd>
722 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
723 <dd>This calling convention attempts to make code in the caller as efficient
724 as possible under the assumption that the call is not commonly executed.
725 As such, these calls often preserve all registers so that the call does
726 not break any live ranges in the caller side. This calling convention
727 does not support varargs and requires the prototype of all callees to
728 exactly match the prototype of the function definition.</dd>
730 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
731 <dd>This calling convention has been implemented specifically for use by the
732 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
733 It passes everything in registers, going to extremes to achieve this by
734 disabling callee save registers. This calling convention should not be
735 used lightly but only for specific situations such as an alternative to
736 the <em>register pinning</em> performance technique often used when
737 implementing functional programming languages.At the moment only X86
738 supports this convention and it has the following limitations:
740 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
741 floating point types are supported.</li>
742 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
743 6 floating point parameters.</li>
745 This calling convention supports
746 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
747 requires both the caller and callee are using it.
750 <dt><b>"<tt>cc <<em>n</em>></tt>" - Numbered convention</b>:</dt>
751 <dd>Any calling convention may be specified by number, allowing
752 target-specific calling conventions to be used. Target specific calling
753 conventions start at 64.</dd>
756 <p>More calling conventions can be added/defined on an as-needed basis, to
757 support Pascal conventions or any other well-known target-independent
762 <!-- ======================================================================= -->
764 <a name="visibility">Visibility Styles</a>
769 <p>All Global Variables and Functions have one of the following visibility
773 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
774 <dd>On targets that use the ELF object file format, default visibility means
775 that the declaration is visible to other modules and, in shared libraries,
776 means that the declared entity may be overridden. On Darwin, default
777 visibility means that the declaration is visible to other modules. Default
778 visibility corresponds to "external linkage" in the language.</dd>
780 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
781 <dd>Two declarations of an object with hidden visibility refer to the same
782 object if they are in the same shared object. Usually, hidden visibility
783 indicates that the symbol will not be placed into the dynamic symbol
784 table, so no other module (executable or shared library) can reference it
787 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
788 <dd>On ELF, protected visibility indicates that the symbol will be placed in
789 the dynamic symbol table, but that references within the defining module
790 will bind to the local symbol. That is, the symbol cannot be overridden by
796 <!-- ======================================================================= -->
798 <a name="namedtypes">Named Types</a>
803 <p>LLVM IR allows you to specify name aliases for certain types. This can make
804 it easier to read the IR and make the IR more condensed (particularly when
805 recursive types are involved). An example of a name specification is:</p>
807 <pre class="doc_code">
808 %mytype = type { %mytype*, i32 }
811 <p>You may give a name to any <a href="#typesystem">type</a> except
812 "<a href="#t_void">void</a>". Type name aliases may be used anywhere a type
813 is expected with the syntax "%mytype".</p>
815 <p>Note that type names are aliases for the structural type that they indicate,
816 and that you can therefore specify multiple names for the same type. This
817 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
818 uses structural typing, the name is not part of the type. When printing out
819 LLVM IR, the printer will pick <em>one name</em> to render all types of a
820 particular shape. This means that if you have code where two different
821 source types end up having the same LLVM type, that the dumper will sometimes
822 print the "wrong" or unexpected type. This is an important design point and
823 isn't going to change.</p>
827 <!-- ======================================================================= -->
829 <a name="globalvars">Global Variables</a>
834 <p>Global variables define regions of memory allocated at compilation time
835 instead of run-time. Global variables may optionally be initialized, may
836 have an explicit section to be placed in, and may have an optional explicit
837 alignment specified. A variable may be defined as "thread_local", which
838 means that it will not be shared by threads (each thread will have a
839 separated copy of the variable). A variable may be defined as a global
840 "constant," which indicates that the contents of the variable
841 will <b>never</b> be modified (enabling better optimization, allowing the
842 global data to be placed in the read-only section of an executable, etc).
843 Note that variables that need runtime initialization cannot be marked
844 "constant" as there is a store to the variable.</p>
846 <p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
847 constant, even if the final definition of the global is not. This capability
848 can be used to enable slightly better optimization of the program, but
849 requires the language definition to guarantee that optimizations based on the
850 'constantness' are valid for the translation units that do not include the
853 <p>As SSA values, global variables define pointer values that are in scope
854 (i.e. they dominate) all basic blocks in the program. Global variables
855 always define a pointer to their "content" type because they describe a
856 region of memory, and all memory objects in LLVM are accessed through
859 <p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
860 that the address is not significant, only the content. Constants marked
861 like this can be merged with other constants if they have the same
862 initializer. Note that a constant with significant address <em>can</em>
863 be merged with a <tt>unnamed_addr</tt> constant, the result being a
864 constant whose address is significant.</p>
866 <p>A global variable may be declared to reside in a target-specific numbered
867 address space. For targets that support them, address spaces may affect how
868 optimizations are performed and/or what target instructions are used to
869 access the variable. The default address space is zero. The address space
870 qualifier must precede any other attributes.</p>
872 <p>LLVM allows an explicit section to be specified for globals. If the target
873 supports it, it will emit globals to the section specified.</p>
875 <p>An explicit alignment may be specified for a global, which must be a power
876 of 2. If not present, or if the alignment is set to zero, the alignment of
877 the global is set by the target to whatever it feels convenient. If an
878 explicit alignment is specified, the global is forced to have exactly that
879 alignment. Targets and optimizers are not allowed to over-align the global
880 if the global has an assigned section. In this case, the extra alignment
881 could be observable: for example, code could assume that the globals are
882 densely packed in their section and try to iterate over them as an array,
883 alignment padding would break this iteration.</p>
885 <p>For example, the following defines a global in a numbered address space with
886 an initializer, section, and alignment:</p>
888 <pre class="doc_code">
889 @G = addrspace(5) constant float 1.0, section "foo", align 4
895 <!-- ======================================================================= -->
897 <a name="functionstructure">Functions</a>
902 <p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
903 optional <a href="#linkage">linkage type</a>, an optional
904 <a href="#visibility">visibility style</a>, an optional
905 <a href="#callingconv">calling convention</a>,
906 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
907 <a href="#paramattrs">parameter attribute</a> for the return type, a function
908 name, a (possibly empty) argument list (each with optional
909 <a href="#paramattrs">parameter attributes</a>), optional
910 <a href="#fnattrs">function attributes</a>, an optional section, an optional
911 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
912 curly brace, a list of basic blocks, and a closing curly brace.</p>
914 <p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
915 optional <a href="#linkage">linkage type</a>, an optional
916 <a href="#visibility">visibility style</a>, an optional
917 <a href="#callingconv">calling convention</a>,
918 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
919 <a href="#paramattrs">parameter attribute</a> for the return type, a function
920 name, a possibly empty list of arguments, an optional alignment, and an
921 optional <a href="#gc">garbage collector name</a>.</p>
923 <p>A function definition contains a list of basic blocks, forming the CFG
924 (Control Flow Graph) for the function. Each basic block may optionally start
925 with a label (giving the basic block a symbol table entry), contains a list
926 of instructions, and ends with a <a href="#terminators">terminator</a>
927 instruction (such as a branch or function return).</p>
929 <p>The first basic block in a function is special in two ways: it is immediately
930 executed on entrance to the function, and it is not allowed to have
931 predecessor basic blocks (i.e. there can not be any branches to the entry
932 block of a function). Because the block can have no predecessors, it also
933 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
935 <p>LLVM allows an explicit section to be specified for functions. If the target
936 supports it, it will emit functions to the section specified.</p>
938 <p>An explicit alignment may be specified for a function. If not present, or if
939 the alignment is set to zero, the alignment of the function is set by the
940 target to whatever it feels convenient. If an explicit alignment is
941 specified, the function is forced to have at least that much alignment. All
942 alignments must be a power of 2.</p>
944 <p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
945 be significant and two identical functions can be merged</p>.
948 <pre class="doc_code">
949 define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
950 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
951 <ResultType> @<FunctionName> ([argument list])
952 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
953 [<a href="#gc">gc</a>] { ... }
958 <!-- ======================================================================= -->
960 <a name="aliasstructure">Aliases</a>
965 <p>Aliases act as "second name" for the aliasee value (which can be either
966 function, global variable, another alias or bitcast of global value). Aliases
967 may have an optional <a href="#linkage">linkage type</a>, and an
968 optional <a href="#visibility">visibility style</a>.</p>
971 <pre class="doc_code">
972 @<Name> = alias [Linkage] [Visibility] <AliaseeTy> @<Aliasee>
977 <!-- ======================================================================= -->
979 <a name="namedmetadatastructure">Named Metadata</a>
984 <p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
985 nodes</a> (but not metadata strings) are the only valid operands for
986 a named metadata.</p>
989 <pre class="doc_code">
990 ; Some unnamed metadata nodes, which are referenced by the named metadata.
991 !0 = metadata !{metadata !"zero"}
992 !1 = metadata !{metadata !"one"}
993 !2 = metadata !{metadata !"two"}
995 !name = !{!0, !1, !2}
1000 <!-- ======================================================================= -->
1002 <a name="paramattrs">Parameter Attributes</a>
1007 <p>The return type and each parameter of a function type may have a set of
1008 <i>parameter attributes</i> associated with them. Parameter attributes are
1009 used to communicate additional information about the result or parameters of
1010 a function. Parameter attributes are considered to be part of the function,
1011 not of the function type, so functions with different parameter attributes
1012 can have the same function type.</p>
1014 <p>Parameter attributes are simple keywords that follow the type specified. If
1015 multiple parameter attributes are needed, they are space separated. For
1018 <pre class="doc_code">
1019 declare i32 @printf(i8* noalias nocapture, ...)
1020 declare i32 @atoi(i8 zeroext)
1021 declare signext i8 @returns_signed_char()
1024 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
1025 <tt>readonly</tt>) come immediately after the argument list.</p>
1027 <p>Currently, only the following parameter attributes are defined:</p>
1030 <dt><tt><b>zeroext</b></tt></dt>
1031 <dd>This indicates to the code generator that the parameter or return value
1032 should be zero-extended to the extent required by the target's ABI (which
1033 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1034 parameter) or the callee (for a return value).</dd>
1036 <dt><tt><b>signext</b></tt></dt>
1037 <dd>This indicates to the code generator that the parameter or return value
1038 should be sign-extended to the extent required by the target's ABI (which
1039 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1042 <dt><tt><b>inreg</b></tt></dt>
1043 <dd>This indicates that this parameter or return value should be treated in a
1044 special target-dependent fashion during while emitting code for a function
1045 call or return (usually, by putting it in a register as opposed to memory,
1046 though some targets use it to distinguish between two different kinds of
1047 registers). Use of this attribute is target-specific.</dd>
1049 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
1050 <dd><p>This indicates that the pointer parameter should really be passed by
1051 value to the function. The attribute implies that a hidden copy of the
1053 is made between the caller and the callee, so the callee is unable to
1054 modify the value in the callee. This attribute is only valid on LLVM
1055 pointer arguments. It is generally used to pass structs and arrays by
1056 value, but is also valid on pointers to scalars. The copy is considered
1057 to belong to the caller not the callee (for example,
1058 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1059 <tt>byval</tt> parameters). This is not a valid attribute for return
1062 <p>The byval attribute also supports specifying an alignment with
1063 the align attribute. It indicates the alignment of the stack slot to
1064 form and the known alignment of the pointer specified to the call site. If
1065 the alignment is not specified, then the code generator makes a
1066 target-specific assumption.</p></dd>
1068 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
1069 <dd>This indicates that the pointer parameter specifies the address of a
1070 structure that is the return value of the function in the source program.
1071 This pointer must be guaranteed by the caller to be valid: loads and
1072 stores to the structure may be assumed by the callee to not to trap. This
1073 may only be applied to the first parameter. This is not a valid attribute
1074 for return values. </dd>
1076 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
1077 <dd>This indicates that pointer values
1078 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
1079 value do not alias pointer values which are not <i>based</i> on it,
1080 ignoring certain "irrelevant" dependencies.
1081 For a call to the parent function, dependencies between memory
1082 references from before or after the call and from those during the call
1083 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1084 return value used in that call.
1085 The caller shares the responsibility with the callee for ensuring that
1086 these requirements are met.
1087 For further details, please see the discussion of the NoAlias response in
1088 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1090 Note that this definition of <tt>noalias</tt> is intentionally
1091 similar to the definition of <tt>restrict</tt> in C99 for function
1092 arguments, though it is slightly weaker.
1094 For function return values, C99's <tt>restrict</tt> is not meaningful,
1095 while LLVM's <tt>noalias</tt> is.
1098 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
1099 <dd>This indicates that the callee does not make any copies of the pointer
1100 that outlive the callee itself. This is not a valid attribute for return
1103 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
1104 <dd>This indicates that the pointer parameter can be excised using the
1105 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1106 attribute for return values.</dd>
1111 <!-- ======================================================================= -->
1113 <a name="gc">Garbage Collector Names</a>
1118 <p>Each function may specify a garbage collector name, which is simply a
1121 <pre class="doc_code">
1122 define void @f() gc "name" { ... }
1125 <p>The compiler declares the supported values of <i>name</i>. Specifying a
1126 collector which will cause the compiler to alter its output in order to
1127 support the named garbage collection algorithm.</p>
1131 <!-- ======================================================================= -->
1133 <a name="fnattrs">Function Attributes</a>
1138 <p>Function attributes are set to communicate additional information about a
1139 function. Function attributes are considered to be part of the function, not
1140 of the function type, so functions with different parameter attributes can
1141 have the same function type.</p>
1143 <p>Function attributes are simple keywords that follow the type specified. If
1144 multiple attributes are needed, they are space separated. For example:</p>
1146 <pre class="doc_code">
1147 define void @f() noinline { ... }
1148 define void @f() alwaysinline { ... }
1149 define void @f() alwaysinline optsize { ... }
1150 define void @f() optsize { ... }
1154 <dt><tt><b>alignstack(<<em>n</em>>)</b></tt></dt>
1155 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1156 the backend should forcibly align the stack pointer. Specify the
1157 desired alignment, which must be a power of two, in parentheses.
1159 <dt><tt><b>alwaysinline</b></tt></dt>
1160 <dd>This attribute indicates that the inliner should attempt to inline this
1161 function into callers whenever possible, ignoring any active inlining size
1162 threshold for this caller.</dd>
1164 <dt><tt><b>hotpatch</b></tt></dt>
1165 <dd>This attribute indicates that the function should be 'hotpatchable',
1166 meaning the function can be patched and/or hooked even while it is
1167 loaded into memory. On x86, the function prologue will be preceded
1168 by six bytes of padding and will begin with a two-byte instruction.
1169 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1170 higher were compiled in this fashion.</dd>
1172 <dt><tt><b>nonlazybind</b></tt></dt>
1173 <dd>This attribute suppresses lazy symbol binding for the function. This
1174 may make calls to the function faster, at the cost of extra program
1175 startup time if the function is not called during program startup.</dd>
1177 <dt><tt><b>inlinehint</b></tt></dt>
1178 <dd>This attribute indicates that the source code contained a hint that inlining
1179 this function is desirable (such as the "inline" keyword in C/C++). It
1180 is just a hint; it imposes no requirements on the inliner.</dd>
1182 <dt><tt><b>naked</b></tt></dt>
1183 <dd>This attribute disables prologue / epilogue emission for the function.
1184 This can have very system-specific consequences.</dd>
1186 <dt><tt><b>noimplicitfloat</b></tt></dt>
1187 <dd>This attributes disables implicit floating point instructions.</dd>
1189 <dt><tt><b>noinline</b></tt></dt>
1190 <dd>This attribute indicates that the inliner should never inline this
1191 function in any situation. This attribute may not be used together with
1192 the <tt>alwaysinline</tt> attribute.</dd>
1194 <dt><tt><b>noredzone</b></tt></dt>
1195 <dd>This attribute indicates that the code generator should not use a red
1196 zone, even if the target-specific ABI normally permits it.</dd>
1198 <dt><tt><b>noreturn</b></tt></dt>
1199 <dd>This function attribute indicates that the function never returns
1200 normally. This produces undefined behavior at runtime if the function
1201 ever does dynamically return.</dd>
1203 <dt><tt><b>nounwind</b></tt></dt>
1204 <dd>This function attribute indicates that the function never returns with an
1205 unwind or exceptional control flow. If the function does unwind, its
1206 runtime behavior is undefined.</dd>
1208 <dt><tt><b>optsize</b></tt></dt>
1209 <dd>This attribute suggests that optimization passes and code generator passes
1210 make choices that keep the code size of this function low, and otherwise
1211 do optimizations specifically to reduce code size.</dd>
1213 <dt><tt><b>readnone</b></tt></dt>
1214 <dd>This attribute indicates that the function computes its result (or decides
1215 to unwind an exception) based strictly on its arguments, without
1216 dereferencing any pointer arguments or otherwise accessing any mutable
1217 state (e.g. memory, control registers, etc) visible to caller functions.
1218 It does not write through any pointer arguments
1219 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1220 changes any state visible to callers. This means that it cannot unwind
1221 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1222 could use the <tt>unwind</tt> instruction.</dd>
1224 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
1225 <dd>This attribute indicates that the function does not write through any
1226 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1227 arguments) or otherwise modify any state (e.g. memory, control registers,
1228 etc) visible to caller functions. It may dereference pointer arguments
1229 and read state that may be set in the caller. A readonly function always
1230 returns the same value (or unwinds an exception identically) when called
1231 with the same set of arguments and global state. It cannot unwind an
1232 exception by calling the <tt>C++</tt> exception throwing methods, but may
1233 use the <tt>unwind</tt> instruction.</dd>
1235 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
1236 <dd>This attribute indicates that the function should emit a stack smashing
1237 protector. It is in the form of a "canary"—a random value placed on
1238 the stack before the local variables that's checked upon return from the
1239 function to see if it has been overwritten. A heuristic is used to
1240 determine if a function needs stack protectors or not.<br>
1242 If a function that has an <tt>ssp</tt> attribute is inlined into a
1243 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1244 function will have an <tt>ssp</tt> attribute.</dd>
1246 <dt><tt><b>sspreq</b></tt></dt>
1247 <dd>This attribute indicates that the function should <em>always</em> emit a
1248 stack smashing protector. This overrides
1249 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1251 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1252 function that doesn't have an <tt>sspreq</tt> attribute or which has
1253 an <tt>ssp</tt> attribute, then the resulting function will have
1254 an <tt>sspreq</tt> attribute.</dd>
1256 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1257 <dd>This attribute indicates that the ABI being targeted requires that
1258 an unwind table entry be produce for this function even if we can
1259 show that no exceptions passes by it. This is normally the case for
1260 the ELF x86-64 abi, but it can be disabled for some compilation
1267 <!-- ======================================================================= -->
1269 <a name="moduleasm">Module-Level Inline Assembly</a>
1274 <p>Modules may contain "module-level inline asm" blocks, which corresponds to
1275 the GCC "file scope inline asm" blocks. These blocks are internally
1276 concatenated by LLVM and treated as a single unit, but may be separated in
1277 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
1279 <pre class="doc_code">
1280 module asm "inline asm code goes here"
1281 module asm "more can go here"
1284 <p>The strings can contain any character by escaping non-printable characters.
1285 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1288 <p>The inline asm code is simply printed to the machine code .s file when
1289 assembly code is generated.</p>
1293 <!-- ======================================================================= -->
1295 <a name="datalayout">Data Layout</a>
1300 <p>A module may specify a target specific data layout string that specifies how
1301 data is to be laid out in memory. The syntax for the data layout is
1304 <pre class="doc_code">
1305 target datalayout = "<i>layout specification</i>"
1308 <p>The <i>layout specification</i> consists of a list of specifications
1309 separated by the minus sign character ('-'). Each specification starts with
1310 a letter and may include other information after the letter to define some
1311 aspect of the data layout. The specifications accepted are as follows:</p>
1315 <dd>Specifies that the target lays out data in big-endian form. That is, the
1316 bits with the most significance have the lowest address location.</dd>
1319 <dd>Specifies that the target lays out data in little-endian form. That is,
1320 the bits with the least significance have the lowest address
1323 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1324 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1325 <i>preferred</i> alignments. All sizes are in bits. Specifying
1326 the <i>pref</i> alignment is optional. If omitted, the
1327 preceding <tt>:</tt> should be omitted too.</dd>
1329 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1330 <dd>This specifies the alignment for an integer type of a given bit
1331 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1333 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1334 <dd>This specifies the alignment for a vector type of a given bit
1337 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1338 <dd>This specifies the alignment for a floating point type of a given bit
1339 <i>size</i>. Only values of <i>size</i> that are supported by the target
1340 will work. 32 (float) and 64 (double) are supported on all targets;
1341 80 or 128 (different flavors of long double) are also supported on some
1344 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1345 <dd>This specifies the alignment for an aggregate type of a given bit
1348 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1349 <dd>This specifies the alignment for a stack object of a given bit
1352 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1353 <dd>This specifies a set of native integer widths for the target CPU
1354 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1355 "n32:64" for PowerPC 64, or "n8:16:32:64" for X86-64. Elements of
1356 this set are considered to support most general arithmetic
1357 operations efficiently.</dd>
1360 <p>When constructing the data layout for a given target, LLVM starts with a
1361 default set of specifications which are then (possibly) overridden by the
1362 specifications in the <tt>datalayout</tt> keyword. The default specifications
1363 are given in this list:</p>
1366 <li><tt>E</tt> - big endian</li>
1367 <li><tt>p:64:64:64</tt> - 64-bit pointers with 64-bit alignment</li>
1368 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1369 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1370 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1371 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
1372 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
1373 alignment of 64-bits</li>
1374 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1375 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1376 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1377 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1378 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1379 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
1382 <p>When LLVM is determining the alignment for a given type, it uses the
1383 following rules:</p>
1386 <li>If the type sought is an exact match for one of the specifications, that
1387 specification is used.</li>
1389 <li>If no match is found, and the type sought is an integer type, then the
1390 smallest integer type that is larger than the bitwidth of the sought type
1391 is used. If none of the specifications are larger than the bitwidth then
1392 the the largest integer type is used. For example, given the default
1393 specifications above, the i7 type will use the alignment of i8 (next
1394 largest) while both i65 and i256 will use the alignment of i64 (largest
1397 <li>If no match is found, and the type sought is a vector type, then the
1398 largest vector type that is smaller than the sought vector type will be
1399 used as a fall back. This happens because <128 x double> can be
1400 implemented in terms of 64 <2 x double>, for example.</li>
1405 <!-- ======================================================================= -->
1407 <a name="pointeraliasing">Pointer Aliasing Rules</a>
1412 <p>Any memory access must be done through a pointer value associated
1413 with an address range of the memory access, otherwise the behavior
1414 is undefined. Pointer values are associated with address ranges
1415 according to the following rules:</p>
1418 <li>A pointer value is associated with the addresses associated with
1419 any value it is <i>based</i> on.
