1 ==============================
2 LLVM Language Reference Manual
3 ==============================
12 This document is a reference manual for the LLVM assembly language. LLVM
13 is a Static Single Assignment (SSA) based representation that provides
14 type safety, low-level operations, flexibility, and the capability of
15 representing 'all' high-level languages cleanly. It is the common code
16 representation used throughout all phases of the LLVM compilation
22 The LLVM code representation is designed to be used in three different
23 forms: as an in-memory compiler IR, as an on-disk bitcode representation
24 (suitable for fast loading by a Just-In-Time compiler), and as a human
25 readable assembly language representation. This allows LLVM to provide a
26 powerful intermediate representation for efficient compiler
27 transformations and analysis, while providing a natural means to debug
28 and visualize the transformations. The three different forms of LLVM are
29 all equivalent. This document describes the human readable
30 representation and notation.
32 The LLVM representation aims to be light-weight and low-level while
33 being expressive, typed, and extensible at the same time. It aims to be
34 a "universal IR" of sorts, by being at a low enough level that
35 high-level ideas may be cleanly mapped to it (similar to how
36 microprocessors are "universal IR's", allowing many source languages to
37 be mapped to them). By providing type information, LLVM can be used as
38 the target of optimizations: for example, through pointer analysis, it
39 can be proven that a C automatic variable is never accessed outside of
40 the current function, allowing it to be promoted to a simple SSA value
41 instead of a memory location.
48 It is important to note that this document describes 'well formed' LLVM
49 assembly language. There is a difference between what the parser accepts
50 and what is considered 'well formed'. For example, the following
51 instruction is syntactically okay, but not well formed:
57 because the definition of ``%x`` does not dominate all of its uses. The
58 LLVM infrastructure provides a verification pass that may be used to
59 verify that an LLVM module is well formed. This pass is automatically
60 run by the parser after parsing input assembly and by the optimizer
61 before it outputs bitcode. The violations pointed out by the verifier
62 pass indicate bugs in transformation passes or input to the parser.
69 LLVM identifiers come in two basic types: global and local. Global
70 identifiers (functions, global variables) begin with the ``'@'``
71 character. Local identifiers (register names, types) begin with the
72 ``'%'`` character. Additionally, there are three different formats for
73 identifiers, for different purposes:
75 #. Named values are represented as a string of characters with their
76 prefix. For example, ``%foo``, ``@DivisionByZero``,
77 ``%a.really.long.identifier``. The actual regular expression used is
78 '``[%@][a-zA-Z$._][a-zA-Z$._0-9]*``'. Identifiers which require other
79 characters in their names can be surrounded with quotes. Special
80 characters may be escaped using ``"\xx"`` where ``xx`` is the ASCII
81 code for the character in hexadecimal. In this way, any character can
82 be used in a name value, even quotes themselves.
83 #. Unnamed values are represented as an unsigned numeric value with
84 their prefix. For example, ``%12``, ``@2``, ``%44``.
85 #. Constants, which are described in the section Constants_ below.
87 LLVM requires that values start with a prefix for two reasons: Compilers
88 don't need to worry about name clashes with reserved words, and the set
89 of reserved words may be expanded in the future without penalty.
90 Additionally, unnamed identifiers allow a compiler to quickly come up
91 with a temporary variable without having to avoid symbol table
94 Reserved words in LLVM are very similar to reserved words in other
95 languages. There are keywords for different opcodes ('``add``',
96 '``bitcast``', '``ret``', etc...), for primitive type names ('``void``',
97 '``i32``', etc...), and others. These reserved words cannot conflict
98 with variable names, because none of them start with a prefix character
101 Here is an example of LLVM code to multiply the integer variable
108 %result = mul i32 %X, 8
110 After strength reduction:
114 %result = shl i32 %X, 3
120 %0 = add i32 %X, %X ; yields {i32}:%0
121 %1 = add i32 %0, %0 ; yields {i32}:%1
122 %result = add i32 %1, %1
124 This last way of multiplying ``%X`` by 8 illustrates several important
125 lexical features of LLVM:
127 #. Comments are delimited with a '``;``' and go until the end of line.
128 #. Unnamed temporaries are created when the result of a computation is
129 not assigned to a named value.
130 #. Unnamed temporaries are numbered sequentially (using a per-function
131 incrementing counter, starting with 0). Note that basic blocks are
132 included in this numbering. For example, if the entry basic block is not
133 given a label name, then it will get number 0.
135 It also shows a convention that we follow in this document. When
136 demonstrating instructions, we will follow an instruction with a comment
137 that defines the type and name of value produced.
145 LLVM programs are composed of ``Module``'s, each of which is a
146 translation unit of the input programs. Each module consists of
147 functions, global variables, and symbol table entries. Modules may be
148 combined together with the LLVM linker, which merges function (and
149 global variable) definitions, resolves forward declarations, and merges
150 symbol table entries. Here is an example of the "hello world" module:
154 ; Declare the string constant as a global constant.
155 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
157 ; External declaration of the puts function
158 declare i32 @puts(i8* nocapture) nounwind
160 ; Definition of main function
161 define i32 @main() { ; i32()*
162 ; Convert [13 x i8]* to i8 *...
163 %cast210 = getelementptr [13 x i8]* @.str, i64 0, i64 0
165 ; Call puts function to write out the string to stdout.
166 call i32 @puts(i8* %cast210)
171 !1 = metadata !{i32 42}
174 This example is made up of a :ref:`global variable <globalvars>` named
175 "``.str``", an external declaration of the "``puts``" function, a
176 :ref:`function definition <functionstructure>` for "``main``" and
177 :ref:`named metadata <namedmetadatastructure>` "``foo``".
179 In general, a module is made up of a list of global values (where both
180 functions and global variables are global values). Global values are
181 represented by a pointer to a memory location (in this case, a pointer
182 to an array of char, and a pointer to a function), and have one of the
183 following :ref:`linkage types <linkage>`.
190 All Global Variables and Functions have one of the following types of
194 Global values with "``private``" linkage are only directly
195 accessible by objects in the current module. In particular, linking
196 code into a module with an private global value may cause the
197 private to be renamed as necessary to avoid collisions. Because the
198 symbol is private to the module, all references can be updated. This
199 doesn't show up in any symbol table in the object file.
201 Similar to ``private``, but the symbol is passed through the
202 assembler and evaluated by the linker. Unlike normal strong symbols,
203 they are removed by the linker from the final linked image
204 (executable or dynamic library).
205 ``linker_private_weak``
206 Similar to "``linker_private``", but the symbol is weak. Note that
207 ``linker_private_weak`` symbols are subject to coalescing by the
208 linker. The symbols are removed by the linker from the final linked
209 image (executable or dynamic library).
211 Similar to private, but the value shows as a local symbol
212 (``STB_LOCAL`` in the case of ELF) in the object file. This
213 corresponds to the notion of the '``static``' keyword in C.
214 ``available_externally``
215 Globals with "``available_externally``" linkage are never emitted
216 into the object file corresponding to the LLVM module. They exist to
217 allow inlining and other optimizations to take place given knowledge
218 of the definition of the global, which is known to be somewhere
219 outside the module. Globals with ``available_externally`` linkage
220 are allowed to be discarded at will, and are otherwise the same as
221 ``linkonce_odr``. This linkage type is only allowed on definitions,
224 Globals with "``linkonce``" linkage are merged with other globals of
225 the same name when linkage occurs. This can be used to implement
226 some forms of inline functions, templates, or other code which must
227 be generated in each translation unit that uses it, but where the
228 body may be overridden with a more definitive