1420 <li>An address of a global variable is associated with the address
1421 range of the variable's storage.</li>
1422 <li>The result value of an allocation instruction is associated with
1423 the address range of the allocated storage.</li>
1424 <li>A null pointer in the default address-space is associated with
1426 <li>An integer constant other than zero or a pointer value returned
1427 from a function not defined within LLVM may be associated with address
1428 ranges allocated through mechanisms other than those provided by
1429 LLVM. Such ranges shall not overlap with any ranges of addresses
1430 allocated by mechanisms provided by LLVM.</li>
1433 <p>A pointer value is <i>based</i> on another pointer value according
1434 to the following rules:</p>
1437 <li>A pointer value formed from a
1438 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1439 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1440 <li>The result value of a
1441 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1442 of the <tt>bitcast</tt>.</li>
1443 <li>A pointer value formed by an
1444 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1445 pointer values that contribute (directly or indirectly) to the
1446 computation of the pointer's value.</li>
1447 <li>The "<i>based</i> on" relationship is transitive.</li>
1450 <p>Note that this definition of <i>"based"</i> is intentionally
1451 similar to the definition of <i>"based"</i> in C99, though it is
1452 slightly weaker.</p>
1454 <p>LLVM IR does not associate types with memory. The result type of a
1455 <tt><a href="#i_load">load</a></tt> merely indicates the size and
1456 alignment of the memory from which to load, as well as the
1457 interpretation of the value. The first operand type of a
1458 <tt><a href="#i_store">store</a></tt> similarly only indicates the size
1459 and alignment of the store.</p>
1461 <p>Consequently, type-based alias analysis, aka TBAA, aka
1462 <tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1463 LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1464 additional information which specialized optimization passes may use
1465 to implement type-based alias analysis.</p>
1469 <!-- ======================================================================= -->
1471 <a name="volatile">Volatile Memory Accesses</a>
1476 <p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1477 href="#i_store"><tt>store</tt></a>s, and <a
1478 href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1479 The optimizers must not change the number of volatile operations or change their
1480 order of execution relative to other volatile operations. The optimizers
1481 <i>may</i> change the order of volatile operations relative to non-volatile
1482 operations. This is not Java's "volatile" and has no cross-thread
1483 synchronization behavior.</p>
1487 <!-- ======================================================================= -->
1489 <a name="memmodel">Memory Model for Concurrent Operations</a>
1494 <p>The LLVM IR does not define any way to start parallel threads of execution
1495 or to register signal handlers. Nonetheless, there are platform-specific
1496 ways to create them, and we define LLVM IR's behavior in their presence. This
1497 model is inspired by the C++0x memory model.</p>
1499 <p>We define a <i>happens-before</i> partial order as the least partial order
1502 <li>Is a superset of single-thread program order, and</li>
1503 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1504 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1505 by platform-specific techniques, like pthread locks, thread
1506 creation, thread joining, etc., and by atomic instructions.
1507 (See also <a href="#ordering">Atomic Memory Ordering Constraints</a>).
1511 <p>Note that program order does not introduce <i>happens-before</i> edges
1512 between a thread and signals executing inside that thread.</p>
1514 <p>Every (defined) read operation (load instructions, memcpy, atomic
1515 loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1516 (defined) write operations (store instructions, atomic
1517 stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1518 initialized globals are considered to have a write of the initializer which is
1519 atomic and happens before any other read or write of the memory in question.
1520 For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1521 any write to the same byte, except:</p>
1524 <li>If <var>write<sub>1</sub></var> happens before
1525 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1526 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
1527 does not see <var>write<sub>1</sub></var>.
1528 <li>If <var>R<sub>byte</sub></var> happens before
1529 <var>write<sub>3</sub></var>, then <var>R<sub>byte</sub></var> does not
1530 see <var>write<sub>3</sub></var>.
1533 <p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1535 <li>If there is no write to the same byte that happens before
1536 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1537 <tt>undef</tt> for that byte.
1538 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
1539 <var>R<sub>byte</sub></var> returns the value written by that
1541 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1542 <var>R<sub>byte</sub></var> may see are atomic, it chooses one of the
1543 values written. See the <a href="#ordering">Atomic Memory Ordering
1544 Constraints</a> section for additional constraints on how the choice
1546 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1549 <p><var>R</var> returns the value composed of the series of bytes it read.
1550 This implies that some bytes within the value may be <tt>undef</tt>
1551 <b>without</b> the entire value being <tt>undef</tt>. Note that this only
1552 defines the semantics of the operation; it doesn't mean that targets will
1553 emit more than one instruction to read the series of bytes.</p>
1555 <p>Note that in cases where none of the atomic intrinsics are used, this model
1556 places only one restriction on IR transformations on top of what is required
1557 for single-threaded execution: introducing a store to a byte which might not
1558 otherwise be stored is not allowed in general. (Specifically, in the case
1559 where another thread might write to and read from an address, introducing a
1560 store can change a load that may see exactly one write into a load that may
1561 see multiple writes.)</p>
1563 <!-- FIXME: This model assumes all targets where concurrency is relevant have
1564 a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1565 none of the backends currently in the tree fall into this category; however,
1566 there might be targets which care. If there are, we want a paragraph
1569 Targets may specify that stores narrower than a certain width are not
1570 available; on such a target, for the purposes of this model, treat any
1571 non-atomic write with an alignment or width less than the minimum width
1572 as if it writes to the relevant surrounding bytes.
1577 <!-- ======================================================================= -->
1578 <div class="doc_subsection">
1579 <a name="ordering">Atomic Memory Ordering Constraints</a>
1582 <div class="doc_text">
1584 <p>Atomic instructions (<a href="#i_cmpxchg"><code>cmpxchg</code></a>,
1585 <a href="#i_atomicrmw"><code>atomicrmw</code></a>, and
1586 <a href="#i_fence"><code>fence</code></a>) take an ordering parameter
1587 that determines which other atomic instructions on the same address they
1588 <i>synchronize with</i>. These semantics are borrowed from Java and C++0x,
1589 but are somewhat more colloquial. If these descriptions aren't precise enough,
1590 check those specs. <a href="#i_fence"><code>fence</code></a> instructions
1591 treat these orderings somewhat differently since they don't take an address.
1592 See that instruction's documentation for details.</p>
1594 <!-- FIXME Note atomic load+store here once those get added. -->
1597 <!-- FIXME: unordered is intended to be used for atomic load and store;
1598 it isn't allowed for any instruction yet. -->
1599 <dt><code>unordered</code></dt>
1600 <dd>The set of values that can be read is governed by the happens-before
1601 partial order. A value cannot be read unless some operation wrote it.
1602 This is intended to provide a guarantee strong enough to model Java's
1603 non-volatile shared variables. This ordering cannot be specified for
1604 read-modify-write operations; it is not strong enough to make them atomic
1605 in any interesting way.</dd>
1606 <dt><code>monotonic</code></dt>
1607 <dd>In addition to the guarantees of <code>unordered</code>, there is a single
1608 total order for modifications by <code>monotonic</code> operations on each
1609 address. All modification orders must be compatible with the happens-before
1610 order. There is no guarantee that the modification orders can be combined to
1611 a global total order for the whole program (and this often will not be
1612 possible). The read in an atomic read-modify-write operation
1613 (<a href="#i_cmpxchg"><code>cmpxchg</code></a> and
1614 <a href="#i_atomicrmw"><code>atomicrmw</code></a>)
1615 reads the value in the modification order immediately before the value it
1616 writes. If one atomic read happens before another atomic read of the same
1617 address, the later read must see the same value or a later value in the
1618 address's modification order. This disallows reordering of
1619 <code>monotonic</code> (or stronger) operations on the same address. If an
1620 address is written <code>monotonic</code>ally by one thread, and other threads
1621 <code>monotonic</code>ally read that address repeatedly, the other threads must
1622 eventually see the write. This is intended to model C++'s relaxed atomic
1624 <dt><code>acquire</code></dt>
1625 <dd>In addition to the guarantees of <code>monotonic</code>, if this operation
1626 reads a value written by a <code>release</code> atomic operation, it
1627 <i>synchronizes-with</i> that operation.</dd>
1628 <dt><code>release</code></dt>
1629 <dd>In addition to the guarantees of <code>monotonic</code>,
1630 a <i>synchronizes-with</i> edge may be formed by an <code>acquire</code>
1632 <dt><code>acq_rel</code> (acquire+release)</dt><dd>Acts as both an
1633 <code>acquire</code> and <code>release</code> operation on its address.</dd>
1634 <dt><code>seq_cst</code> (sequentially consistent)</dt><dd>
1635 <dd>In addition to the guarantees of <code>acq_rel</code>
1636 (<code>acquire</code> for an operation which only reads, <code>release</code>
1637 for an operation which only writes), there is a global total order on all
1638 sequentially-consistent operations on all addresses, which is consistent with
1639 the <i>happens-before</i> partial order and with the modification orders of
1640 all the affected addresses. Each sequentially-consistent read sees the last
1641 preceding write to the same address in this global order. This is intended
1642 to model C++'s sequentially-consistent atomic variables and Java's volatile
1643 shared variables.</dd>
1646 <p id="singlethread">If an atomic operation is marked <code>singlethread</code>,
1647 it only <i>synchronizes with</i> or participates in modification and seq_cst
1648 total orderings with other operations running in the same thread (for example,
1649 in signal handlers).</p>
1655 <!-- *********************************************************************** -->
1656 <h2><a name="typesystem">Type System</a></h2>
1657 <!-- *********************************************************************** -->
1661 <p>The LLVM type system is one of the most important features of the
1662 intermediate representation. Being typed enables a number of optimizations
1663 to be performed on the intermediate representation directly, without having
1664 to do extra analyses on the side before the transformation. A strong type
1665 system makes it easier to read the generated code and enables novel analyses
1666 and transformations that are not feasible to perform on normal three address
1667 code representations.</p>
1669 <!-- ======================================================================= -->
1671 <a name="t_classifications">Type Classifications</a>
1676 <p>The types fall into a few useful classifications:</p>
1678 <table border="1" cellspacing="0" cellpadding="4">
1680 <tr><th>Classification</th><th>Types</th></tr>
1682 <td><a href="#t_integer">integer</a></td>
1683 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
1686 <td><a href="#t_floating">floating point</a></td>
1687 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
1690 <td><a name="t_firstclass">first class</a></td>
1691 <td><a href="#t_integer">integer</a>,
1692 <a href="#t_floating">floating point</a>,
1693 <a href="#t_pointer">pointer</a>,
1694 <a href="#t_vector">vector</a>,
1695 <a href="#t_struct">structure</a>,
1696 <a href="#t_array">array</a>,
1697 <a href="#t_label">label</a>,
1698 <a href="#t_metadata">metadata</a>.
1702 <td><a href="#t_primitive">primitive</a></td>
1703 <td><a href="#t_label">label</a>,
1704 <a href="#t_void">void</a>,
1705 <a href="#t_integer">integer</a>,
1706 <a href="#t_floating">floating point</a>,
1707 <a href="#t_x86mmx">x86mmx</a>,
1708 <a href="#t_metadata">metadata</a>.</td>
1711 <td><a href="#t_derived">derived</a></td>
1712 <td><a href="#t_array">array</a>,
1713 <a href="#t_function">function</a>,
1714 <a href="#t_pointer">pointer</a>,
1715 <a href="#t_struct">structure</a>,
1716 <a href="#t_vector">vector</a>,
1717 <a href="#t_opaque">opaque</a>.
1723 <p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1724 important. Values of these types are the only ones which can be produced by
1729 <!-- ======================================================================= -->
1731 <a name="t_primitive">Primitive Types</a>
1736 <p>The primitive types are the fundamental building blocks of the LLVM
1739 <!-- _______________________________________________________________________ -->
1741 <a name="t_integer">Integer Type</a>
1747 <p>The integer type is a very simple type that simply specifies an arbitrary
1748 bit width for the integer type desired. Any bit width from 1 bit to
1749 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1756 <p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1760 <table class="layout">
1762 <td class="left"><tt>i1</tt></td>
1763 <td class="left">a single-bit integer.</td>
1766 <td class="left"><tt>i32</tt></td>
1767 <td class="left">a 32-bit integer.</td>
1770 <td class="left"><tt>i1942652</tt></td>
1771 <td class="left">a really big integer of over 1 million bits.</td>
1777 <!-- _______________________________________________________________________ -->
1779 <a name="t_floating">Floating Point Types</a>
1786 <tr><th>Type</th><th>Description</th></tr>
1787 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1788 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1789 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1790 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1791 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1797 <!-- _______________________________________________________________________ -->
1799 <a name="t_x86mmx">X86mmx Type</a>
1805 <p>The x86mmx type represents a value held in an MMX register on an x86 machine. The operations allowed on it are quite limited: parameters and return values, load and store, and bitcast. User-specified MMX instructions are represented as intrinsic or asm calls with arguments and/or results of this type. There are no arrays, vectors or constants of this type.</p>
1814 <!-- _______________________________________________________________________ -->
1816 <a name="t_void">Void Type</a>
1822 <p>The void type does not represent any value and has no size.</p>
1831 <!-- _______________________________________________________________________ -->
1833 <a name="t_label">Label Type</a>
1839 <p>The label type represents code labels.</p>
1848 <!-- _______________________________________________________________________ -->
1850 <a name="t_metadata">Metadata Type</a>
1856 <p>The metadata type represents embedded metadata. No derived types may be
1857 created from metadata except for <a href="#t_function">function</a>
1869 <!-- ======================================================================= -->
1871 <a name="t_derived">Derived Types</a>
1876 <p>The real power in LLVM comes from the derived types in the system. This is
1877 what allows a programmer to represent arrays, functions, pointers, and other
1878 useful types. Each of these types contain one or more element types which
1879 may be a primitive type, or another derived type. For example, it is
1880 possible to have a two dimensional array, using an array as the element type
1881 of another array.</p>
1886 <!-- _______________________________________________________________________ -->
1888 <a name="t_aggregate">Aggregate Types</a>
1893 <p>Aggregate Types are a subset of derived types that can contain multiple
1894 member types. <a href="#t_array">Arrays</a>,
1895 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1896 aggregate types.</p>
1900 <!-- _______________________________________________________________________ -->
1902 <a name="t_array">Array Type</a>
1908 <p>The array type is a very simple derived type that arranges elements
1909 sequentially in memory. The array type requires a size (number of elements)
1910 and an underlying data type.</p>
1914 [<# elements> x <elementtype>]
1917 <p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1918 be any type with a size.</p>
1921 <table class="layout">
1923 <td class="left"><tt>[40 x i32]</tt></td>
1924 <td class="left">Array of 40 32-bit integer values.</td>
1927 <td class="left"><tt>[41 x i32]</tt></td>
1928 <td class="left">Array of 41 32-bit integer values.</td>
1931 <td class="left"><tt>[4 x i8]</tt></td>
1932 <td class="left">Array of 4 8-bit integer values.</td>
1935 <p>Here are some examples of multidimensional arrays:</p>
1936 <table class="layout">
1938 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1939 <td class="left">3x4 array of 32-bit integer values.</td>
1942 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1943 <td class="left">12x10 array of single precision floating point values.</td>
1946 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1947 <td class="left">2x3x4 array of 16-bit integer values.</td>
1951 <p>There is no restriction on indexing beyond the end of the array implied by
1952 a static type (though there are restrictions on indexing beyond the bounds
1953 of an allocated object in some cases). This means that single-dimension
1954 'variable sized array' addressing can be implemented in LLVM with a zero
1955 length array type. An implementation of 'pascal style arrays' in LLVM could
1956 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
1960 <!-- _______________________________________________________________________ -->
1962 <a name="t_function">Function Type</a>
1968 <p>The function type can be thought of as a function signature. It consists of
1969 a return type and a list of formal parameter types. The return type of a
1970 function type is a first class type or a void type.</p>
1974 <returntype> (<parameter list>)
1977 <p>...where '<tt><parameter list></tt>' is a comma-separated list of type
1978 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1979 which indicates that the function takes a variable number of arguments.
1980 Variable argument functions can access their arguments with
1981 the <a href="#int_varargs">variable argument handling intrinsic</a>
1982 functions. '<tt><returntype></tt>' is any type except
1983 <a href="#t_label">label</a>.</p>
1986 <table class="layout">
1988 <td class="left"><tt>i32 (i32)</tt></td>
1989 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
1991 </tr><tr class="layout">
1992 <td class="left"><tt>float (i16, i32 *) *
1994 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1995 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1996 returning <tt>float</tt>.
1998 </tr><tr class="layout">
1999 <td class="left"><tt>i32 (i8*, ...)</tt></td>
2000 <td class="left">A vararg function that takes at least one
2001 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
2002 which returns an integer. This is the signature for <tt>printf</tt> in
2005 </tr><tr class="layout">
2006 <td class="left"><tt>{i32, i32} (i32)</tt></td>
2007 <td class="left">A function taking an <tt>i32</tt>, returning a
2008 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
2015 <!-- _______________________________________________________________________ -->
2017 <a name="t_struct">Structure Type</a>
2023 <p>The structure type is used to represent a collection of data members together
2024 in memory. The elements of a structure may be any type that has a size.</p>
2026 <p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
2027 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
2028 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2029 Structures in registers are accessed using the
2030 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
2031 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
2033 <p>Structures may optionally be "packed" structures, which indicate that the
2034 alignment of the struct is one byte, and that there is no padding between
2035 the elements. In non-packed structs, padding between field types is defined
2036 by the target data string to match the underlying processor.</p>
2038 <p>Structures can either be "anonymous" or "named". An anonymous structure is
2039 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) and a named types
2040 are always defined at the top level with a name. Anonmyous types are uniqued
2041 by their contents and can never be recursive since there is no way to write
2042 one. Named types can be recursive.
2047 %T1 = type { <type list> } <i>; Named normal struct type</i>
2048 %T2 = type <{ <type list> }> <i>; Named packed struct type</i>
2052 <table class="layout">
2054 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
2055 <td class="left">A triple of three <tt>i32</tt> values</td>
2058 <td class="left"><tt>{ float, i32 (i32) * }</tt></td>
2059 <td class="left">A pair, where the first element is a <tt>float</tt> and the
2060 second element is a <a href="#t_pointer">pointer</a> to a
2061 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
2062 an <tt>i32</tt>.</td>
2065 <td class="left"><tt><{ i8, i32 }></tt></td>
2066 <td class="left">A packed struct known to be 5 bytes in size.</td>
2072 <!-- _______________________________________________________________________ -->
2074 <a name="t_opaque">Opaque Structure Types</a>
2080 <p>Opaque structure types are used to represent named structure types that do
2081 not have a body specified. This corresponds (for example) to the C notion of
2082 a forward declared structure.</p>
2091 <table class="layout">
2093 <td class="left"><tt>opaque</tt></td>
2094 <td class="left">An opaque type.</td>
2102 <!-- _______________________________________________________________________ -->
2104 <a name="t_pointer">Pointer Type</a>
2110 <p>The pointer type is used to specify memory locations.
2111 Pointers are commonly used to reference objects in memory.</p>
2113 <p>Pointer types may have an optional address space attribute defining the
2114 numbered address space where the pointed-to object resides. The default
2115 address space is number zero. The semantics of non-zero address
2116 spaces are target-specific.</p>
2118 <p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2119 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
2127 <table class="layout">
2129 <td class="left"><tt>[4 x i32]*</tt></td>
2130 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2131 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2134 <td class="left"><tt>i32 (i32*) *</tt></td>
2135 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
2136 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
2140 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2141 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2142 that resides in address space #5.</td>
2148 <!-- _______________________________________________________________________ -->
2150 <a name="t_vector">Vector Type</a>
2156 <p>A vector type is a simple derived type that represents a vector of elements.
2157 Vector types are used when multiple primitive data are operated in parallel
2158 using a single instruction (SIMD). A vector type requires a size (number of
2159 elements) and an underlying primitive data type. Vector types are considered
2160 <a href="#t_firstclass">first class</a>.</p>
2164 < <# elements> x <elementtype> >
2167 <p>The number of elements is a constant integer value larger than 0; elementtype
2168 may be any integer or floating point type. Vectors of size zero are not
2169 allowed, and pointers are not allowed as the element type.</p>
2172 <table class="layout">
2174 <td class="left"><tt><4 x i32></tt></td>
2175 <td class="left">Vector of 4 32-bit integer values.</td>
2178 <td class="left"><tt><8 x float></tt></td>
2179 <td class="left">Vector of 8 32-bit floating-point values.</td>
2182 <td class="left"><tt><2 x i64></tt></td>
2183 <td class="left">Vector of 2 64-bit integer values.</td>
2191 <!-- *********************************************************************** -->
2192 <h2><a name="constants">Constants</a></h2>
2193 <!-- *********************************************************************** -->
2197 <p>LLVM has several different basic types of constants. This section describes
2198 them all and their syntax.</p>
2200 <!-- ======================================================================= -->
2202 <a name="simpleconstants">Simple Constants</a>
2208 <dt><b>Boolean constants</b></dt>
2209 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
2210 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
2212 <dt><b>Integer constants</b></dt>
2213 <dd>Standard integers (such as '4') are constants of
2214 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2215 with integer types.</dd>
2217 <dt><b>Floating point constants</b></dt>
2218 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
2219 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2220 notation (see below). The assembler requires the exact decimal value of a
2221 floating-point constant. For example, the assembler accepts 1.25 but
2222 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2223 constants must have a <a href="#t_floating">floating point</a> type. </dd>
2225 <dt><b>Null pointer constants</b></dt>
2226 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
2227 and must be of <a href="#t_pointer">pointer type</a>.</dd>
2230 <p>The one non-intuitive notation for constants is the hexadecimal form of
2231 floating point constants. For example, the form '<tt>double
2232 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2233 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2234 constants are required (and the only time that they are generated by the
2235 disassembler) is when a floating point constant must be emitted but it cannot
2236 be represented as a decimal floating point number in a reasonable number of
2237 digits. For example, NaN's, infinities, and other special values are
2238 represented in their IEEE hexadecimal format so that assembly and disassembly
2239 do not cause any bits to change in the constants.</p>
2241 <p>When using the hexadecimal form, constants of types float and double are
2242 represented using the 16-digit form shown above (which matches the IEEE754
2243 representation for double); float values must, however, be exactly
2244 representable as IEE754 single precision. Hexadecimal format is always used
2245 for long double, and there are three forms of long double. The 80-bit format
2246 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2247 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2248 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2249 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2250 currently supported target uses this format. Long doubles will only work if
2251 they match the long double format on your target. All hexadecimal formats
2252 are big-endian (sign bit at the left).</p>
2254 <p>There are no constants of type x86mmx.</p>
2257 <!-- ======================================================================= -->
2259 <a name="aggregateconstants"></a> <!-- old anchor -->
2260 <a name="complexconstants">Complex Constants</a>
2265 <p>Complex constants are a (potentially recursive) combination of simple
2266 constants and smaller complex constants.</p>
2269 <dt><b>Structure constants</b></dt>
2270 <dd>Structure constants are represented with notation similar to structure
2271 type definitions (a comma separated list of elements, surrounded by braces
2272 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2273 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2274 Structure constants must have <a href="#t_struct">structure type</a>, and
2275 the number and types of elements must match those specified by the
2278 <dt><b>Array constants</b></dt>
2279 <dd>Array constants are represented with notation similar to array type
2280 definitions (a comma separated list of elements, surrounded by square
2281 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2282 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2283 the number and types of elements must match those specified by the
2286 <dt><b>Vector constants</b></dt>
2287 <dd>Vector constants are represented with notation similar to vector type
2288 definitions (a comma separated list of elements, surrounded by
2289 less-than/greater-than's (<tt><></tt>)). For example: "<tt>< i32
2290 42, i32 11, i32 74, i32 100 ></tt>". Vector constants must
2291 have <a href="#t_vector">vector type</a>, and the number and types of
2292 elements must match those specified by the type.</dd>
2294 <dt><b>Zero initialization</b></dt>
2295 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
2296 value to zero of <em>any</em> type, including scalar and
2297 <a href="#t_aggregate">aggregate</a> types.
2298 This is often used to avoid having to print large zero initializers
2299 (e.g. for large arrays) and is always exactly equivalent to using explicit
2300 zero initializers.</dd>
2302 <dt><b>Metadata node</b></dt>
2303 <dd>A metadata node is a structure-like constant with
2304 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2305 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2306 be interpreted as part of the instruction stream, metadata is a place to
2307 attach additional information such as debug info.</dd>
2312 <!-- ======================================================================= -->
2314 <a name="globalconstants">Global Variable and Function Addresses</a>
2319 <p>The addresses of <a href="#globalvars">global variables</a>
2320 and <a href="#functionstructure">functions</a> are always implicitly valid
2321 (link-time) constants. These constants are explicitly referenced when
2322 the <a href="#identifiers">identifier for the global</a> is used and always
2323 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2324 legal LLVM file:</p>
2326 <pre class="doc_code">
2329 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2334 <!-- ======================================================================= -->
2336 <a name="undefvalues">Undefined Values</a>
2341 <p>The string '<tt>undef</tt>' can be used anywhere a constant is expected, and
2342 indicates that the user of the value may receive an unspecified bit-pattern.
2343 Undefined values may be of any type (other than '<tt>label</tt>'
2344 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
2346 <p>Undefined values are useful because they indicate to the compiler that the
2347 program is well defined no matter what value is used. This gives the
2348 compiler more freedom to optimize. Here are some examples of (potentially
2349 surprising) transformations that are valid (in pseudo IR):</p>
2352 <pre class="doc_code">
2362 <p>This is safe because all of the output bits are affected by the undef bits.
2363 Any output bit can have a zero or one depending on the input bits.</p>
2365 <pre class="doc_code">
2376 <p>These logical operations have bits that are not always affected by the input.
2377 For example, if <tt>%X</tt> has a zero bit, then the output of the
2378 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2379 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2380 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2381 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2382 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2383 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2384 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
2386 <pre class="doc_code">
2387 %A = select undef, %X, %Y
2388 %B = select undef, 42, %Y
2389 %C = select %X, %Y, undef
2400 <p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2401 branch) conditions can go <em>either way</em>, but they have to come from one
2402 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2403 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2404 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2405 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2406 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2409 <pre class="doc_code">
2410 %A = xor undef, undef
2428 <p>This example points out that two '<tt>undef</tt>' operands are not
2429 necessarily the same. This can be surprising to people (and also matches C
2430 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2431 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2432 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2433 its value over its "live range". This is true because the variable doesn't
2434 actually <em>have a live range</em>. Instead, the value is logically read
2435 from arbitrary registers that happen to be around when needed, so the value
2436 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2437 need to have the same semantics or the core LLVM "replace all uses with"
2438 concept would not hold.</p>
2440 <pre class="doc_code">
2448 <p>These examples show the crucial difference between an <em>undefined
2449 value</em> and <em>undefined behavior</em>. An undefined value (like
2450 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2451 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2452 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2453 defined on SNaN's. However, in the second example, we can make a more
2454 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2455 arbitrary value, we are allowed to assume that it could be zero. Since a
2456 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2457 the operation does not execute at all. This allows us to delete the divide and
2458 all code after it. Because the undefined operation "can't happen", the
2459 optimizer can assume that it occurs in dead code.</p>
2461 <pre class="doc_code">
2462 a: store undef -> %X
2463 b: store %X -> undef
2469 <p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2470 undefined value can be assumed to not have any effect; we can assume that the
2471 value is overwritten with bits that happen to match what was already there.
2472 However, a store <em>to</em> an undefined location could clobber arbitrary
2473 memory, therefore, it has undefined behavior.</p>
2477 <!-- ======================================================================= -->
2479 <a name="trapvalues">Trap Values</a>
2484 <p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
2485 instead of representing an unspecified bit pattern, they represent the
2486 fact that an instruction or constant expression which cannot evoke side
2487 effects has nevertheless detected a condition which results in undefined
2490 <p>There is currently no way of representing a trap value in the IR; they
2491 only exist when produced by operations such as
2492 <a href="#i_add"><tt>add</tt></a> with the <tt>nsw</tt> flag.</p>
2494 <p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
2497 <li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2498 their operands.</li>
2500 <li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2501 to their dynamic predecessor basic block.</li>
2503 <li>Function arguments depend on the corresponding actual argument values in
2504 the dynamic callers of their functions.</li>
2506 <li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2507 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2508 control back to them.</li>
2510 <li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2511 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2512 or exception-throwing call instructions that dynamically transfer control
2515 <li>Non-volatile loads and stores depend on the most recent stores to all of the
2516 referenced memory addresses, following the order in the IR
2517 (including loads and stores implied by intrinsics such as
2518 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2520 <!-- TODO: In the case of multiple threads, this only applies if the store
2521 "happens-before" the load or store. -->
2523 <!-- TODO: floating-point exception state -->
2525 <li>An instruction with externally visible side effects depends on the most
2526 recent preceding instruction with externally visible side effects, following
2527 the order in the IR. (This includes
2528 <a href="#volatile">volatile operations</a>.)</li>
2530 <li>An instruction <i>control-depends</i> on a
2531 <a href="#terminators">terminator instruction</a>
2532 if the terminator instruction has multiple successors and the instruction
2533 is always executed when control transfers to one of the successors, and
2534 may not be executed when control is transferred to another.</li>
2536 <li>Additionally, an instruction also <i>control-depends</i> on a terminator
2537 instruction if the set of instructions it otherwise depends on would be
2538 different if the terminator had transferred control to a different
2541 <li>Dependence is transitive.</li>
2545 <p>Whenever a trap value is generated, all values which depend on it evaluate
2546 to trap. If they have side effects, the evoke their side effects as if each
2547 operand with a trap value were undef. If they have externally-visible side
2548 effects, the behavior is undefined.</p>
2550 <p>Here are some examples:</p>
2552 <pre class="doc_code">
2554 %trap = sub nuw i32 0, 1 ; Results in a trap value.
2555 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2556 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2557 store i32 0, i32* %trap_yet_again ; undefined behavior
2559 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2560 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2562 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2564 %narrowaddr = bitcast i32* @g to i16*
2565 %wideaddr = bitcast i32* @g to i64*
2566 %trap3 = load i16* %narrowaddr ; Returns a trap value.
2567 %trap4 = load i64* %wideaddr ; Returns a trap value.
2569 %cmp = icmp slt i32 %trap, 0 ; Returns a trap value.
2570 br i1 %cmp, label %true, label %end ; Branch to either destination.
2573 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2574 ; it has undefined behavior.
2578 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2579 ; Both edges into this PHI are
2580 ; control-dependent on %cmp, so this
2581 ; always results in a trap value.
2583 volatile store i32 0, i32* @g ; This would depend on the store in %true
2584 ; if %cmp is true, or the store in %entry
2585 ; otherwise, so this is undefined behavior.
2587 br i1 %cmp, label %second_true, label %second_end
2588 ; The same branch again, but this time the
2589 ; true block doesn't have side effects.
2596 volatile store i32 0, i32* @g ; This time, the instruction always depends
2597 ; on the store in %end. Also, it is
2598 ; control-equivalent to %end, so this is
2599 ; well-defined (again, ignoring earlier
2600 ; undefined behavior in this example).
2605 <!-- ======================================================================= -->
2607 <a name="blockaddress">Addresses of Basic Blocks</a>
2612 <p><b><tt>blockaddress(@function, %block)</tt></b></p>
2614 <p>The '<tt>blockaddress</tt>' constant computes the address of the specified
2615 basic block in the specified function, and always has an i8* type. Taking
2616 the address of the entry block is illegal.</p>
2618 <p>This value only has defined behavior when used as an operand to the
2619 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2620 comparisons against null. Pointer equality tests between labels addresses
2621 results in undefined behavior — though, again, comparison against null
2622 is ok, and no label is equal to the null pointer. This may be passed around
2623 as an opaque pointer sized value as long as the bits are not inspected. This
2624 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2625 long as the original value is reconstituted before the <tt>indirectbr</tt>
2628 <p>Finally, some targets may provide defined semantics when using the value as
2629 the operand to an inline assembly, but that is target specific.</p>
2634 <!-- ======================================================================= -->
2636 <a name="constantexprs">Constant Expressions</a>
2641 <p>Constant expressions are used to allow expressions involving other constants
2642 to be used as constants. Constant expressions may be of
2643 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2644 operation that does not have side effects (e.g. load and call are not
2645 supported). The following is the syntax for constant expressions:</p>
2648 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
2649 <dd>Truncate a constant to another type. The bit size of CST must be larger
2650 than the bit size of TYPE. Both types must be integers.</dd>
2652 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
2653 <dd>Zero extend a constant to another type. The bit size of CST must be
2654 smaller than the bit size of TYPE. Both types must be integers.</dd>
2656 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
2657 <dd>Sign extend a constant to another type. The bit size of CST must be
2658 smaller than the bit size of TYPE. Both types must be integers.</dd>
2660 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
2661 <dd>Truncate a floating point constant to another floating point type. The
2662 size of CST must be larger than the size of TYPE. Both types must be
2663 floating point.</dd>
2665 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
2666 <dd>Floating point extend a constant to another type. The size of CST must be
2667 smaller or equal to the size of TYPE. Both types must be floating
2670 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
2671 <dd>Convert a floating point constant to the corresponding unsigned integer
2672 constant. TYPE must be a scalar or vector integer type. CST must be of
2673 scalar or vector floating point type. Both CST and TYPE must be scalars,
2674 or vectors of the same number of elements. If the value won't fit in the
2675 integer type, the results are undefined.</dd>
2677 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
2678 <dd>Convert a floating point constant to the corresponding signed integer
2679 constant. TYPE must be a scalar or vector integer type. CST must be of
2680 scalar or vector floating point type. Both CST and TYPE must be scalars,
2681 or vectors of the same number of elements. If the value won't fit in the
2682 integer type, the results are undefined.</dd>
2684 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
2685 <dd>Convert an unsigned integer constant to the corresponding floating point
2686 constant. TYPE must be a scalar or vector floating point type. CST must be
2687 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2688 vectors of the same number of elements. If the value won't fit in the
2689 floating point type, the results are undefined.</dd>
2691 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
2692 <dd>Convert a signed integer constant to the corresponding floating point
2693 constant. TYPE must be a scalar or vector floating point type. CST must be
2694 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2695 vectors of the same number of elements. If the value won't fit in the
2696 floating point type, the results are undefined.</dd>
2698 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
2699 <dd>Convert a pointer typed constant to the corresponding integer constant
2700 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2701 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2702 make it fit in <tt>TYPE</tt>.</dd>
2704 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
2705 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2706 type. CST must be of integer type. The CST value is zero extended,
2707 truncated, or unchanged to make it fit in a pointer size. This one is
2708 <i>really</i> dangerous!</dd>
2710 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
2711 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2712 are the same as those for the <a href="#i_bitcast">bitcast
2713 instruction</a>.</dd>
2715 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2716 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2717 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
2718 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2719 instruction, the index list may have zero or more indexes, which are
2720 required to make sense for the type of "CSTPTR".</dd>
2722 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
2723 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
2725 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
2726 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2728 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
2729 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
2731 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
2732 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2735 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
2736 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2739 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
2740 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2743 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2744 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2745 constants. The index list is interpreted in a similar manner as indices in
2746 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2747 index value must be specified.</dd>
2749 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2750 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2751 constants. The index list is interpreted in a similar manner as indices in
2752 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2753 index value must be specified.</dd>
2755 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
2756 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2757 be any of the <a href="#binaryops">binary</a>
2758 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2759 on operands are the same as those for the corresponding instruction
2760 (e.g. no bitwise operations on floating point values are allowed).</dd>
2767 <!-- *********************************************************************** -->
2768 <h2><a name="othervalues">Other Values</a></h2>
2769 <!-- *********************************************************************** -->
2771 <!-- ======================================================================= -->
2773 <a name="inlineasm">Inline Assembler Expressions</a>
2778 <p>LLVM supports inline assembler expressions (as opposed
2779 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2780 a special value. This value represents the inline assembler as a string
2781 (containing the instructions to emit), a list of operand constraints (stored
2782 as a string), a flag that indicates whether or not the inline asm
2783 expression has side effects, and a flag indicating whether the function
2784 containing the asm needs to align its stack conservatively. An example
2785 inline assembler expression is:</p>
2787 <pre class="doc_code">
2788 i32 (i32) asm "bswap $0", "=r,r"
2791 <p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2792 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2795 <pre class="doc_code">
2796 %X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
2799 <p>Inline asms with side effects not visible in the constraint list must be
2800 marked as having side effects. This is done through the use of the
2801 '<tt>sideeffect</tt>' keyword, like so:</p>
2803 <pre class="doc_code">
2804 call void asm sideeffect "eieio", ""()
2807 <p>In some cases inline asms will contain code that will not work unless the
2808 stack is aligned in some way, such as calls or SSE instructions on x86,
2809 yet will not contain code that does that alignment within the asm.
2810 The compiler should make conservative assumptions about what the asm might
2811 contain and should generate its usual stack alignment code in the prologue
2812 if the '<tt>alignstack</tt>' keyword is present:</p>
2814 <pre class="doc_code">
2815 call void asm alignstack "eieio", ""()
2818 <p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2821 <p>TODO: The format of the asm and constraints string still need to be
2822 documented here. Constraints on what can be done (e.g. duplication, moving,
2823 etc need to be documented). This is probably best done by reference to
2824 another document that covers inline asm from a holistic perspective.</p>
2827 <a name="inlineasm_md">Inline Asm Metadata</a>
2832 <p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
2833 attached to it that contains a list of constant integers. If present, the
2834 code generator will use the integer as the location cookie value when report
2835 errors through the LLVMContext error reporting mechanisms. This allows a
2836 front-end to correlate backend errors that occur with inline asm back to the
2837 source code that produced it. For example:</p>
2839 <pre class="doc_code">
2840 call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2842 !42 = !{ i32 1234567 }
2845 <p>It is up to the front-end to make sense of the magic numbers it places in the
2846 IR. If the MDNode contains multiple constants, the code generator will use
2847 the one that corresponds to the line of the asm that the error occurs on.</p>
2853 <!-- ======================================================================= -->
2855 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2860 <p>LLVM IR allows metadata to be attached to instructions in the program that
2861 can convey extra information about the code to the optimizers and code
2862 generator. One example application of metadata is source-level debug
2863 information. There are two metadata primitives: strings and nodes. All
2864 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2865 preceding exclamation point ('<tt>!</tt>').</p>
2867 <p>A metadata string is a string surrounded by double quotes. It can contain
2868 any character by escaping non-printable characters with "\xx" where "xx" is
2869 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2871 <p>Metadata nodes are represented with notation similar to structure constants
2872 (a comma separated list of elements, surrounded by braces and preceded by an
2873 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2874 10}</tt>". Metadata nodes can have any values as their operand.</p>
2876 <p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2877 metadata nodes, which can be looked up in the module symbol table. For
2878 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2880 <p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
2881 function is using two metadata arguments.</p>
2883 <div class="doc_code">
2885 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2889 <p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
2890 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
2892 <div class="doc_code">
2894 %indvar.next = add i64 %indvar, 1, !dbg !21
2902 <!-- *********************************************************************** -->
2904 <a name="intrinsic_globals">Intrinsic Global Variables</a>
2906 <!-- *********************************************************************** -->
2908 <p>LLVM has a number of "magic" global variables that contain data that affect
2909 code generation or other IR semantics. These are documented here. All globals
2910 of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2911 section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2914 <!-- ======================================================================= -->
2916 <a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
2921 <p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2922 href="#linkage_appending">appending linkage</a>. This array contains a list of
2923 pointers to global variables and functions which may optionally have a pointer
2924 cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2930 @llvm.used = appending global [2 x i8*] [
2932 i8* bitcast (i32* @Y to i8*)
2933 ], section "llvm.metadata"
2936 <p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2937 compiler, assembler, and linker are required to treat the symbol as if there is
2938 a reference to the global that it cannot see. For example, if a variable has
2939 internal linkage and no references other than that from the <tt>@llvm.used</tt>
2940 list, it cannot be deleted. This is commonly used to represent references from
2941 inline asms and other things the compiler cannot "see", and corresponds to
2942 "attribute((used))" in GNU C.</p>
2944 <p>On some targets, the code generator must emit a directive to the assembler or
2945 object file to prevent the assembler and linker from molesting the symbol.</p>
2949 <!-- ======================================================================= -->
2951 <a name="intg_compiler_used">
2952 The '<tt>llvm.compiler.used</tt>' Global Variable
2958 <p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2959 <tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2960 touching the symbol. On targets that support it, this allows an intelligent
2961 linker to optimize references to the symbol without being impeded as it would be
2962 by <tt>@llvm.used</tt>.</p>
2964 <p>This is a rare construct that should only be used in rare circumstances, and
2965 should not be exposed to source languages.</p>
2969 <!-- ======================================================================= -->
2971 <a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
2976 %0 = type { i32, void ()* }
2977 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
2979 <p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor functions and associated priorities. The functions referenced by this array will be called in ascending order of priority (i.e. lowest first) when the module is loaded. The order of functions with the same priority is not defined.
2984 <!-- ======================================================================= -->
2986 <a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
2991 %0 = type { i32, void ()* }
2992 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
2995 <p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions and associated priorities. The functions referenced by this array will be called in descending order of priority (i.e. highest first) when the module is loaded. The order of functions with the same priority is not defined.
3002 <!-- *********************************************************************** -->
3003 <h2><a name="instref">Instruction Reference</a></h2>
3004 <!-- *********************************************************************** -->
3008 <p>The LLVM instruction set consists of several different classifications of
3009 instructions: <a href="#terminators">terminator
3010 instructions</a>, <a href="#binaryops">binary instructions</a>,
3011 <a href="#bitwiseops">bitwise binary instructions</a>,
3012 <a href="#memoryops">memory instructions</a>, and
3013 <a href="#otherops">other instructions</a>.</p>
3015 <!-- ======================================================================= -->
3017 <a name="terminators">Terminator Instructions</a>
3022 <p>As mentioned <a href="#functionstructure">previously</a>, every basic block
3023 in a program ends with a "Terminator" instruction, which indicates which
3024 block should be executed after the current block is finished. These
3025 terminator instructions typically yield a '<tt>void</tt>' value: they produce
3026 control flow, not values (the one exception being the
3027 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
3029 <p>There are eight different terminator instructions: the
3030 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
3031 '<a href="#i_br"><tt>br</tt></a>' instruction, the
3032 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
3033 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
3034 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
3035 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, the
3036 '<a href="#i_resume"><tt>resume</tt></a>' instruction, and the
3037 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
3039 <!-- _______________________________________________________________________ -->
3041 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
3048 ret <type> <value> <i>; Return a value from a non-void function</i>
3049 ret void <i>; Return from void function</i>
3053 <p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
3054 a value) from a function back to the caller.</p>
3056 <p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
3057 value and then causes control flow, and one that just causes control flow to
3061 <p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
3062 return value. The type of the return value must be a
3063 '<a href="#t_firstclass">first class</a>' type.</p>
3065 <p>A function is not <a href="#wellformed">well formed</a> if it it has a
3066 non-void return type and contains a '<tt>ret</tt>' instruction with no return
3067 value or a return value with a type that does not match its type, or if it
3068 has a void return type and contains a '<tt>ret</tt>' instruction with a
3072 <p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
3073 the calling function's context. If the caller is a
3074 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
3075 instruction after the call. If the caller was an
3076 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
3077 the beginning of the "normal" destination block. If the instruction returns
3078 a value, that value shall set the call or invoke instruction's return
3083 ret i32 5 <i>; Return an integer value of 5</i>
3084 ret void <i>; Return from a void function</i>
3085 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
3089 <!-- _______________________________________________________________________ -->
3091 <a name="i_br">'<tt>br</tt>' Instruction</a>
3098 br i1 <cond>, label <iftrue>, label <iffalse>
3099 br label <dest> <i>; Unconditional branch</i>
3103 <p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3104 different basic block in the current function. There are two forms of this
3105 instruction, corresponding to a conditional branch and an unconditional
3109 <p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3110 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3111 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3115 <p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
3116 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3117 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3118 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3123 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3124 br i1 %cond, label %IfEqual, label %IfUnequal
3126 <a href="#i_ret">ret</a> i32 1
3128 <a href="#i_ret">ret</a> i32 0
3133 <!-- _______________________________________________________________________ -->
3135 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
3142 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
3146 <p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
3147 several different places. It is a generalization of the '<tt>br</tt>'
3148 instruction, allowing a branch to occur to one of many possible
3152 <p>The '<tt>switch</tt>' instruction uses three parameters: an integer
3153 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3154 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3155 The table is not allowed to contain duplicate constant entries.</p>
3158 <p>The <tt>switch</tt> instruction specifies a table of values and
3159 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3160 is searched for the given value. If the value is found, control flow is
3161 transferred to the corresponding destination; otherwise, control flow is
3162 transferred to the default destination.</p>
3164 <h5>Implementation:</h5>
3165 <p>Depending on properties of the target machine and the particular
3166 <tt>switch</tt> instruction, this instruction may be code generated in
3167 different ways. For example, it could be generated as a series of chained
3168 conditional branches or with a lookup table.</p>
3172 <i>; Emulate a conditional br instruction</i>
3173 %Val = <a href="#i_zext">zext</a> i1 %value to i32
3174 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
3176 <i>; Emulate an unconditional br instruction</i>
3177 switch i32 0, label %dest [ ]
3179 <i>; Implement a jump table:</i>
3180 switch i32 %val, label %otherwise [ i32 0, label %onzero
3182 i32 2, label %ontwo ]
3188 <!-- _______________________________________________________________________ -->
3190 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
3197 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
3202 <p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
3203 within the current function, whose address is specified by
3204 "<tt>address</tt>". Address must be derived from a <a
3205 href="#blockaddress">blockaddress</a> constant.</p>
3209 <p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3210 rest of the arguments indicate the full set of possible destinations that the
3211 address may point to. Blocks are allowed to occur multiple times in the
3212 destination list, though this isn't particularly useful.</p>
3214 <p>This destination list is required so that dataflow analysis has an accurate
3215 understanding of the CFG.</p>
3219 <p>Control transfers to the block specified in the address argument. All
3220 possible destination blocks must be listed in the label list, otherwise this
3221 instruction has undefined behavior. This implies that jumps to labels
3222 defined in other functions have undefined behavior as well.</p>
3224 <h5>Implementation:</h5>
3226 <p>This is typically implemented with a jump through a register.</p>
3230 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
3236 <!-- _______________________________________________________________________ -->
3238 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
3245 <result> = invoke [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] <ptr to function ty> <function ptr val>(<function args>) [<a href="#fnattrs">fn attrs</a>]
3246 to label <normal label> unwind label <exception label>
3250 <p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
3251 function, with the possibility of control flow transfer to either the
3252 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3253 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3254 control flow will return to the "normal" label. If the callee (or any
3255 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3256 instruction, control is interrupted and continued at the dynamically nearest
3257 "exception" label.</p>
3260 <p>This instruction requires several arguments:</p>
3263 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3264 convention</a> the call should use. If none is specified, the call
3265 defaults to using C calling conventions.</li>
3267 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
3268 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3269 '<tt>inreg</tt>' attributes are valid here.</li>
3271 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
3272 function value being invoked. In most cases, this is a direct function
3273 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3274 off an arbitrary pointer to function value.</li>
3276 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
3277 function to be invoked. </li>
3279 <li>'<tt>function args</tt>': argument list whose types match the function
3280 signature argument types and parameter attributes. All arguments must be
3281 of <a href="#t_firstclass">first class</a> type. If the function
3282 signature indicates the function accepts a variable number of arguments,
3283 the extra arguments can be specified.</li>
3285 <li>'<tt>normal label</tt>': the label reached when the called function
3286 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
3288 <li>'<tt>exception label</tt>': the label reached when a callee returns with
3289 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
3291 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
3292 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3293 '<tt>readnone</tt>' attributes are valid here.</li>
3297 <p>This instruction is designed to operate as a standard
3298 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3299 primary difference is that it establishes an association with a label, which
3300 is used by the runtime library to unwind the stack.</p>
3302 <p>This instruction is used in languages with destructors to ensure that proper
3303 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3304 exception. Additionally, this is important for implementation of
3305 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
3307 <p>For the purposes of the SSA form, the definition of the value returned by the
3308 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3309 block to the "normal" label. If the callee unwinds then no return value is
3312 <p>Note that the code generator does not yet completely support unwind, and
3313 that the invoke/unwind semantics are likely to change in future versions.</p>
3317 %retval = invoke i32 @Test(i32 15) to label %Continue
3318 unwind label %TestCleanup <i>; {i32}:retval set</i>
3319 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
3320 unwind label %TestCleanup <i>; {i32}:retval set</i>
3325 <!-- _______________________________________________________________________ -->
3328 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3339 <p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
3340 at the first callee in the dynamic call stack which used
3341 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3342 This is primarily used to implement exception handling.</p>
3345 <p>The '<tt>unwind</tt>' instruction causes execution of the current function to
3346 immediately halt. The dynamic call stack is then searched for the
3347 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3348 Once found, execution continues at the "exceptional" destination block
3349 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3350 instruction in the dynamic call chain, undefined behavior results.</p>
3352 <p>Note that the code generator does not yet completely support unwind, and
3353 that the invoke/unwind semantics are likely to change in future versions.</p>
3357 <!-- _______________________________________________________________________ -->
3360 <a name="i_resume">'<tt>resume</tt>' Instruction</a>
3367 resume <type> <value>
3371 <p>The '<tt>resume</tt>' instruction is a terminator instruction that has no
3375 <p>The '<tt>resume</tt>' instruction's argument must have the same type as the
3376 result of any '<tt>landingpad</tt>' instruction in the same function.</p>
3379 <p>The '<tt>resume</tt>' instruction resumes propagation of an existing
3380 (in-flight) exception whose unwinding was interrupted with
3381 a landingpad instruction.</p>
3385 resume { i8*, i32 } %exn
3390 <!-- _______________________________________________________________________ -->
3393 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3404 <p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
3405 instruction is used to inform the optimizer that a particular portion of the
3406 code is not reachable. This can be used to indicate that the code after a
3407 no-return function cannot be reached, and other facts.</p>
3410 <p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
3416 <!-- ======================================================================= -->
3418 <a name="binaryops">Binary Operations</a>
3423 <p>Binary operators are used to do most of the computation in a program. They
3424 require two operands of the same type, execute an operation on them, and
3425 produce a single value. The operands might represent multiple data, as is
3426 the case with the <a href="#t_vector">vector</a> data type. The result value
3427 has the same type as its operands.</p>
3429 <p>There are several different binary operators:</p>
3431 <!-- _______________________________________________________________________ -->
3433 <a name="i_add">'<tt>add</tt>' Instruction</a>
3440 <result> = add <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3441 <result> = add nuw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3442 <result> = add nsw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3443 <result> = add nuw nsw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3447 <p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
3450 <p>The two arguments to the '<tt>add</tt>' instruction must
3451 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3452 integer values. Both arguments must have identical types.</p>
3455 <p>The value produced is the integer sum of the two operands.</p>
3457 <p>If the sum has unsigned overflow, the result returned is the mathematical
3458 result modulo 2<sup>n</sup>, where n is the bit width of the result.</p>
3460 <p>Because LLVM integers use a two's complement representation, this instruction
3461 is appropriate for both signed and unsigned integers.</p>
3463 <p><tt>nuw</tt> and <tt>nsw</tt> stand for "No Unsigned Wrap"
3464 and "No Signed Wrap", respectively. If the <tt>nuw</tt> and/or
3465 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
3466 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3467 respectively, occurs.</p>
3471 <result> = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
3476 <!-- _______________________________________________________________________ -->
3478 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
3485 <result> = fadd <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3489 <p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3492 <p>The two arguments to the '<tt>fadd</tt>' instruction must be
3493 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3494 floating point values. Both arguments must have identical types.</p>
3497 <p>The value produced is the floating point sum of the two operands.</p>
3501 <result> = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3506 <!-- _______________________________________________________________________ -->
3508 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
3515 <result> = sub <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3516 <result> = sub nuw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3517 <result> = sub nsw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3518 <result> = sub nuw nsw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3522 <p>The '<tt>sub</tt>' instruction returns the difference of its two
3525 <p>Note that the '<tt>sub</tt>' instruction is used to represent the
3526 '<tt>neg</tt>' instruction present in most other intermediate
3527 representations.</p>
3530 <p>The two arguments to the '<tt>sub</tt>' instruction must
3531 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3532 integer values. Both arguments must have identical types.</p>
3535 <p>The value produced is the integer difference of the two operands.</p>
3537 <p>If the difference has unsigned overflow, the result returned is the
3538 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3541 <p>Because LLVM integers use a two's complement representation, this instruction
3542 is appropriate for both signed and unsigned integers.</p>
3544 <p><tt>nuw</tt> and <tt>nsw</tt> stand for "No Unsigned Wrap"
3545 and "No Signed Wrap", respectively. If the <tt>nuw</tt> and/or
3546 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
3547 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3548 respectively, occurs.</p>
3552 <result> = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
3553 <result> = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
3558 <!-- _______________________________________________________________________ -->
3560 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
3567 <result> = fsub <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3571 <p>The '<tt>fsub</tt>' instruction returns the difference of its two
3574 <p>Note that the '<tt>fsub</tt>' instruction is used to represent the
3575 '<tt>fneg</tt>' instruction present in most other intermediate
3576 representations.</p>
3579 <p>The two arguments to the '<tt>fsub</tt>' instruction must be
3580 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3581 floating point values. Both arguments must have identical types.</p>
3584 <p>The value produced is the floating point difference of the two operands.</p>
3588 <result> = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3589 <result> = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3594 <!-- _______________________________________________________________________ -->
3596 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
3603 <result> = mul <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3604 <result> = mul nuw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3605 <result> = mul nsw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3606 <result> = mul nuw nsw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3610 <p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
3613 <p>The two arguments to the '<tt>mul</tt>' instruction must
3614 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3615 integer values. Both arguments must have identical types.</p>
3618 <p>The value produced is the integer product of the two operands.</p>
3620 <p>If the result of the multiplication has unsigned overflow, the result
3621 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3622 width of the result.</p>
3624 <p>Because LLVM integers use a two's complement representation, and the result
3625 is the same width as the operands, this instruction returns the correct
3626 result for both signed and unsigned integers. If a full product
3627 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3628 be sign-extended or zero-extended as appropriate to the width of the full
3631 <p><tt>nuw</tt> and <tt>nsw</tt> stand for "No Unsigned Wrap"
3632 and "No Signed Wrap", respectively. If the <tt>nuw</tt> and/or
3633 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
3634 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3635 respectively, occurs.</p>
3639 <result> = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
3644 <!-- _______________________________________________________________________ -->
3646 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
3653 <result> = fmul <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3657 <p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
3660 <p>The two arguments to the '<tt>fmul</tt>' instruction must be
3661 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3662 floating point values. Both arguments must have identical types.</p>
3665 <p>The value produced is the floating point product of the two operands.</p>
3669 <result> = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
3674 <!-- _______________________________________________________________________ -->
3676 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3683 <result> = udiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3684 <result> = udiv exact <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3688 <p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
3691 <p>The two arguments to the '<tt>udiv</tt>' instruction must be
3692 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3693 values. Both arguments must have identical types.</p>
3696 <p>The value produced is the unsigned integer quotient of the two operands.</p>
3698 <p>Note that unsigned integer division and signed integer division are distinct
3699 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3701 <p>Division by zero leads to undefined behavior.</p>
3703 <p>If the <tt>exact</tt> keyword is present, the result value of the
3704 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3705 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3710 <result> = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
3715 <!-- _______________________________________________________________________ -->
3717 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3724 <result> = sdiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3725 <result> = sdiv exact <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3729 <p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
3732 <p>The two arguments to the '<tt>sdiv</tt>' instruction must be
3733 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3734 values. Both arguments must have identical types.</p>
3737 <p>The value produced is the signed integer quotient of the two operands rounded
3740 <p>Note that signed integer division and unsigned integer division are distinct
3741 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3743 <p>Division by zero leads to undefined behavior. Overflow also leads to
3744 undefined behavior; this is a rare case, but can occur, for example, by doing
3745 a 32-bit division of -2147483648 by -1.</p>
3747 <p>If the <tt>exact</tt> keyword is present, the result value of the
3748 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
3753 <result> = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
3758 <!-- _______________________________________________________________________ -->
3760 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3767 <result> = fdiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3771 <p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
3774 <p>The two arguments to the '<tt>fdiv</tt>' instruction must be
3775 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3776 floating point values. Both arguments must have identical types.</p>
3779 <p>The value produced is the floating point quotient of the two operands.</p>
3783 <result> = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
3788 <!-- _______________________________________________________________________ -->
3790 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3797 <result> = urem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3801 <p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3802 division of its two arguments.</p>
3805 <p>The two arguments to the '<tt>urem</tt>' instruction must be
3806 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3807 values. Both arguments must have identical types.</p>
3810 <p>This instruction returns the unsigned integer <i>remainder</i> of a division.
3811 This instruction always performs an unsigned division to get the
3814 <p>Note that unsigned integer remainder and signed integer remainder are
3815 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3817 <p>Taking the remainder of a division by zero leads to undefined behavior.</p>
3821 <result> = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
3826 <!-- _______________________________________________________________________ -->
3828 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
3835 <result> = srem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3839 <p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3840 division of its two operands. This instruction can also take
3841 <a href="#t_vector">vector</a> versions of the values in which case the
3842 elements must be integers.</p>
3845 <p>The two arguments to the '<tt>srem</tt>' instruction must be
3846 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3847 values. Both arguments must have identical types.</p>
3850 <p>This instruction returns the <i>remainder</i> of a division (where the result
3851 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3852 <i>modulo</i> operator (where the result is either zero or has the same sign
3853 as the divisor, <tt>op2</tt>) of a value.
3854 For more information about the difference,
3855 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3856 Math Forum</a>. For a table of how this is implemented in various languages,
3857 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3858 Wikipedia: modulo operation</a>.</p>
3860 <p>Note that signed integer remainder and unsigned integer remainder are
3861 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3863 <p>Taking the remainder of a division by zero leads to undefined behavior.
3864 Overflow also leads to undefined behavior; this is a rare case, but can
3865 occur, for example, by taking the remainder of a 32-bit division of
3866 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3867 lets srem be implemented using instructions that return both the result of
3868 the division and the remainder.)</p>
3872 <result> = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
3877 <!-- _______________________________________________________________________ -->
3879 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
3886 <result> = frem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3890 <p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3891 its two operands.</p>
3894 <p>The two arguments to the '<tt>frem</tt>' instruction must be
3895 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3896 floating point values. Both arguments must have identical types.</p>
3899 <p>This instruction returns the <i>remainder</i> of a division. The remainder
3900 has the same sign as the dividend.</p>
3904 <result> = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
3911 <!-- ======================================================================= -->
3913 <a name="bitwiseops">Bitwise Binary Operations</a>
3918 <p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3919 program. They are generally very efficient instructions and can commonly be
3920 strength reduced from other instructions. They require two operands of the
3921 same type, execute an operation on them, and produce a single value. The
3922 resulting value is the same type as its operands.</p>
3924 <!-- _______________________________________________________________________ -->
3926 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
3933 <result> = shl <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3934 <result> = shl nuw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3935 <result> = shl nsw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3936 <result> = shl nuw nsw <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3940 <p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3941 a specified number of bits.</p>
3944 <p>Both arguments to the '<tt>shl</tt>' instruction must be the
3945 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3946 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
3949 <p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3950 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3951 is (statically or dynamically) negative or equal to or larger than the number
3952 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3953 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3954 shift amount in <tt>op2</tt>.</p>
3956 <p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3957 <a href="#trapvalues">trap value</a> if it shifts out any non-zero bits. If
3958 the <tt>nsw</tt> keyword is present, then the shift produces a
3959 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3960 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3961 they would if the shift were expressed as a mul instruction with the same
3962 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3966 <result> = shl i32 4, %var <i>; yields {i32}: 4 << %var</i>
3967 <result> = shl i32 4, 2 <i>; yields {i32}: 16</i>
3968 <result> = shl i32 1, 10 <i>; yields {i32}: 1024</i>
3969 <result> = shl i32 1, 32 <i>; undefined</i>
3970 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> <i>; yields: result=<2 x i32> < i32 2, i32 4></i>
3975 <!-- _______________________________________________________________________ -->
3977 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
3984 <result> = lshr <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3985 <result> = lshr exact <ty> <op1>, <op2> <i>; yields {ty}:result</i>
3989 <p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3990 operand shifted to the right a specified number of bits with zero fill.</p>
3993 <p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
3994 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3995 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
3998 <p>This instruction always performs a logical shift right operation. The most
3999 significant bits of the result will be filled with zero bits after the shift.
4000 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
4001 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4002 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4003 shift amount in <tt>op2</tt>.</p>
4005 <p>If the <tt>exact</tt> keyword is present, the result value of the
4006 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
4007 shifted out are non-zero.</p>
4012 <result> = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
4013 <result> = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
4014 <result> = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
4015 <result> = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
4016 <result> = lshr i32 1, 32 <i>; undefined</i>
4017 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> <i>; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1></i>
4022 <!-- _______________________________________________________________________ -->
4024 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
4031 <result> = ashr <ty> <op1>, <op2> <i>; yields {ty}:result</i>
4032 <result> = ashr exact <ty> <op1>, <op2> <i>; yields {ty}:result</i>
4036 <p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
4037 operand shifted to the right a specified number of bits with sign
4041 <p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
4042 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4043 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
4046 <p>This instruction always performs an arithmetic shift right operation, The
4047 most significant bits of the result will be filled with the sign bit
4048 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
4049 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
4050 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
4051 the corresponding shift amount in <tt>op2</tt>.</p>
4053 <p>If the <tt>exact</tt> keyword is present, the result value of the
4054 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
4055 shifted out are non-zero.</p>
4059 <result> = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
4060 <result> = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
4061 <result> = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
4062 <result> = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
4063 <result> = ashr i32 1, 32 <i>; undefined</i>
4064 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> <i>; yields: result=<2 x i32> < i32 -1, i32 0></i>
4069 <!-- _______________________________________________________________________ -->
4071 <a name="i_and">'<tt>and</tt>' Instruction</a>
4078 <result> = and <ty> <op1>, <op2> <i>; yields {ty}:result</i>
4082 <p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
4086 <p>The two arguments to the '<tt>and</tt>' instruction must be
4087 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4088 values. Both arguments must have identical types.</p>
4091 <p>The truth table used for the '<tt>and</tt>' instruction is:</p>
4093 <table border="1" cellspacing="0" cellpadding="4">
4125 <result> = and i32 4, %var <i>; yields {i32}:result = 4 & %var</i>
4126 <result> = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4127 <result> = and i32 4, 8 <i>; yields {i32}:result = 0</i>
4130 <!-- _______________________________________________________________________ -->
4132 <a name="i_or">'<tt>or</tt>' Instruction</a>
4139 <result> = or <ty> <op1>, <op2> <i>; yields {ty}:result</i>
4143 <p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4147 <p>The two arguments to the '<tt>or</tt>' instruction must be
4148 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4149 values. Both arguments must have identical types.</p>
4152 <p>The truth table used for the '<tt>or</tt>' instruction is:</p>
4154 <table border="1" cellspacing="0" cellpadding="4">
4186 <result> = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
4187 <result> = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4188 <result> = or i32 4, 8 <i>; yields {i32}:result = 12</i>
4193 <!-- _______________________________________________________________________ -->
4195 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4202 <result> = xor <ty> <op1>, <op2> <i>; yields {ty}:result</i>
4206 <p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4207 its two operands. The <tt>xor</tt> is used to implement the "one's
4208 complement" operation, which is the "~" operator in C.</p>
4211 <p>The two arguments to the '<tt>xor</tt>' instruction must be
4212 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4213 values. Both arguments must have identical types.</p>
4216 <p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
4218 <table border="1" cellspacing="0" cellpadding="4">
4250 <result> = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
4251 <result> = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4252 <result> = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4253 <result> = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
4260 <!-- ======================================================================= -->
4262 <a name="vectorops">Vector Operations</a>
4267 <p>LLVM supports several instructions to represent vector operations in a
4268 target-independent manner. These instructions cover the element-access and
4269 vector-specific operations needed to process vectors effectively. While LLVM
4270 does directly support these vector operations, many sophisticated algorithms
4271 will want to use target-specific intrinsics to take full advantage of a
4272 specific target.</p>
4274 <!-- _______________________________________________________________________ -->
4276 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
4283 <result> = extractelement <n x <ty>> <val>, i32 <idx> <i>; yields <ty></i>
4287 <p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4288 from a vector at a specified index.</p>
4292 <p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4293 of <a href="#t_vector">vector</a> type. The second operand is an index
4294 indicating the position from which to extract the element. The index may be
4298 <p>The result is a scalar of the same type as the element type of
4299 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4300 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4301 results are undefined.</p>
4305 <result> = extractelement <4 x i32> %vec, i32 0 <i>; yields i32</i>
4310 <!-- _______________________________________________________________________ -->
4312 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
4319 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, i32 <idx> <i>; yields <n x <ty>></i>
4323 <p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4324 vector at a specified index.</p>
4327 <p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4328 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4329 whose type must equal the element type of the first operand. The third
4330 operand is an index indicating the position at which to insert the value.
4331 The index may be a variable.</p>
4334 <p>The result is a vector of the same type as <tt>val</tt>. Its element values
4335 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4336 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4337 results are undefined.</p>
4341 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 <i>; yields <4 x i32></i>
4346 <!-- _______________________________________________________________________ -->
4348 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
4355 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> <i>; yields <m x <ty>></i>
4359 <p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4360 from two input vectors, returning a vector with the same element type as the
4361 input and length that is the same as the shuffle mask.</p>
4364 <p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4365 with types that match each other. The third argument is a shuffle mask whose
4366 element type is always 'i32'. The result of the instruction is a vector
4367 whose length is the same as the shuffle mask and whose element type is the
4368 same as the element type of the first two operands.</p>
4370 <p>The shuffle mask operand is required to be a constant vector with either
4371 constant integer or undef values.</p>
4374 <p>The elements of the two input vectors are numbered from left to right across
4375 both of the vectors. The shuffle mask operand specifies, for each element of
4376 the result vector, which element of the two input vectors the result element
4377 gets. The element selector may be undef (meaning "don't care") and the
4378 second operand may be undef if performing a shuffle from only one vector.</p>
4382 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4383 <4 x i32> <i32 0, i32 4, i32 1, i32 5> <i>; yields <4 x i32></i>
4384 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
4385 <4 x i32> <i32 0, i32 1, i32 2, i32 3> <i>; yields <4 x i32></i> - Identity shuffle.
4386 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
4387 <4 x i32> <i32 0, i32 1, i32 2, i32 3> <i>; yields <4 x i32></i>
4388 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4389 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > <i>; yields <8 x i32></i>
4396 <!-- ======================================================================= -->
4398 <a name="aggregateops">Aggregate Operations</a>
4403 <p>LLVM supports several instructions for working with
4404 <a href="#t_aggregate">aggregate</a> values.</p>
4406 <!-- _______________________________________________________________________ -->
4408 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
4415 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
4419 <p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4420 from an <a href="#t_aggregate">aggregate</a> value.</p>
4423 <p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
4424 of <a href="#t_struct">struct</a> or
4425 <a href="#t_array">array</a> type. The operands are constant indices to
4426 specify which value to extract in a similar manner as indices in a
4427 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
4428 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4430 <li>Since the value being indexed is not a pointer, the first index is
4431 omitted and assumed to be zero.</li>
4432 <li>At least one index must be specified.</li>
4433 <li>Not only struct indices but also array indices must be in
4438 <p>The result is the value at the position in the aggregate specified by the
4443 <result> = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
4448 <!-- _______________________________________________________________________ -->
4450 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
4457 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* <i>; yields <aggregate type></i>
4461 <p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4462 in an <a href="#t_aggregate">aggregate</a> value.</p>
4465 <p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
4466 of <a href="#t_struct">struct</a> or
4467 <a href="#t_array">array</a> type. The second operand is a first-class
4468 value to insert. The following operands are constant indices indicating
4469 the position at which to insert the value in a similar manner as indices in a
4470 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
4471 value to insert must have the same type as the value identified by the
4475 <p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4476 that of <tt>val</tt> except that the value at the position specified by the
4477 indices is that of <tt>elt</tt>.</p>
4481 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4482 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4483 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 <i>; yields {i32 1, float %val}</i>
4490 <!-- ======================================================================= -->
4492 <a name="memoryops">Memory Access and Addressing Operations</a>
4497 <p>A key design point of an SSA-based representation is how it represents
4498 memory. In LLVM, no memory locations are in SSA form, which makes things
4499 very simple. This section describes how to read, write, and allocate
4502 <!-- _______________________________________________________________________ -->
4504 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
4511 <result> = alloca <type>[, <ty> <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
4515 <p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
4516 currently executing function, to be automatically released when this function
4517 returns to its caller. The object is always allocated in the generic address
4518 space (address space zero).</p>
4521 <p>The '<tt>alloca</tt>' instruction
4522 allocates <tt>sizeof(<type>)*NumElements</tt> bytes of memory on the
4523 runtime stack, returning a pointer of the appropriate type to the program.
4524 If "NumElements" is specified, it is the number of elements allocated,
4525 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4526 specified, the value result of the allocation is guaranteed to be aligned to
4527 at least that boundary. If not specified, or if zero, the target can choose
4528 to align the allocation on any convenient boundary compatible with the
4531 <p>'<tt>type</tt>' may be any sized type.</p>
4534 <p>Memory is allocated; a pointer is returned. The operation is undefined if
4535 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4536 memory is automatically released when the function returns. The
4537 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4538 variables that must have an address available. When the function returns
4539 (either with the <tt><a href="#i_ret">ret</a></tt>
4540 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4541 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
4545 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4546 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4547 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4548 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
4553 <!-- _______________________________________________________________________ -->
4555 <a name="i_load">'<tt>load</tt>' Instruction</a>
4562 <result> = load <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>]
4563 <result> = volatile load <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>]
4564 !<index> = !{ i32 1 }
4568 <p>The '<tt>load</tt>' instruction is used to read from memory.</p>
4571 <p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4572 from which to load. The pointer must point to
4573 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4574 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
4575 number or order of execution of this <tt>load</tt> with other <a
4576 href="#volatile">volatile operations</a>.</p>
4578 <p>The optional constant <tt>align</tt> argument specifies the alignment of the
4579 operation (that is, the alignment of the memory address). A value of 0 or an
4580 omitted <tt>align</tt> argument means that the operation has the preferential
4581 alignment for the target. It is the responsibility of the code emitter to
4582 ensure that the alignment information is correct. Overestimating the
4583 alignment results in undefined behavior. Underestimating the alignment may
4584 produce less efficient code. An alignment of 1 is always safe.</p>
4586 <p>The optional <tt>!nontemporal</tt> metadata must reference a single
4587 metatadata name <index> corresponding to a metadata node with
4588 one <tt>i32</tt> entry of value 1. The existence of
4589 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4590 and code generator that this load is not expected to be reused in the cache.
4591 The code generator may select special instructions to save cache bandwidth,
4592 such as the <tt>MOVNT</tt> instruction on x86.</p>
4595 <p>The location of memory pointed to is loaded. If the value being loaded is of
4596 scalar type then the number of bytes read does not exceed the minimum number
4597 of bytes needed to hold all bits of the type. For example, loading an
4598 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4599 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4600 is undefined if the value was not originally written using a store of the
4605 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4606 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
4607 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
4612 <!-- _______________________________________________________________________ -->
4614 <a name="i_store">'<tt>store</tt>' Instruction</a>
4621 store <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] <i>; yields {void}</i>
4622 volatile store <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] <i>; yields {void}</i>
4626 <p>The '<tt>store</tt>' instruction is used to write to memory.</p>
4629 <p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4630 and an address at which to store it. The type of the
4631 '<tt><pointer></tt>' operand must be a pointer to
4632 the <a href="#t_firstclass">first class</a> type of the
4633 '<tt><value></tt>' operand. If the <tt>store</tt> is marked as
4634 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4635 order of execution of this <tt>store</tt> with other <a
4636 href="#volatile">volatile operations</a>.</p>
4638 <p>The optional constant "align" argument specifies the alignment of the
4639 operation (that is, the alignment of the memory address). A value of 0 or an
4640 omitted "align" argument means that the operation has the preferential
4641 alignment for the target. It is the responsibility of the code emitter to
4642 ensure that the alignment information is correct. Overestimating the
4643 alignment results in an undefined behavior. Underestimating the alignment may
4644 produce less efficient code. An alignment of 1 is always safe.</p>
4646 <p>The optional !nontemporal metadata must reference a single metatadata
4647 name <index> corresponding to a metadata node with one i32 entry of
4648 value 1. The existence of the !nontemporal metatadata on the
4649 instruction tells the optimizer and code generator that this load is
4650 not expected to be reused in the cache. The code generator may
4651 select special instructions to save cache bandwidth, such as the
4652 MOVNT instruction on x86.</p>
4656 <p>The contents of memory are updated to contain '<tt><value></tt>' at the
4657 location specified by the '<tt><pointer></tt>' operand. If
4658 '<tt><value></tt>' is of scalar type then the number of bytes written
4659 does not exceed the minimum number of bytes needed to hold all bits of the
4660 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4661 writing a value of a type like <tt>i20</tt> with a size that is not an
4662 integral number of bytes, it is unspecified what happens to the extra bits
4663 that do not belong to the type, but they will typically be overwritten.</p>
4667 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4668 store i32 3, i32* %ptr <i>; yields {void}</i>
4669 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
4674 <!-- _______________________________________________________________________ -->
4675 <div class="doc_subsubsection"> <a name="i_fence">'<tt>fence</tt>'
4676 Instruction</a> </div>
4678 <div class="doc_text">
4682 fence [singlethread] <ordering> <i>; yields {void}</i>
4686 <p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
4687 between operations.</p>
4689 <h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
4690 href="#ordering">ordering</a> argument which defines what
4691 <i>synchronizes-with</i> edges they add. They can only be given
4692 <code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
4693 <code>seq_cst</code> orderings.</p>
4696 <p>A fence <var>A</var> which has (at least) <code>release</code> ordering
4697 semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
4698 <code>acquire</code> ordering semantics if and only if there exist atomic
4699 operations <var>X</var> and <var>Y</var>, both operating on some atomic object
4700 <var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
4701 <var>X</var> modifies <var>M</var> (either directly or through some side effect
4702 of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
4703 <var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
4704 <i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
4705 than an explicit <code>fence</code>, one (but not both) of the atomic operations
4706 <var>X</var> or <var>Y</var> might provide a <code>release</code> or
4707 <code>acquire</code> (resp.) ordering constraint and still
4708 <i>synchronize-with</i> the explicit <code>fence</code> and establish the
4709 <i>happens-before</i> edge.</p>
4711 <p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
4712 having both <code>acquire</code> and <code>release</code> semantics specified
4713 above, participates in the global program order of other <code>seq_cst</code>
4714 operations and/or fences.</p>
4716 <p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
4717 specifies that the fence only synchronizes with other fences in the same
4718 thread. (This is useful for interacting with signal handlers.)</p>
4720 <p>FIXME: This instruction is a work in progress; until it is finished, use
4721 llvm.memory.barrier.
4725 fence acquire <i>; yields {void}</i>
4726 fence singlethread seq_cst <i>; yields {void}</i>
4731 <!-- _______________________________________________________________________ -->
4732 <div class="doc_subsubsection"> <a name="i_cmpxchg">'<tt>cmpxchg</tt>'
4733 Instruction</a> </div>
4735 <div class="doc_text">
4739 [volatile] cmpxchg <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <ordering> <i>; yields {ty}</i>
4743 <p>The '<tt>cmpxchg</tt>' instruction is used to atomically modify memory.
4744 It loads a value in memory and compares it to a given value. If they are
4745 equal, it stores a new value into the memory.</p>
4748 <p>There are three arguments to the '<code>cmpxchg</code>' instruction: an
4749 address to operate on, a value to compare to the value currently be at that
4750 address, and a new value to place at that address if the compared values are
4751 equal. The type of '<var><cmp></var>' must be an integer type whose
4752 bit width is a power of two greater than or equal to eight and less than
4753 or equal to a target-specific size limit. '<var><cmp></var>' and
4754 '<var><new></var>' must have the same type, and the type of
4755 '<var><pointer></var>' must be a pointer to that type. If the
4756 <code>cmpxchg</code> is marked as <code>volatile</code>, then the
4757 optimizer is not allowed to modify the number or order of execution
4758 of this <code>cmpxchg</code> with other <a href="#volatile">volatile
4761 <!-- FIXME: Extend allowed types. -->
4763 <p>The <a href="#ordering"><var>ordering</var></a> argument specifies how this
4764 <code>cmpxchg</code> synchronizes with other atomic operations.</p>
4766 <p>The optional "<code>singlethread</code>" argument declares that the
4767 <code>cmpxchg</code> is only atomic with respect to code (usually signal
4768 handlers) running in the same thread as the <code>cmpxchg</code>. Otherwise the
4769 cmpxchg is atomic with respect to all other code in the system.</p>
4771 <p>The pointer passed into cmpxchg must have alignment greater than or equal to
4772 the size in memory of the operand.
4775 <p>The contents of memory at the location specified by the
4776 '<tt><pointer></tt>' operand is read and compared to
4777 '<tt><cmp></tt>'; if the read value is the equal,
4778 '<tt><new></tt>' is written. The original value at the location
4781 <p>A successful <code>cmpxchg</code> is a read-modify-write instruction for the
4782 purpose of identifying <a href="#release_sequence">release sequences</a>. A
4783 failed <code>cmpxchg</code> is equivalent to an atomic load with an ordering
4784 parameter determined by dropping any <code>release</code> part of the
4785 <code>cmpxchg</code>'s ordering.</p>
4788 FIXME: Is compare_exchange_weak() necessary? (Consider after we've done
4789 optimization work on ARM.)
4791 FIXME: Is a weaker ordering constraint on failure helpful in practice?
4797 %orig = atomic <a href="#i_load">load</a> i32* %ptr unordered <i>; yields {i32}</i>
4798 <a href="#i_br">br</a> label %loop
4801 %cmp = <a href="#i_phi">phi</a> i32 [ %orig, %entry ], [%old, %loop]
4802 %squared = <a href="#i_mul">mul</a> i32 %cmp, %cmp
4803 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared <i>; yields {i32}</i>
4804 %success = <a href="#i_icmp">icmp</a> eq i32 %cmp, %old
4805 <a href="#i_br">br</a> i1 %success, label %done, label %loop
4813 <!-- _______________________________________________________________________ -->
4814 <div class="doc_subsubsection"> <a name="i_atomicrmw">'<tt>atomicrmw</tt>'
4815 Instruction</a> </div>
4817 <div class="doc_text">
4821 [volatile] atomicrmw <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> <i>; yields {ty}</i>
4825 <p>The '<tt>atomicrmw</tt>' instruction is used to atomically modify memory.</p>
4828 <p>There are three arguments to the '<code>atomicrmw</code>' instruction: an
4829 operation to apply, an address whose value to modify, an argument to the
4830 operation. The operation must be one of the following keywords:</p>
4845 <p>The type of '<var><value></var>' must be an integer type whose
4846 bit width is a power of two greater than or equal to eight and less than
4847 or equal to a target-specific size limit. The type of the
4848 '<code><pointer></code>' operand must be a pointer to that type.
4849 If the <code>atomicrmw</code> is marked as <code>volatile</code>, then the
4850 optimizer is not allowed to modify the number or order of execution of this
4851 <code>atomicrmw</code> with other <a href="#volatile">volatile
4854 <!-- FIXME: Extend allowed types. -->
4857 <p>The contents of memory at the location specified by the
4858 '<tt><pointer></tt>' operand are atomically read, modified, and written
4859 back. The original value at the location is returned. The modification is
4860 specified by the <var>operation</var> argument:</p>
4863 <li>xchg: <code>*ptr = val</code></li>
4864 <li>add: <code>*ptr = *ptr + val</code></li>
4865 <li>sub: <code>*ptr = *ptr - val</code></li>
4866 <li>and: <code>*ptr = *ptr & val</code></li>
4867 <li>nand: <code>*ptr = ~(*ptr & val)</code></li>
4868 <li>or: <code>*ptr = *ptr | val</code></li>
4869 <li>xor: <code>*ptr = *ptr ^ val</code></li>
4870 <li>max: <code>*ptr = *ptr > val ? *ptr : val</code> (using a signed comparison)</li>
4871 <li>min: <code>*ptr = *ptr < val ? *ptr : val</code> (using a signed comparison)</li>
4872 <li>umax: <code>*ptr = *ptr > val ? *ptr : val</code> (using an unsigned comparison)</li>
4873 <li>umin: <code>*ptr = *ptr < val ? *ptr : val</code> (using an unsigned comparison)</li>
4878 %old = atomicrmw add i32* %ptr, i32 1 acquire <i>; yields {i32}</i>
4883 <!-- _______________________________________________________________________ -->
4885 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
4892 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
4893 <result> = getelementptr inbounds <pty>* <ptrval>{, <ty> <idx>}*
4897 <p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
4898 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4899 It performs address calculation only and does not access memory.</p>
4902 <p>The first argument is always a pointer, and forms the basis of the
4903 calculation. The remaining arguments are indices that indicate which of the
4904 elements of the aggregate object are indexed. The interpretation of each
4905 index is dependent on the type being indexed into. The first index always
4906 indexes the pointer value given as the first argument, the second index
4907 indexes a value of the type pointed to (not necessarily the value directly
4908 pointed to, since the first index can be non-zero), etc. The first type
4909 indexed into must be a pointer value, subsequent types can be arrays,
4910 vectors, and structs. Note that subsequent types being indexed into
4911 can never be pointers, since that would require loading the pointer before
4912 continuing calculation.</p>
4914 <p>The type of each index argument depends on the type it is indexing into.
4915 When indexing into a (optionally packed) structure, only <tt>i32</tt>
4916 integer <b>constants</b> are allowed. When indexing into an array, pointer
4917 or vector, integers of any width are allowed, and they are not required to be
4920 <p>For example, let's consider a C code fragment and how it gets compiled to
4923 <pre class="doc_code">
4935 int *foo(struct ST *s) {
4936 return &s[1].Z.B[5][13];
4940 <p>The LLVM code generated by the GCC frontend is:</p>
4942 <pre class="doc_code">
4943 %RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4944 %ST = <a href="#namedtypes">type</a> { i32, double, %RT }
4946 define i32* @foo(%ST* %s) {
4948 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4954 <p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
4955 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4956 }</tt>' type, a structure. The second index indexes into the third element
4957 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4958 i8 }</tt>' type, another structure. The third index indexes into the second
4959 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4960 array. The two dimensions of the array are subscripted into, yielding an
4961 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4962 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
4964 <p>Note that it is perfectly legal to index partially through a structure,
4965 returning a pointer to an inner element. Because of this, the LLVM code for
4966 the given testcase is equivalent to:</p>
4969 define i32* @foo(%ST* %s) {
4970 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
4971 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4972 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
4973 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4974 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4979 <p>If the <tt>inbounds</tt> keyword is present, the result value of the
4980 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4981 base pointer is not an <i>in bounds</i> address of an allocated object,
4982 or if any of the addresses that would be formed by successive addition of
4983 the offsets implied by the indices to the base address with infinitely
4984 precise arithmetic are not an <i>in bounds</i> address of that allocated
4985 object. The <i>in bounds</i> addresses for an allocated object are all
4986 the addresses that point into the object, plus the address one byte past
4989 <p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4990 the base address with silently-wrapping two's complement arithmetic, and
4991 the result value of the <tt>getelementptr</tt> may be outside the object
4992 pointed to by the base pointer. The result value may not necessarily be
4993 used to access memory though, even if it happens to point into allocated
4994 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4995 section for more information.</p>
4997 <p>The getelementptr instruction is often confusing. For some more insight into
4998 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
5002 <i>; yields [12 x i8]*:aptr</i>
5003 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5004 <i>; yields i8*:vptr</i>
5005 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
5006 <i>; yields i8*:eptr</i>
5007 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
5008 <i>; yields i32*:iptr</i>
5009 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
5016 <!-- ======================================================================= -->
5018 <a name="convertops">Conversion Operations</a>
5023 <p>The instructions in this category are the conversion instructions (casting)
5024 which all take a single operand and a type. They perform various bit
5025 conversions on the operand.</p>
5027 <!-- _______________________________________________________________________ -->
5029 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
5036 <result> = trunc <ty> <value> to <ty2> <i>; yields ty2</i>
5040 <p>The '<tt>trunc</tt>' instruction truncates its operand to the
5041 type <tt>ty2</tt>.</p>
5044 <p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
5045 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5046 of the same number of integers.
5047 The bit size of the <tt>value</tt> must be larger than
5048 the bit size of the destination type, <tt>ty2</tt>.
5049 Equal sized types are not allowed.</p>
5052 <p>The '<tt>trunc</tt>' instruction truncates the high order bits
5053 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
5054 source size must be larger than the destination size, <tt>trunc</tt> cannot
5055 be a <i>no-op cast</i>. It will always truncate bits.</p>
5059 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
5060 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
5061 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
5062 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> <i>; yields <i8 8, i8 7></i>
5067 <!-- _______________________________________________________________________ -->
5069 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
5076 <result> = zext <ty> <value> to <ty2> <i>; yields ty2</i>
5080 <p>The '<tt>zext</tt>' instruction zero extends its operand to type
5085 <p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
5086 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5087 of the same number of integers.
5088 The bit size of the <tt>value</tt> must be smaller than
5089 the bit size of the destination type,
5093 <p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
5094 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
5096 <p>When zero extending from i1, the result will always be either 0 or 1.</p>
5100 %X = zext i32 257 to i64 <i>; yields i64:257</i>
5101 %Y = zext i1 true to i32 <i>; yields i32:1</i>
5102 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> <i>; yields <i32 8, i32 7></i>
5107 <!-- _______________________________________________________________________ -->
5109 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
5116 <result> = sext <ty> <value> to <ty2> <i>; yields ty2</i>
5120 <p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
5123 <p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
5124 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5125 of the same number of integers.
5126 The bit size of the <tt>value</tt> must be smaller than
5127 the bit size of the destination type,
5131 <p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
5132 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
5133 of the type <tt>ty2</tt>.</p>
5135 <p>When sign extending from i1, the extension always results in -1 or 0.</p>
5139 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
5140 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
5141 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> <i>; yields <i32 8, i32 7></i>
5146 <!-- _______________________________________________________________________ -->
5148 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
5155 <result> = fptrunc <ty> <value> to <ty2> <i>; yields ty2</i>
5159 <p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
5163 <p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
5164 point</a> value to cast and a <a href="#t_floating">floating point</a> type
5165 to cast it to. The size of <tt>value</tt> must be larger than the size of
5166 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
5167 <i>no-op cast</i>.</p>
5170 <p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
5171 <a href="#t_floating">floating point</a> type to a smaller
5172 <a href="#t_floating">floating point</a> type. If the value cannot fit
5173 within the destination type, <tt>ty2</tt>, then the results are
5178 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
5179 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
5184 <!-- _______________________________________________________________________ -->
5186 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
5193 <result> = fpext <ty> <value> to <ty2> <i>; yields ty2</i>
5197 <p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
5198 floating point value.</p>
5201 <p>The '<tt>fpext</tt>' instruction takes a
5202 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
5203 a <a href="#t_floating">floating point</a> type to cast it to. The source
5204 type must be smaller than the destination type.</p>
5207 <p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
5208 <a href="#t_floating">floating point</a> type to a larger
5209 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
5210 used to make a <i>no-op cast</i> because it always changes bits. Use
5211 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
5215 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
5216 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
5221 <!-- _______________________________________________________________________ -->
5223 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
5230 <result> = fptoui <ty> <value> to <ty2> <i>; yields ty2</i>
5234 <p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
5235 unsigned integer equivalent of type <tt>ty2</tt>.</p>
5238 <p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
5239 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5240 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5241 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5242 vector integer type with the same number of elements as <tt>ty</tt></p>
5245 <p>The '<tt>fptoui</tt>' instruction converts its
5246 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5247 towards zero) unsigned integer value. If the value cannot fit
5248 in <tt>ty2</tt>, the results are undefined.</p>
5252 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
5253 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
5254 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
5259 <!-- _______________________________________________________________________ -->
5261 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
5268 <result> = fptosi <ty> <value> to <ty2> <i>; yields ty2</i>
5272 <p>The '<tt>fptosi</tt>' instruction converts
5273 <a href="#t_floating">floating point</a> <tt>value</tt> to
5274 type <tt>ty2</tt>.</p>
5277 <p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5278 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5279 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5280 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5281 vector integer type with the same number of elements as <tt>ty</tt></p>
5284 <p>The '<tt>fptosi</tt>' instruction converts its
5285 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5286 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5287 the results are undefined.</p>
5291 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
5292 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
5293 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
5298 <!-- _______________________________________________________________________ -->
5300 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
5307 <result> = uitofp <ty> <value> to <ty2> <i>; yields ty2</i>
5311 <p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
5312 integer and converts that value to the <tt>ty2</tt> type.</p>
5315 <p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
5316 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5317 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5318 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5319 floating point type with the same number of elements as <tt>ty</tt></p>
5322 <p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
5323 integer quantity and converts it to the corresponding floating point
5324 value. If the value cannot fit in the floating point value, the results are
5329 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
5330 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
5335 <!-- _______________________________________________________________________ -->
5337 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
5344 <result> = sitofp <ty> <value> to <ty2> <i>; yields ty2</i>
5348 <p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5349 and converts that value to the <tt>ty2</tt> type.</p>
5352 <p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
5353 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5354 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5355 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5356 floating point type with the same number of elements as <tt>ty</tt></p>
5359 <p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5360 quantity and converts it to the corresponding floating point value. If the
5361 value cannot fit in the floating point value, the results are undefined.</p>
5365 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
5366 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
5371 <!-- _______________________________________________________________________ -->
5373 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
5380 <result> = ptrtoint <ty> <value> to <ty2> <i>; yields ty2</i>
5384 <p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
5385 the integer type <tt>ty2</tt>.</p>
5388 <p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
5389 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
5390 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
5393 <p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
5394 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5395 truncating or zero extending that value to the size of the integer type. If
5396 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5397 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5398 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5403 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
5404 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
5409 <!-- _______________________________________________________________________ -->
5411 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
5418 <result> = inttoptr <ty> <value> to <ty2> <i>; yields ty2</i>
5422 <p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5423 pointer type, <tt>ty2</tt>.</p>
5426 <p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
5427 value to cast, and a type to cast it to, which must be a
5428 <a href="#t_pointer">pointer</a> type.</p>
5431 <p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
5432 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5433 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5434 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5435 than the size of a pointer then a zero extension is done. If they are the
5436 same size, nothing is done (<i>no-op cast</i>).</p>
5440 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
5441 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5442 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
5447 <!-- _______________________________________________________________________ -->
5449 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
5456 <result> = bitcast <ty> <value> to <ty2> <i>; yields ty2</i>
5460 <p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
5461 <tt>ty2</tt> without changing any bits.</p>
5464 <p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5465 non-aggregate first class value, and a type to cast it to, which must also be
5466 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5467 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5468 identical. If the source type is a pointer, the destination type must also be
5469 a pointer. This instruction supports bitwise conversion of vectors to
5470 integers and to vectors of other types (as long as they have the same
5474 <p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
5475 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5476 this conversion. The conversion is done as if the <tt>value</tt> had been
5477 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5478 be converted to other pointer types with this instruction. To convert
5479 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5480 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
5484 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
5485 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
5486 %Z = bitcast <2 x int> %V to i64; <i>; yields i64: %V</i>
5493 <!-- ======================================================================= -->
5495 <a name="otherops">Other Operations</a>
5500 <p>The instructions in this category are the "miscellaneous" instructions, which
5501 defy better classification.</p>
5503 <!-- _______________________________________________________________________ -->
5505 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5512 <result> = icmp <cond> <ty> <op1>, <op2> <i>; yields {i1} or {<N x i1>}:result</i>
5516 <p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5517 boolean values based on comparison of its two integer, integer vector, or
5518 pointer operands.</p>
5521 <p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
5522 the condition code indicating the kind of comparison to perform. It is not a
5523 value, just a keyword. The possible condition code are:</p>
5526 <li><tt>eq</tt>: equal</li>
5527 <li><tt>ne</tt>: not equal </li>
5528 <li><tt>ugt</tt>: unsigned greater than</li>
5529 <li><tt>uge</tt>: unsigned greater or equal</li>
5530 <li><tt>ult</tt>: unsigned less than</li>
5531 <li><tt>ule</tt>: unsigned less or equal</li>
5532 <li><tt>sgt</tt>: signed greater than</li>
5533 <li><tt>sge</tt>: signed greater or equal</li>
5534 <li><tt>slt</tt>: signed less than</li>
5535 <li><tt>sle</tt>: signed less or equal</li>
5538 <p>The remaining two arguments must be <a href="#t_integer">integer</a> or
5539 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5540 typed. They must also be identical types.</p>
5543 <p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5544 condition code given as <tt>cond</tt>. The comparison performed always yields
5545 either an <a href="#t_integer"><tt>i1</tt></a> or vector of <tt>i1</tt>
5546 result, as follows:</p>
5549 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
5550 <tt>false</tt> otherwise. No sign interpretation is necessary or
5553 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
5554 <tt>false</tt> otherwise. No sign interpretation is necessary or
5557 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
5558 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5560 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
5561 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5562 to <tt>op2</tt>.</li>
5564 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
5565 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5567 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
5568 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5570 <li><tt>sgt</tt>: interprets the operands as signed values and yields
5571 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5573 <li><tt>sge</tt>: interprets the operands as signed values and yields
5574 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5575 to <tt>op2</tt>.</li>
5577 <li><tt>slt</tt>: interprets the operands as signed values and yields
5578 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5580 <li><tt>sle</tt>: interprets the operands as signed values and yields
5581 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5584 <p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
5585 values are compared as if they were integers.</p>
5587 <p>If the operands are integer vectors, then they are compared element by
5588 element. The result is an <tt>i1</tt> vector with the same number of elements
5589 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
5593 <result> = icmp eq i32 4, 5 <i>; yields: result=false</i>
5594 <result> = icmp ne float* %X, %X <i>; yields: result=false</i>
5595 <result> = icmp ult i16 4, 5 <i>; yields: result=true</i>
5596 <result> = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5597 <result> = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5598 <result> = icmp sge i16 4, 5 <i>; yields: result=false</i>
5601 <p>Note that the code generator does not yet support vector types with
5602 the <tt>icmp</tt> instruction.</p>
5606 <!-- _______________________________________________________________________ -->
5608 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5615 <result> = fcmp <cond> <ty> <op1>, <op2> <i>; yields {i1} or {<N x i1>}:result</i>
5619 <p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5620 values based on comparison of its operands.</p>
5622 <p>If the operands are floating point scalars, then the result type is a boolean
5623 (<a href="#t_integer"><tt>i1</tt></a>).</p>
5625 <p>If the operands are floating point vectors, then the result type is a vector
5626 of boolean with the same number of elements as the operands being
5630 <p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
5631 the condition code indicating the kind of comparison to perform. It is not a
5632 value, just a keyword. The possible condition code are:</p>
5635 <li><tt>false</tt>: no comparison, always returns false</li>
5636 <li><tt>oeq</tt>: ordered and equal</li>
5637 <li><tt>ogt</tt>: ordered and greater than </li>
5638 <li><tt>oge</tt>: ordered and greater than or equal</li>
5639 <li><tt>olt</tt>: ordered and less than </li>
5640 <li><tt>ole</tt>: ordered and less than or equal</li>
5641 <li><tt>one</tt>: ordered and not equal</li>
5642 <li><tt>ord</tt>: ordered (no nans)</li>
5643 <li><tt>ueq</tt>: unordered or equal</li>
5644 <li><tt>ugt</tt>: unordered or greater than </li>
5645 <li><tt>uge</tt>: unordered or greater than or equal</li>
5646 <li><tt>ult</tt>: unordered or less than </li>
5647 <li><tt>ule</tt>: unordered or less than or equal</li>
5648 <li><tt>une</tt>: unordered or not equal</li>
5649 <li><tt>uno</tt>: unordered (either nans)</li>
5650 <li><tt>true</tt>: no comparison, always returns true</li>
5653 <p><i>Ordered</i> means that neither operand is a QNAN while
5654 <i>unordered</i> means that either operand may be a QNAN.</p>
5656 <p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5657 a <a href="#t_floating">floating point</a> type or
5658 a <a href="#t_vector">vector</a> of floating point type. They must have
5659 identical types.</p>
5662 <p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
5663 according to the condition code given as <tt>cond</tt>. If the operands are
5664 vectors, then the vectors are compared element by element. Each comparison
5665 performed always yields an <a href="#t_integer">i1</a> result, as
5669 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
5671 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
5672 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5674 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
5675 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5677 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
5678 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5680 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
5681 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5683 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
5684 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5686 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
5687 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5689 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
5691 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
5692 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5694 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
5695 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5697 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
5698 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5700 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
5701 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5703 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
5704 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5706 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
5707 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5709 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
5711 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5716 <result> = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
5717 <result> = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5718 <result> = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5719 <result> = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
5722 <p>Note that the code generator does not yet support vector types with
5723 the <tt>fcmp</tt> instruction.</p>
5727 <!-- _______________________________________________________________________ -->
5729 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
5736 <result> = phi <ty> [ <val0>, <label0>], ...
5740 <p>The '<tt>phi</tt>' instruction is used to implement the φ node in the
5741 SSA graph representing the function.</p>
5744 <p>The type of the incoming values is specified with the first type field. After
5745 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5746 one pair for each predecessor basic block of the current block. Only values
5747 of <a href="#t_firstclass">first class</a> type may be used as the value
5748 arguments to the PHI node. Only labels may be used as the label
5751 <p>There must be no non-phi instructions between the start of a basic block and
5752 the PHI instructions: i.e. PHI instructions must be first in a basic
5755 <p>For the purposes of the SSA form, the use of each incoming value is deemed to
5756 occur on the edge from the corresponding predecessor block to the current
5757 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5758 value on the same edge).</p>
5761 <p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
5762 specified by the pair corresponding to the predecessor basic block that
5763 executed just prior to the current block.</p>
5767 Loop: ; Infinite loop that counts from 0 on up...
5768 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5769 %nextindvar = add i32 %indvar, 1
5775 <!-- _______________________________________________________________________ -->
5777 <a name="i_select">'<tt>select</tt>' Instruction</a>
5784 <result> = select <i>selty</i> <cond>, <ty> <val1>, <ty> <val2> <i>; yields ty</i>
5786 <i>selty</i> is either i1 or {<N x i1>}
5790 <p>The '<tt>select</tt>' instruction is used to choose one value based on a
5791 condition, without branching.</p>
5795 <p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5796 values indicating the condition, and two values of the
5797 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5798 vectors and the condition is a scalar, then entire vectors are selected, not
5799 individual elements.</p>
5802 <p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5803 first value argument; otherwise, it returns the second value argument.</p>
5805 <p>If the condition is a vector of i1, then the value arguments must be vectors
5806 of the same size, and the selection is done element by element.</p>
5810 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
5813 <p>Note that the code generator does not yet support conditions
5814 with vector type.</p>
5818 <!-- _______________________________________________________________________ -->
5820 <a name="i_call">'<tt>call</tt>' Instruction</a>
5827 <result> = [tail] call [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] <ty> [<fnty>*] <fnptrval>(<function args>) [<a href="#fnattrs">fn attrs</a>]
5831 <p>The '<tt>call</tt>' instruction represents a simple function call.</p>
5834 <p>This instruction requires several arguments:</p>
5837 <li>The optional "tail" marker indicates that the callee function does not
5838 access any allocas or varargs in the caller. Note that calls may be
5839 marked "tail" even if they do not occur before
5840 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5841 present, the function call is eligible for tail call optimization,
5842 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
5843 optimized into a jump</a>. The code generator may optimize calls marked
5844 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5845 sibling call optimization</a> when the caller and callee have
5846 matching signatures, or 2) forced tail call optimization when the
5847 following extra requirements are met:
5849 <li>Caller and callee both have the calling
5850 convention <tt>fastcc</tt>.</li>
5851 <li>The call is in tail position (ret immediately follows call and ret
5852 uses value of call or is void).</li>
5853 <li>Option <tt>-tailcallopt</tt> is enabled,
5854 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
5855 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5856 constraints are met.</a></li>
5860 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5861 convention</a> the call should use. If none is specified, the call
5862 defaults to using C calling conventions. The calling convention of the
5863 call must match the calling convention of the target function, or else the
5864 behavior is undefined.</li>
5866 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5867 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5868 '<tt>inreg</tt>' attributes are valid here.</li>
5870 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5871 type of the return value. Functions that return no value are marked
5872 <tt><a href="#t_void">void</a></tt>.</li>
5874 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5875 being invoked. The argument types must match the types implied by this
5876 signature. This type can be omitted if the function is not varargs and if
5877 the function type does not return a pointer to a function.</li>
5879 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5880 be invoked. In most cases, this is a direct function invocation, but
5881 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5882 to function value.</li>
5884 <li>'<tt>function args</tt>': argument list whose types match the function
5885 signature argument types and parameter attributes. All arguments must be
5886 of <a href="#t_firstclass">first class</a> type. If the function
5887 signature indicates the function accepts a variable number of arguments,
5888 the extra arguments can be specified.</li>
5890 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5891 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5892 '<tt>readnone</tt>' attributes are valid here.</li>
5896 <p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5897 a specified function, with its incoming arguments bound to the specified
5898 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5899 function, control flow continues with the instruction after the function
5900 call, and the return value of the function is bound to the result
5905 %retval = call i32 @test(i32 %argc)
5906 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
5907 %X = tail call i32 @foo() <i>; yields i32</i>
5908 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5909 call void %foo(i8 97 signext)
5911 %struct.A = type { i32, i8 }
5912 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
5913 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5914 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
5915 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
5916 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
5919 <p>llvm treats calls to some functions with names and arguments that match the
5920 standard C99 library as being the C99 library functions, and may perform
5921 optimizations or generate code for them under that assumption. This is
5922 something we'd like to change in the future to provide better support for
5923 freestanding environments and non-C-based languages.</p>
5927 <!-- _______________________________________________________________________ -->
5929 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
5936 <resultval> = va_arg <va_list*> <arglist>, <argty>
5940 <p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
5941 the "variable argument" area of a function call. It is used to implement the
5942 <tt>va_arg</tt> macro in C.</p>
5945 <p>This instruction takes a <tt>va_list*</tt> value and the type of the
5946 argument. It returns a value of the specified argument type and increments
5947 the <tt>va_list</tt> to point to the next argument. The actual type
5948 of <tt>va_list</tt> is target specific.</p>
5951 <p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5952 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5953 to the next argument. For more information, see the variable argument
5954 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
5956 <p>It is legal for this instruction to be called in a function which does not
5957 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5960 <p><tt>va_arg</tt> is an LLVM instruction instead of
5961 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5965 <p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5967 <p>Note that the code generator does not yet fully support va_arg on many
5968 targets. Also, it does not currently support va_arg with aggregate types on
5977 <!-- *********************************************************************** -->
5978 <h2><a name="intrinsics">Intrinsic Functions</a></h2>
5979 <!-- *********************************************************************** -->
5983 <p>LLVM supports the notion of an "intrinsic function". These functions have
5984 well known names and semantics and are required to follow certain
5985 restrictions. Overall, these intrinsics represent an extension mechanism for
5986 the LLVM language that does not require changing all of the transformations
5987 in LLVM when adding to the language (or the bitcode reader/writer, the
5988 parser, etc...).</p>
5990 <p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
5991 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5992 begin with this prefix. Intrinsic functions must always be external
5993 functions: you cannot define the body of intrinsic functions. Intrinsic
5994 functions may only be used in call or invoke instructions: it is illegal to
5995 take the address of an intrinsic function. Additionally, because intrinsic
5996 functions are part of the LLVM language, it is required if any are added that
5997 they be documented here.</p>
5999 <p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
6000 family of functions that perform the same operation but on different data
6001 types. Because LLVM can represent over 8 million different integer types,
6002 overloading is used commonly to allow an intrinsic function to operate on any
6003 integer type. One or more of the argument types or the result type can be
6004 overloaded to accept any integer type. Argument types may also be defined as
6005 exactly matching a previous argument's type or the result type. This allows
6006 an intrinsic function which accepts multiple arguments, but needs all of them
6007 to be of the same type, to only be overloaded with respect to a single
6008 argument or the result.</p>
6010 <p>Overloaded intrinsics will have the names of its overloaded argument types
6011 encoded into its function name, each preceded by a period. Only those types
6012 which are overloaded result in a name suffix. Arguments whose type is matched
6013 against another type do not. For example, the <tt>llvm.ctpop</tt> function
6014 can take an integer of any width and returns an integer of exactly the same
6015 integer width. This leads to a family of functions such as
6016 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
6017 %val)</tt>. Only one type, the return type, is overloaded, and only one type
6018 suffix is required. Because the argument's type is matched against the return
6019 type, it does not require its own name suffix.</p>
6021 <p>To learn how to add an intrinsic function, please see the
6022 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
6024 <!-- ======================================================================= -->
6026 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
6031 <p>Variable argument support is defined in LLVM with
6032 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
6033 intrinsic functions. These functions are related to the similarly named
6034 macros defined in the <tt><stdarg.h></tt> header file.</p>
6036 <p>All of these functions operate on arguments that use a target-specific value
6037 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
6038 not define what this type is, so all transformations should be prepared to
6039 handle these functions regardless of the type used.</p>
6041 <p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
6042 instruction and the variable argument handling intrinsic functions are
6045 <pre class="doc_code">
6046 define i32 @test(i32 %X, ...) {
6047 ; Initialize variable argument processing
6049 %ap2 = bitcast i8** %ap to i8*
6050 call void @llvm.va_start(i8* %ap2)
6052 ; Read a single integer argument
6053 %tmp = va_arg i8** %ap, i32
6055 ; Demonstrate usage of llvm.va_copy and llvm.va_end
6057 %aq2 = bitcast i8** %aq to i8*
6058 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
6059 call void @llvm.va_end(i8* %aq2)
6061 ; Stop processing of arguments.
6062 call void @llvm.va_end(i8* %ap2)
6066 declare void @llvm.va_start(i8*)
6067 declare void @llvm.va_copy(i8*, i8*)
6068 declare void @llvm.va_end(i8*)
6071 <!-- _______________________________________________________________________ -->
6073 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
6081 declare void %llvm.va_start(i8* <arglist>)
6085 <p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*<arglist></tt>
6086 for subsequent use by <tt><a href="#i_va_arg">va_arg</a></tt>.</p>
6089 <p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
6092 <p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
6093 macro available in C. In a target-dependent way, it initializes
6094 the <tt>va_list</tt> element to which the argument points, so that the next
6095 call to <tt>va_arg</tt> will produce the first variable argument passed to
6096 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
6097 need to know the last argument of the function as the compiler can figure
6102 <!-- _______________________________________________________________________ -->
6104 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
6111 declare void @llvm.va_end(i8* <arglist>)
6115 <p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*<arglist></tt>,
6116 which has been initialized previously
6117 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
6118 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
6121 <p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
6124 <p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
6125 macro available in C. In a target-dependent way, it destroys
6126 the <tt>va_list</tt> element to which the argument points. Calls
6127 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
6128 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
6129 with calls to <tt>llvm.va_end</tt>.</p>
6133 <!-- _______________________________________________________________________ -->
6135 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
6142 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
6146 <p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
6147 from the source argument list to the destination argument list.</p>
6150 <p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
6151 The second argument is a pointer to a <tt>va_list</tt> element to copy
6155 <p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
6156 macro available in C. In a target-dependent way, it copies the
6157 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
6158 element. This intrinsic is necessary because
6159 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
6160 arbitrarily complex and require, for example, memory allocation.</p>
6168 <!-- ======================================================================= -->
6170 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
6175 <p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
6176 Collection</a> (GC) requires the implementation and generation of these
6177 intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
6178 roots on the stack</a>, as well as garbage collector implementations that
6179 require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
6180 barriers. Front-ends for type-safe garbage collected languages should generate
6181 these intrinsics to make use of the LLVM garbage collectors. For more details,
6182 see <a href="GarbageCollection.html">Accurate Garbage Collection with
6185 <p>The garbage collection intrinsics only operate on objects in the generic
6186 address space (address space zero).</p>
6188 <!-- _______________________________________________________________________ -->
6190 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
6197 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
6201 <p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
6202 the code generator, and allows some metadata to be associated with it.</p>
6205 <p>The first argument specifies the address of a stack object that contains the
6206 root pointer. The second pointer (which must be either a constant or a
6207 global value address) contains the meta-data to be associated with the
6211 <p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
6212 location. At compile-time, the code generator generates information to allow
6213 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
6214 intrinsic may only be used in a function which <a href="#gc">specifies a GC
6219 <!-- _______________________________________________________________________ -->
6221 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
6228 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
6232 <p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
6233 locations, allowing garbage collector implementations that require read
6237 <p>The second argument is the address to read from, which should be an address
6238 allocated from the garbage collector. The first object is a pointer to the
6239 start of the referenced object, if needed by the language runtime (otherwise
6243 <p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
6244 instruction, but may be replaced with substantially more complex code by the
6245 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
6246 may only be used in a function which <a href="#gc">specifies a GC
6251 <!-- _______________________________________________________________________ -->
6253 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
6260 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
6264 <p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
6265 locations, allowing garbage collector implementations that require write
6266 barriers (such as generational or reference counting collectors).</p>
6269 <p>The first argument is the reference to store, the second is the start of the
6270 object to store it to, and the third is the address of the field of Obj to
6271 store to. If the runtime does not require a pointer to the object, Obj may
6275 <p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
6276 instruction, but may be replaced with substantially more complex code by the
6277 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6278 may only be used in a function which <a href="#gc">specifies a GC
6285 <!-- ======================================================================= -->
6287 <a name="int_codegen">Code Generator Intrinsics</a>
6292 <p>These intrinsics are provided by LLVM to expose special features that may
6293 only be implemented with code generator support.</p>
6295 <!-- _______________________________________________________________________ -->
6297 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
6304 declare i8 *@llvm.returnaddress(i32 <level>)
6308 <p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6309 target-specific value indicating the return address of the current function
6310 or one of its callers.</p>
6313 <p>The argument to this intrinsic indicates which function to return the address
6314 for. Zero indicates the calling function, one indicates its caller, etc.
6315 The argument is <b>required</b> to be a constant integer value.</p>
6318 <p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6319 indicating the return address of the specified call frame, or zero if it
6320 cannot be identified. The value returned by this intrinsic is likely to be
6321 incorrect or 0 for arguments other than zero, so it should only be used for
6322 debugging purposes.</p>
6324 <p>Note that calling this intrinsic does not prevent function inlining or other
6325 aggressive transformations, so the value returned may not be that of the
6326 obvious source-language caller.</p>
6330 <!-- _______________________________________________________________________ -->
6332 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
6339 declare i8* @llvm.frameaddress(i32 <level>)
6343 <p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6344 target-specific frame pointer value for the specified stack frame.</p>
6347 <p>The argument to this intrinsic indicates which function to return the frame
6348 pointer for. Zero indicates the calling function, one indicates its caller,
6349 etc. The argument is <b>required</b> to be a constant integer value.</p>
6352 <p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6353 indicating the frame address of the specified call frame, or zero if it
6354 cannot be identified. The value returned by this intrinsic is likely to be
6355 incorrect or 0 for arguments other than zero, so it should only be used for
6356 debugging purposes.</p>
6358 <p>Note that calling this intrinsic does not prevent function inlining or other
6359 aggressive transformations, so the value returned may not be that of the
6360 obvious source-language caller.</p>
6364 <!-- _______________________________________________________________________ -->
6366 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
6373 declare i8* @llvm.stacksave()
6377 <p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6378 of the function stack, for use
6379 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6380 useful for implementing language features like scoped automatic variable
6381 sized arrays in C99.</p>
6384 <p>This intrinsic returns a opaque pointer value that can be passed
6385 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6386 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6387 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6388 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6389 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6390 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
6394 <!-- _______________________________________________________________________ -->
6396 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
6403 declare void @llvm.stackrestore(i8* %ptr)
6407 <p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6408 the function stack to the state it was in when the
6409 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6410 executed. This is useful for implementing language features like scoped
6411 automatic variable sized arrays in C99.</p>
6414 <p>See the description
6415 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
6419 <!-- _______________________________________________________________________ -->
6421 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
6428 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
6432 <p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6433 insert a prefetch instruction if supported; otherwise, it is a noop.
6434 Prefetches have no effect on the behavior of the program but can change its
6435 performance characteristics.</p>
6438 <p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6439 specifier determining if the fetch should be for a read (0) or write (1),
6440 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
6441 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6442 specifies whether the prefetch is performed on the data (1) or instruction (0)
6443 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6444 must be constant integers.</p>
6447 <p>This intrinsic does not modify the behavior of the program. In particular,
6448 prefetches cannot trap and do not produce a value. On targets that support
6449 this intrinsic, the prefetch can provide hints to the processor cache for
6450 better performance.</p>
6454 <!-- _______________________________________________________________________ -->
6456 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
6463 declare void @llvm.pcmarker(i32 <id>)
6467 <p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6468 Counter (PC) in a region of code to simulators and other tools. The method
6469 is target specific, but it is expected that the marker will use exported
6470 symbols to transmit the PC of the marker. The marker makes no guarantees
6471 that it will remain with any specific instruction after optimizations. It is
6472 possible that the presence of a marker will inhibit optimizations. The
6473 intended use is to be inserted after optimizations to allow correlations of
6474 simulation runs.</p>
6477 <p><tt>id</tt> is a numerical id identifying the marker.</p>
6480 <p>This intrinsic does not modify the behavior of the program. Backends that do
6481 not support this intrinsic may ignore it.</p>
6485 <!-- _______________________________________________________________________ -->
6487 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
6494 declare i64 @llvm.readcyclecounter()
6498 <p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6499 counter register (or similar low latency, high accuracy clocks) on those
6500 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6501 should map to RPCC. As the backing counters overflow quickly (on the order
6502 of 9 seconds on alpha), this should only be used for small timings.</p>
6505 <p>When directly supported, reading the cycle counter should not modify any
6506 memory. Implementations are allowed to either return a application specific
6507 value or a system wide value. On backends without support, this is lowered
6508 to a constant 0.</p>
6514 <!-- ======================================================================= -->
6516 <a name="int_libc">Standard C Library Intrinsics</a>
6521 <p>LLVM provides intrinsics for a few important standard C library functions.
6522 These intrinsics allow source-language front-ends to pass information about
6523 the alignment of the pointer arguments to the code generator, providing
6524 opportunity for more efficient code generation.</p>
6526 <!-- _______________________________________________________________________ -->
6528 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
6534 <p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
6535 integer bit width and for different address spaces. Not all targets support
6536 all bit widths however.</p>
6539 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
6540 i32 <len>, i32 <align>, i1 <isvolatile>)
6541 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
6542 i64 <len>, i32 <align>, i1 <isvolatile>)
6546 <p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6547 source location to the destination location.</p>
6549 <p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
6550 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6551 and the pointers can be in specified address spaces.</p>
6555 <p>The first argument is a pointer to the destination, the second is a pointer
6556 to the source. The third argument is an integer argument specifying the
6557 number of bytes to copy, the fourth argument is the alignment of the
6558 source and destination locations, and the fifth is a boolean indicating a
6559 volatile access.</p>
6561 <p>If the call to this intrinsic has an alignment value that is not 0 or 1,
6562 then the caller guarantees that both the source and destination pointers are
6563 aligned to that boundary.</p>
6565 <p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6566 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6567 The detailed access behavior is not very cleanly specified and it is unwise
6568 to depend on it.</p>
6572 <p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6573 source location to the destination location, which are not allowed to
6574 overlap. It copies "len" bytes of memory over. If the argument is known to
6575 be aligned to some boundary, this can be specified as the fourth argument,
6576 otherwise it should be set to 0 or 1.</p>
6580 <!-- _______________________________________________________________________ -->
6582 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
6588 <p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
6589 width and for different address space. Not all targets support all bit
6593 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
6594 i32 <len>, i32 <align>, i1 <isvolatile>)
6595 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
6596 i64 <len>, i32 <align>, i1 <isvolatile>)
6600 <p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6601 source location to the destination location. It is similar to the
6602 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6605 <p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
6606 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6607 and the pointers can be in specified address spaces.</p>
6611 <p>The first argument is a pointer to the destination, the second is a pointer
6612 to the source. The third argument is an integer argument specifying the
6613 number of bytes to copy, the fourth argument is the alignment of the
6614 source and destination locations, and the fifth is a boolean indicating a
6615 volatile access.</p>
6617 <p>If the call to this intrinsic has an alignment value that is not 0 or 1,
6618 then the caller guarantees that the source and destination pointers are
6619 aligned to that boundary.</p>
6621 <p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6622 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6623 The detailed access behavior is not very cleanly specified and it is unwise
6624 to depend on it.</p>
6628 <p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6629 source location to the destination location, which may overlap. It copies
6630 "len" bytes of memory over. If the argument is known to be aligned to some
6631 boundary, this can be specified as the fourth argument, otherwise it should
6632 be set to 0 or 1.</p>
6636 <!-- _______________________________________________________________________ -->
6638 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
6644 <p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
6645 width and for different address spaces. However, not all targets support all
6649 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
6650 i32 <len>, i32 <align>, i1 <isvolatile>)
6651 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
6652 i64 <len>, i32 <align>, i1 <isvolatile>)
6656 <p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6657 particular byte value.</p>
6659 <p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
6660 intrinsic does not return a value and takes extra alignment/volatile
6661 arguments. Also, the destination can be in an arbitrary address space.</p>
6664 <p>The first argument is a pointer to the destination to fill, the second is the
6665 byte value with which to fill it, the third argument is an integer argument
6666 specifying the number of bytes to fill, and the fourth argument is the known
6667 alignment of the destination location.</p>
6669 <p>If the call to this intrinsic has an alignment value that is not 0 or 1,
6670 then the caller guarantees that the destination pointer is aligned to that
6673 <p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6674 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6675 The detailed access behavior is not very cleanly specified and it is unwise
6676 to depend on it.</p>
6679 <p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6680 at the destination location. If the argument is known to be aligned to some
6681 boundary, this can be specified as the fourth argument, otherwise it should
6682 be set to 0 or 1.</p>
6686 <!-- _______________________________________________________________________ -->
6688 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
6694 <p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6695 floating point or vector of floating point type. Not all targets support all
6699 declare float @llvm.sqrt.f32(float %Val)
6700 declare double @llvm.sqrt.f64(double %Val)
6701 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6702 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6703 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
6707 <p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6708 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6709 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6710 behavior for negative numbers other than -0.0 (which allows for better
6711 optimization, because there is no need to worry about errno being
6712 set). <tt>llvm.sqrt(-0.0)</tt> is defined to return -0.0 like IEEE sqrt.</p>
6715 <p>The argument and return value are floating point numbers of the same
6719 <p>This function returns the sqrt of the specified operand if it is a
6720 nonnegative floating point number.</p>
6724 <!-- _______________________________________________________________________ -->
6726 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
6732 <p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6733 floating point or vector of floating point type. Not all targets support all
6737 declare float @llvm.powi.f32(float %Val, i32 %power)
6738 declare double @llvm.powi.f64(double %Val, i32 %power)
6739 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6740 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6741 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
6745 <p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6746 specified (positive or negative) power. The order of evaluation of
6747 multiplications is not defined. When a vector of floating point type is
6748 used, the second argument remains a scalar integer value.</p>
6751 <p>The second argument is an integer power, and the first is a value to raise to
6755 <p>This function returns the first value raised to the second power with an
6756 unspecified sequence of rounding operations.</p>
6760 <!-- _______________________________________________________________________ -->
6762 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
6768 <p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6769 floating point or vector of floating point type. Not all targets support all
6773 declare float @llvm.sin.f32(float %Val)
6774 declare double @llvm.sin.f64(double %Val)
6775 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6776 declare fp128 @llvm.sin.f128(fp128 %Val)
6777 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6781 <p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
6784 <p>The argument and return value are floating point numbers of the same
6788 <p>This function returns the sine of the specified operand, returning the same
6789 values as the libm <tt>sin</tt> functions would, and handles error conditions
6790 in the same way.</p>
6794 <!-- _______________________________________________________________________ -->
6796 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
6802 <p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6803 floating point or vector of floating point type. Not all targets support all
6807 declare float @llvm.cos.f32(float %Val)
6808 declare double @llvm.cos.f64(double %Val)
6809 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6810 declare fp128 @llvm.cos.f128(fp128 %Val)
6811 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6815 <p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
6818 <p>The argument and return value are floating point numbers of the same
6822 <p>This function returns the cosine of the specified operand, returning the same
6823 values as the libm <tt>cos</tt> functions would, and handles error conditions
6824 in the same way.</p>
6828 <!-- _______________________________________________________________________ -->
6830 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
6836 <p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6837 floating point or vector of floating point type. Not all targets support all
6841 declare float @llvm.pow.f32(float %Val, float %Power)
6842 declare double @llvm.pow.f64(double %Val, double %Power)
6843 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6844 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6845 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6849 <p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6850 specified (positive or negative) power.</p>
6853 <p>The second argument is a floating point power, and the first is a value to
6854 raise to that power.</p>
6857 <p>This function returns the first value raised to the second power, returning
6858 the same values as the libm <tt>pow</tt> functions would, and handles error
6859 conditions in the same way.</p>
6865 <!-- _______________________________________________________________________ -->
6867 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
6873 <p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
6874 floating point or vector of floating point type. Not all targets support all
6878 declare float @llvm.exp.f32(float %Val)
6879 declare double @llvm.exp.f64(double %Val)
6880 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
6881 declare fp128 @llvm.exp.f128(fp128 %Val)
6882 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
6886 <p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
6889 <p>The argument and return value are floating point numbers of the same
6893 <p>This function returns the same values as the libm <tt>exp</tt> functions
6894 would, and handles error conditions in the same way.</p>
6898 <!-- _______________________________________________________________________ -->
6900 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
6906 <p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
6907 floating point or vector of floating point type. Not all targets support all
6911 declare float @llvm.log.f32(float %Val)
6912 declare double @llvm.log.f64(double %Val)
6913 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
6914 declare fp128 @llvm.log.f128(fp128 %Val)
6915 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
6919 <p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
6922 <p>The argument and return value are floating point numbers of the same
6926 <p>This function returns the same values as the libm <tt>log</tt> functions
6927 would, and handles error conditions in the same way.</p>
6930 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
6936 <p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
6937 floating point or vector of floating point type. Not all targets support all
6941 declare float @llvm.fma.f32(float %a, float %b, float %c)
6942 declare double @llvm.fma.f64(double %a, double %b, double %c)
6943 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
6944 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
6945 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
6949 <p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
6953 <p>The argument and return value are floating point numbers of the same
6957 <p>This function returns the same values as the libm <tt>fma</tt> functions
6962 <!-- ======================================================================= -->
6964 <a name="int_manip">Bit Manipulation Intrinsics</a>
6969 <p>LLVM provides intrinsics for a few important bit manipulation operations.
6970 These allow efficient code generation for some algorithms.</p>
6972 <!-- _______________________________________________________________________ -->
6974 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
6980 <p>This is an overloaded intrinsic function. You can use bswap on any integer
6981 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6984 declare i16 @llvm.bswap.i16(i16 <id>)
6985 declare i32 @llvm.bswap.i32(i32 <id>)
6986 declare i64 @llvm.bswap.i64(i64 <id>)
6990 <p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6991 values with an even number of bytes (positive multiple of 16 bits). These
6992 are useful for performing operations on data that is not in the target's
6993 native byte order.</p>
6996 <p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6997 and low byte of the input i16 swapped. Similarly,
6998 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6999 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
7000 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
7001 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
7002 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
7003 more, respectively).</p>
7007 <!-- _______________________________________________________________________ -->
7009 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
7015 <p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
7016 width, or on any vector with integer elements. Not all targets support all
7017 bit widths or vector types, however.</p>
7020 declare i8 @llvm.ctpop.i8(i8 <src>)
7021 declare i16 @llvm.ctpop.i16(i16 <src>)
7022 declare i32 @llvm.ctpop.i32(i32 <src>)
7023 declare i64 @llvm.ctpop.i64(i64 <src>)
7024 declare i256 @llvm.ctpop.i256(i256 <src>)
7025 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
7029 <p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
7033 <p>The only argument is the value to be counted. The argument may be of any
7034 integer type, or a vector with integer elements.
7035 The return type must match the argument type.</p>
7038 <p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
7039 element of a vector.</p>
7043 <!-- _______________________________________________________________________ -->
7045 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
7051 <p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
7052 integer bit width, or any vector whose elements are integers. Not all
7053 targets support all bit widths or vector types, however.</p>
7056 declare i8 @llvm.ctlz.i8 (i8 <src>)
7057 declare i16 @llvm.ctlz.i16(i16 <src>)
7058 declare i32 @llvm.ctlz.i32(i32 <src>)
7059 declare i64 @llvm.ctlz.i64(i64 <src>)
7060 declare i256 @llvm.ctlz.i256(i256 <src>)
7061 declare <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src;gt)
7065 <p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
7066 leading zeros in a variable.</p>
7069 <p>The only argument is the value to be counted. The argument may be of any
7070 integer type, or any vector type with integer element type.
7071 The return type must match the argument type.</p>
7074 <p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
7075 zeros in a variable, or within each element of the vector if the operation
7076 is of vector type. If the src == 0 then the result is the size in bits of
7077 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
7081 <!-- _______________________________________________________________________ -->
7083 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
7089 <p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
7090 integer bit width, or any vector of integer elements. Not all targets
7091 support all bit widths or vector types, however.</p>
7094 declare i8 @llvm.cttz.i8 (i8 <src>)
7095 declare i16 @llvm.cttz.i16(i16 <src>)
7096 declare i32 @llvm.cttz.i32(i32 <src>)
7097 declare i64 @llvm.cttz.i64(i64 <src>)
7098 declare i256 @llvm.cttz.i256(i256 <src>)
7099 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>)
7103 <p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
7107 <p>The only argument is the value to be counted. The argument may be of any
7108 integer type, or a vectory with integer element type.. The return type
7109 must match the argument type.</p>
7112 <p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
7113 zeros in a variable, or within each element of a vector.
7114 If the src == 0 then the result is the size in bits of
7115 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
7121 <!-- ======================================================================= -->
7123 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
7128 <p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
7130 <!-- _______________________________________________________________________ -->
7132 <a name="int_sadd_overflow">
7133 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
7140 <p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
7141 on any integer bit width.</p>
7144 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7145 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7146 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7150 <p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
7151 a signed addition of the two arguments, and indicate whether an overflow
7152 occurred during the signed summation.</p>
7155 <p>The arguments (%a and %b) and the first element of the result structure may
7156 be of integer types of any bit width, but they must have the same bit
7157 width. The second element of the result structure must be of
7158 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7159 undergo signed addition.</p>
7162 <p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
7163 a signed addition of the two variables. They return a structure — the
7164 first element of which is the signed summation, and the second element of
7165 which is a bit specifying if the signed summation resulted in an
7170 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7171 %sum = extractvalue {i32, i1} %res, 0
7172 %obit = extractvalue {i32, i1} %res, 1
7173 br i1 %obit, label %overflow, label %normal
7178 <!-- _______________________________________________________________________ -->
7180 <a name="int_uadd_overflow">
7181 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
7188 <p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
7189 on any integer bit width.</p>
7192 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7193 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7194 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7198 <p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
7199 an unsigned addition of the two arguments, and indicate whether a carry
7200 occurred during the unsigned summation.</p>
7203 <p>The arguments (%a and %b) and the first element of the result structure may
7204 be of integer types of any bit width, but they must have the same bit
7205 width. The second element of the result structure must be of
7206 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7207 undergo unsigned addition.</p>
7210 <p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
7211 an unsigned addition of the two arguments. They return a structure —
7212 the first element of which is the sum, and the second element of which is a
7213 bit specifying if the unsigned summation resulted in a carry.</p>
7217 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7218 %sum = extractvalue {i32, i1} %res, 0
7219 %obit = extractvalue {i32, i1} %res, 1
7220 br i1 %obit, label %carry, label %normal
7225 <!-- _______________________________________________________________________ -->
7227 <a name="int_ssub_overflow">
7228 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
7235 <p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
7236 on any integer bit width.</p>
7239 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7240 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7241 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7245 <p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
7246 a signed subtraction of the two arguments, and indicate whether an overflow
7247 occurred during the signed subtraction.</p>
7250 <p>The arguments (%a and %b) and the first element of the result structure may
7251 be of integer types of any bit width, but they must have the same bit
7252 width. The second element of the result structure must be of
7253 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7254 undergo signed subtraction.</p>
7257 <p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
7258 a signed subtraction of the two arguments. They return a structure —
7259 the first element of which is the subtraction, and the second element of
7260 which is a bit specifying if the signed subtraction resulted in an
7265 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7266 %sum = extractvalue {i32, i1} %res, 0
7267 %obit = extractvalue {i32, i1} %res, 1
7268 br i1 %obit, label %overflow, label %normal
7273 <!-- _______________________________________________________________________ -->
7275 <a name="int_usub_overflow">
7276 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7283 <p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
7284 on any integer bit width.</p>
7287 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7288 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7289 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7293 <p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
7294 an unsigned subtraction of the two arguments, and indicate whether an
7295 overflow occurred during the unsigned subtraction.</p>
7298 <p>The arguments (%a and %b) and the first element of the result structure may
7299 be of integer types of any bit width, but they must have the same bit
7300 width. The second element of the result structure must be of
7301 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7302 undergo unsigned subtraction.</p>
7305 <p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
7306 an unsigned subtraction of the two arguments. They return a structure —
7307 the first element of which is the subtraction, and the second element of
7308 which is a bit specifying if the unsigned subtraction resulted in an
7313 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7314 %sum = extractvalue {i32, i1} %res, 0
7315 %obit = extractvalue {i32, i1} %res, 1
7316 br i1 %obit, label %overflow, label %normal
7321 <!-- _______________________________________________________________________ -->
7323 <a name="int_smul_overflow">
7324 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7331 <p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
7332 on any integer bit width.</p>
7335 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7336 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7337 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7342 <p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
7343 a signed multiplication of the two arguments, and indicate whether an
7344 overflow occurred during the signed multiplication.</p>
7347 <p>The arguments (%a and %b) and the first element of the result structure may
7348 be of integer types of any bit width, but they must have the same bit
7349 width. The second element of the result structure must be of
7350 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7351 undergo signed multiplication.</p>
7354 <p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
7355 a signed multiplication of the two arguments. They return a structure —
7356 the first element of which is the multiplication, and the second element of
7357 which is a bit specifying if the signed multiplication resulted in an
7362 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7363 %sum = extractvalue {i32, i1} %res, 0
7364 %obit = extractvalue {i32, i1} %res, 1
7365 br i1 %obit, label %overflow, label %normal
7370 <!-- _______________________________________________________________________ -->
7372 <a name="int_umul_overflow">
7373 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7380 <p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
7381 on any integer bit width.</p>
7384 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7385 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7386 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7390 <p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
7391 a unsigned multiplication of the two arguments, and indicate whether an
7392 overflow occurred during the unsigned multiplication.</p>
7395 <p>The arguments (%a and %b) and the first element of the result structure may
7396 be of integer types of any bit width, but they must have the same bit
7397 width. The second element of the result structure must be of
7398 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7399 undergo unsigned multiplication.</p>
7402 <p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
7403 an unsigned multiplication of the two arguments. They return a structure
7404 — the first element of which is the multiplication, and the second
7405 element of which is a bit specifying if the unsigned multiplication resulted
7410 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7411 %sum = extractvalue {i32, i1} %res, 0
7412 %obit = extractvalue {i32, i1} %res, 1
7413 br i1 %obit, label %overflow, label %normal
7420 <!-- ======================================================================= -->
7422 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
7427 <p>Half precision floating point is a storage-only format. This means that it is
7428 a dense encoding (in memory) but does not support computation in the
7431 <p>This means that code must first load the half-precision floating point
7432 value as an i16, then convert it to float with <a
7433 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7434 Computation can then be performed on the float value (including extending to
7435 double etc). To store the value back to memory, it is first converted to
7436 float if needed, then converted to i16 with
7437 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7438 storing as an i16 value.</p>
7440 <!-- _______________________________________________________________________ -->
7442 <a name="int_convert_to_fp16">
7443 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7451 declare i16 @llvm.convert.to.fp16(f32 %a)
7455 <p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7456 a conversion from single precision floating point format to half precision
7457 floating point format.</p>
7460 <p>The intrinsic function contains single argument - the value to be
7464 <p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7465 a conversion from single precision floating point format to half precision
7466 floating point format. The return value is an <tt>i16</tt> which
7467 contains the converted number.</p>
7471 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7472 store i16 %res, i16* @x, align 2
7477 <!-- _______________________________________________________________________ -->
7479 <a name="int_convert_from_fp16">
7480 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7488 declare f32 @llvm.convert.from.fp16(i16 %a)
7492 <p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7493 a conversion from half precision floating point format to single precision
7494 floating point format.</p>
7497 <p>The intrinsic function contains single argument - the value to be
7501 <p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
7502 conversion from half single precision floating point format to single
7503 precision floating point format. The input half-float value is represented by
7504 an <tt>i16</tt> value.</p>
7508 %a = load i16* @x, align 2
7509 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7516 <!-- ======================================================================= -->
7518 <a name="int_debugger">Debugger Intrinsics</a>
7523 <p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7524 prefix), are described in
7525 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7526 Level Debugging</a> document.</p>
7530 <!-- ======================================================================= -->
7532 <a name="int_eh">Exception Handling Intrinsics</a>
7537 <p>The LLVM exception handling intrinsics (which all start with
7538 <tt>llvm.eh.</tt> prefix), are described in
7539 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7540 Handling</a> document.</p>
7544 <!-- ======================================================================= -->
7546 <a name="int_trampoline">Trampoline Intrinsic</a>
7551 <p>This intrinsic makes it possible to excise one parameter, marked with
7552 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7553 The result is a callable
7554 function pointer lacking the nest parameter - the caller does not need to
7555 provide a value for it. Instead, the value to use is stored in advance in a
7556 "trampoline", a block of memory usually allocated on the stack, which also
7557 contains code to splice the nest value into the argument list. This is used
7558 to implement the GCC nested function address extension.</p>
7560 <p>For example, if the function is
7561 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7562 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7565 <pre class="doc_code">
7566 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7567 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
7568 %p = call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval)
7569 %fp = bitcast i8* %p to i32 (i32, i32)*
7572 <p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7573 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
7575 <!-- _______________________________________________________________________ -->
7578 '<tt>llvm.init.trampoline</tt>' Intrinsic
7586 declare i8* @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
7590 <p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
7591 function pointer suitable for executing it.</p>
7594 <p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7595 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7596 sufficiently aligned block of memory; this memory is written to by the
7597 intrinsic. Note that the size and the alignment are target-specific - LLVM
7598 currently provides no portable way of determining them, so a front-end that
7599 generates this intrinsic needs to have some target-specific knowledge.
7600 The <tt>func</tt> argument must hold a function bitcast to
7601 an <tt>i8*</tt>.</p>
7604 <p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7605 dependent code, turning it into a function. A pointer to this function is
7606 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7607 function pointer type</a> before being called. The new function's signature
7608 is the same as that of <tt>func</tt> with any arguments marked with
7609 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7610 is allowed, and it must be of pointer type. Calling the new function is
7611 equivalent to calling <tt>func</tt> with the same argument list, but
7612 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7613 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7614 by <tt>tramp</tt> is modified, then the effect of any later call to the
7615 returned function pointer is undefined.</p>
7621 <!-- ======================================================================= -->
7623 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
7628 <p>These intrinsic functions expand the "universal IR" of LLVM to represent
7629 hardware constructs for atomic operations and memory synchronization. This
7630 provides an interface to the hardware, not an interface to the programmer. It
7631 is aimed at a low enough level to allow any programming models or APIs
7632 (Application Programming Interfaces) which need atomic behaviors to map
7633 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7634 hardware provides a "universal IR" for source languages, it also provides a
7635 starting point for developing a "universal" atomic operation and
7636 synchronization IR.</p>
7638 <p>These do <em>not</em> form an API such as high-level threading libraries,
7639 software transaction memory systems, atomic primitives, and intrinsic
7640 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7641 application libraries. The hardware interface provided by LLVM should allow
7642 a clean implementation of all of these APIs and parallel programming models.
7643 No one model or paradigm should be selected above others unless the hardware
7644 itself ubiquitously does so.</p>
7646 <!-- _______________________________________________________________________ -->
7648 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
7654 declare void @llvm.memory.barrier(i1 <ll>, i1 <ls>, i1 <sl>, i1 <ss>, i1 <device>)
7658 <p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7659 specific pairs of memory access types.</p>
7662 <p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7663 The first four arguments enables a specific barrier as listed below. The
7664 fifth argument specifies that the barrier applies to io or device or uncached
7668 <li><tt>ll</tt>: load-load barrier</li>
7669 <li><tt>ls</tt>: load-store barrier</li>
7670 <li><tt>sl</tt>: store-load barrier</li>
7671 <li><tt>ss</tt>: store-store barrier</li>
7672 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7676 <p>This intrinsic causes the system to enforce some ordering constraints upon
7677 the loads and stores of the program. This barrier does not
7678 indicate <em>when</em> any events will occur, it only enforces
7679 an <em>order</em> in which they occur. For any of the specified pairs of load
7680 and store operations (f.ex. load-load, or store-load), all of the first
7681 operations preceding the barrier will complete before any of the second
7682 operations succeeding the barrier begin. Specifically the semantics for each
7683 pairing is as follows:</p>
7686 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7687 after the barrier begins.</li>
7688 <li><tt>ls</tt>: All loads before the barrier must complete before any
7689 store after the barrier begins.</li>
7690 <li><tt>ss</tt>: All stores before the barrier must complete before any
7691 store after the barrier begins.</li>
7692 <li><tt>sl</tt>: All stores before the barrier must complete before any
7693 load after the barrier begins.</li>
7696 <p>These semantics are applied with a logical "and" behavior when more than one
7697 is enabled in a single memory barrier intrinsic.</p>
7699 <p>Backends may implement stronger barriers than those requested when they do
7700 not support as fine grained a barrier as requested. Some architectures do
7701 not need all types of barriers and on such architectures, these become
7706 %mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7707 %ptr = bitcast i8* %mallocP to i32*
7710 %result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
7711 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false, i1 true)
7712 <i>; guarantee the above finishes</i>
7713 store i32 8, %ptr <i>; before this begins</i>
7718 <!-- _______________________________________________________________________ -->
7720 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
7726 <p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7727 any integer bit width and for different address spaces. Not all targets
7728 support all bit widths however.</p>
7731 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* <ptr>, i8 <cmp>, i8 <val>)
7732 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* <ptr>, i16 <cmp>, i16 <val>)
7733 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* <ptr>, i32 <cmp>, i32 <val>)
7734 declare i64 @llvm.atomic.cmp.swap.i64.p0i64(i64* <ptr>, i64 <cmp>, i64 <val>)
7738 <p>This loads a value in memory and compares it to a given value. If they are
7739 equal, it stores a new value into the memory.</p>
7742 <p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7743 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7744 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7745 this integer type. While any bit width integer may be used, targets may only
7746 lower representations they support in hardware.</p>
7749 <p>This entire intrinsic must be executed atomically. It first loads the value
7750 in memory pointed to by <tt>ptr</tt> and compares it with the
7751 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7752 memory. The loaded value is yielded in all cases. This provides the
7753 equivalent of an atomic compare-and-swap operation within the SSA
7758 %mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7759 %ptr = bitcast i8* %mallocP to i32*
7762 %val1 = add i32 4, 4
7763 %result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
7764 <i>; yields {i32}:result1 = 4</i>
7765 %stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7766 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7768 %val2 = add i32 1, 1
7769 %result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
7770 <i>; yields {i32}:result2 = 8</i>
7771 %stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7773 %memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7778 <!-- _______________________________________________________________________ -->
7780 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
7786 <p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7787 integer bit width. Not all targets support all bit widths however.</p>
7790 declare i8 @llvm.atomic.swap.i8.p0i8(i8* <ptr>, i8 <val>)
7791 declare i16 @llvm.atomic.swap.i16.p0i16(i16* <ptr>, i16 <val>)
7792 declare i32 @llvm.atomic.swap.i32.p0i32(i32* <ptr>, i32 <val>)
7793 declare i64 @llvm.atomic.swap.i64.p0i64(i64* <ptr>, i64 <val>)
7797 <p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7798 the value from memory. It then stores the value in <tt>val</tt> in the memory
7799 at <tt>ptr</tt>.</p>
7802 <p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7803 the <tt>val</tt> argument and the result must be integers of the same bit
7804 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7805 integer type. The targets may only lower integer representations they
7809 <p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7810 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7811 equivalent of an atomic swap operation within the SSA framework.</p>
7815 %mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7816 %ptr = bitcast i8* %mallocP to i32*
7819 %val1 = add i32 4, 4
7820 %result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
7821 <i>; yields {i32}:result1 = 4</i>
7822 %stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7823 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7825 %val2 = add i32 1, 1
7826 %result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
7827 <i>; yields {i32}:result2 = 8</i>
7829 %stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7830 %memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7835 <!-- _______________________________________________________________________ -->
7837 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
7843 <p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7844 any integer bit width. Not all targets support all bit widths however.</p>
7847 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* <ptr>, i8 <delta>)
7848 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* <ptr>, i16 <delta>)
7849 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* <ptr>, i32 <delta>)
7850 declare i64 @llvm.atomic.load.add.i64.p0i64(i64* <ptr>, i64 <delta>)
7854 <p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7855 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7858 <p>The intrinsic takes two arguments, the first a pointer to an integer value
7859 and the second an integer value. The result is also an integer value. These
7860 integer types can have any bit width, but they must all have the same bit
7861 width. The targets may only lower integer representations they support.</p>
7864 <p>This intrinsic does a series of operations atomically. It first loads the
7865 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7866 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
7870 %mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7871 %ptr = bitcast i8* %mallocP to i32*
7873 %result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
7874 <i>; yields {i32}:result1 = 4</i>
7875 %result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
7876 <i>; yields {i32}:result2 = 8</i>
7877 %result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
7878 <i>; yields {i32}:result3 = 10</i>
7879 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
7884 <!-- _______________________________________________________________________ -->
7886 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
7892 <p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7893 any integer bit width and for different address spaces. Not all targets
7894 support all bit widths however.</p>
7897 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* <ptr>, i8 <delta>)
7898 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* <ptr>, i16 <delta>)
7899 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* <ptr>, i32 <delta>)
7900 declare i64 @llvm.atomic.load.sub.i64.p0i32(i64* <ptr>, i64 <delta>)
7904 <p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
7905 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7908 <p>The intrinsic takes two arguments, the first a pointer to an integer value
7909 and the second an integer value. The result is also an integer value. These
7910 integer types can have any bit width, but they must all have the same bit
7911 width. The targets may only lower integer representations they support.</p>
7914 <p>This intrinsic does a series of operations atomically. It first loads the
7915 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7916 result to <tt>ptr</tt>. It yields the original value stored
7917 at <tt>ptr</tt>.</p>
7921 %mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7922 %ptr = bitcast i8* %mallocP to i32*
7924 %result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
7925 <i>; yields {i32}:result1 = 8</i>
7926 %result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
7927 <i>; yields {i32}:result2 = 4</i>
7928 %result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
7929 <i>; yields {i32}:result3 = 2</i>
7930 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7935 <!-- _______________________________________________________________________ -->
7937 <a name="int_atomic_load_and">
7938 '<tt>llvm.atomic.load.and.*</tt>' Intrinsic
7941 <a name="int_atomic_load_nand">
7942 '<tt>llvm.atomic.load.nand.*</tt>' Intrinsic
7945 <a name="int_atomic_load_or">
7946 '<tt>llvm.atomic.load.or.*</tt>' Intrinsic
7949 <a name="int_atomic_load_xor">
7950 '<tt>llvm.atomic.load.xor.*</tt>' Intrinsic
7957 <p>These are overloaded intrinsics. You can
7958 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7959 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7960 bit width and for different address spaces. Not all targets support all bit
7964 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* <ptr>, i8 <delta>)
7965 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* <ptr>, i16 <delta>)
7966 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* <ptr>, i32 <delta>)
7967 declare i64 @llvm.atomic.load.and.i64.p0i64(i64* <ptr>, i64 <delta>)
7971 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* <ptr>, i8 <delta>)
7972 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* <ptr>, i16 <delta>)
7973 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* <ptr>, i32 <delta>)
7974 declare i64 @llvm.atomic.load.or.i64.p0i64(i64* <ptr>, i64 <delta>)
7978 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* <ptr>, i8 <delta>)
7979 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* <ptr>, i16 <delta>)
7980 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* <ptr>, i32 <delta>)
7981 declare i64 @llvm.atomic.load.nand.i64.p0i32(i64* <ptr>, i64 <delta>)
7985 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* <ptr>, i8 <delta>)
7986 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* <ptr>, i16 <delta>)
7987 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* <ptr>, i32 <delta>)
7988 declare i64 @llvm.atomic.load.xor.i64.p0i32(i64* <ptr>, i64 <delta>)
7992 <p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7993 the value stored in memory at <tt>ptr</tt>. It yields the original value
7994 at <tt>ptr</tt>.</p>
7997 <p>These intrinsics take two arguments, the first a pointer to an integer value
7998 and the second an integer value. The result is also an integer value. These
7999 integer types can have any bit width, but they must all have the same bit
8000 width. The targets may only lower integer representations they support.</p>
8003 <p>These intrinsics does a series of operations atomically. They first load the
8004 value stored at <tt>ptr</tt>. They then do the bitwise
8005 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
8006 original value stored at <tt>ptr</tt>.</p>
8010 %mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8011 %ptr = bitcast i8* %mallocP to i32*
8012 store i32 0x0F0F, %ptr
8013 %result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
8014 <i>; yields {i32}:result0 = 0x0F0F</i>
8015 %result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
8016 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
8017 %result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
8018 <i>; yields {i32}:result2 = 0xF0</i>
8019 %result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
8020 <i>; yields {i32}:result3 = FF</i>
8021 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
8026 <!-- _______________________________________________________________________ -->
8028 <a name="int_atomic_load_max">
8029 '<tt>llvm.atomic.load.max.*</tt>' Intrinsic
8032 <a name="int_atomic_load_min">
8033 '<tt>llvm.atomic.load.min.*</tt>' Intrinsic
8036 <a name="int_atomic_load_umax">
8037 '<tt>llvm.atomic.load.umax.*</tt>' Intrinsic
8040 <a name="int_atomic_load_umin">
8041 '<tt>llvm.atomic.load.umin.*</tt>' Intrinsic
8048 <p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
8049 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
8050 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
8051 address spaces. Not all targets support all bit widths however.</p>
8054 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* <ptr>, i8 <delta>)
8055 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* <ptr>, i16 <delta>)
8056 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* <ptr>, i32 <delta>)
8057 declare i64 @llvm.atomic.load.max.i64.p0i64(i64* <ptr>, i64 <delta>)
8061 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* <ptr>, i8 <delta>)
8062 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* <ptr>, i16 <delta>)
8063 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* <ptr>, i32 <delta>)
8064 declare i64 @llvm.atomic.load.min.i64.p0i64(i64* <ptr>, i64 <delta>)
8068 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* <ptr>, i8 <delta>)
8069 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* <ptr>, i16 <delta>)
8070 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* <ptr>, i32 <delta>)
8071 declare i64 @llvm.atomic.load.umax.i64.p0i64(i64* <ptr>, i64 <delta>)
8075 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* <ptr>, i8 <delta>)
8076 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* <ptr>, i16 <delta>)
8077 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* <ptr>, i32 <delta>)
8078 declare i64 @llvm.atomic.load.umin.i64.p0i64(i64* <ptr>, i64 <delta>)
8082 <p>These intrinsics takes the signed or unsigned minimum or maximum of
8083 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
8084 original value at <tt>ptr</tt>.</p>
8087 <p>These intrinsics take two arguments, the first a pointer to an integer value
8088 and the second an integer value. The result is also an integer value. These
8089 integer types can have any bit width, but they must all have the same bit
8090 width. The targets may only lower integer representations they support.</p>
8093 <p>These intrinsics does a series of operations atomically. They first load the
8094 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
8095 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
8096 yield the original value stored at <tt>ptr</tt>.</p>
8100 %mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8101 %ptr = bitcast i8* %mallocP to i32*
8103 %result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
8104 <i>; yields {i32}:result0 = 7</i>
8105 %result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
8106 <i>; yields {i32}:result1 = -2</i>
8107 %result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
8108 <i>; yields {i32}:result2 = 8</i>
8109 %result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
8110 <i>; yields {i32}:result3 = 8</i>
8111 %memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
8118 <!-- ======================================================================= -->
8120 <a name="int_memorymarkers">Memory Use Markers</a>
8125 <p>This class of intrinsics exists to information about the lifetime of memory
8126 objects and ranges where variables are immutable.</p>
8128 <!-- _______________________________________________________________________ -->
8130 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
8137 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
8141 <p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
8142 object's lifetime.</p>
8145 <p>The first argument is a constant integer representing the size of the
8146 object, or -1 if it is variable sized. The second argument is a pointer to
8150 <p>This intrinsic indicates that before this point in the code, the value of the
8151 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
8152 never be used and has an undefined value. A load from the pointer that
8153 precedes this intrinsic can be replaced with
8154 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
8158 <!-- _______________________________________________________________________ -->
8160 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
8167 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
8171 <p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
8172 object's lifetime.</p>
8175 <p>The first argument is a constant integer representing the size of the
8176 object, or -1 if it is variable sized. The second argument is a pointer to
8180 <p>This intrinsic indicates that after this point in the code, the value of the
8181 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
8182 never be used and has an undefined value. Any stores into the memory object
8183 following this intrinsic may be removed as dead.
8187 <!-- _______________________________________________________________________ -->
8189 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
8196 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
8200 <p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
8201 a memory object will not change.</p>
8204 <p>The first argument is a constant integer representing the size of the
8205 object, or -1 if it is variable sized. The second argument is a pointer to
8209 <p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
8210 the return value, the referenced memory location is constant and
8215 <!-- _______________________________________________________________________ -->
8217 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
8224 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
8228 <p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
8229 a memory object are mutable.</p>
8232 <p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
8233 The second argument is a constant integer representing the size of the
8234 object, or -1 if it is variable sized and the third argument is a pointer
8238 <p>This intrinsic indicates that the memory is mutable again.</p>
8244 <!-- ======================================================================= -->
8246 <a name="int_general">General Intrinsics</a>
8251 <p>This class of intrinsics is designed to be generic and has no specific
8254 <!-- _______________________________________________________________________ -->
8256 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
8263 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
8267 <p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
8270 <p>The first argument is a pointer to a value, the second is a pointer to a
8271 global string, the third is a pointer to a global string which is the source
8272 file name, and the last argument is the line number.</p>
8275 <p>This intrinsic allows annotation of local variables with arbitrary strings.
8276 This can be useful for special purpose optimizations that want to look for
8277 these annotations. These have no other defined use, they are ignored by code
8278 generation and optimization.</p>
8282 <!-- _______________________________________________________________________ -->
8284 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
8290 <p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8291 any integer bit width.</p>
8294 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
8295 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
8296 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
8297 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
8298 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
8302 <p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
8305 <p>The first argument is an integer value (result of some expression), the
8306 second is a pointer to a global string, the third is a pointer to a global
8307 string which is the source file name, and the last argument is the line
8308 number. It returns the value of the first argument.</p>
8311 <p>This intrinsic allows annotations to be put on arbitrary expressions with
8312 arbitrary strings. This can be useful for special purpose optimizations that
8313 want to look for these annotations. These have no other defined use, they
8314 are ignored by code generation and optimization.</p>
8318 <!-- _______________________________________________________________________ -->
8320 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
8327 declare void @llvm.trap()
8331 <p>The '<tt>llvm.trap</tt>' intrinsic.</p>
8337 <p>This intrinsics is lowered to the target dependent trap instruction. If the
8338 target does not have a trap instruction, this intrinsic will be lowered to
8339 the call of the <tt>abort()</tt> function.</p>
8343 <!-- _______________________________________________________________________ -->
8345 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
8352 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
8356 <p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8357 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8358 ensure that it is placed on the stack before local variables.</p>
8361 <p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8362 arguments. The first argument is the value loaded from the stack
8363 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8364 that has enough space to hold the value of the guard.</p>
8367 <p>This intrinsic causes the prologue/epilogue inserter to force the position of
8368 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8369 stack. This is to ensure that if a local variable on the stack is
8370 overwritten, it will destroy the value of the guard. When the function exits,
8371 the guard on the stack is checked against the original guard. If they are
8372 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8377 <!-- _______________________________________________________________________ -->
8379 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
8386 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <type>)
8387 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <type>)
8391 <p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8392 the optimizers to determine at compile time whether a) an operation (like
8393 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8394 runtime check for overflow isn't necessary. An object in this context means
8395 an allocation of a specific class, structure, array, or other object.</p>
8398 <p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
8399 argument is a pointer to or into the <tt>object</tt>. The second argument
8400 is a boolean 0 or 1. This argument determines whether you want the
8401 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
8402 1, variables are not allowed.</p>
8405 <p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
8406 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8407 depending on the <tt>type</tt> argument, if the size cannot be determined at
8416 <!-- *********************************************************************** -->
8419 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
8420 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
8421 <a href="http://validator.w3.org/check/referer"><img
8422 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
8424 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
8425 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
8426 Last modified: $Date$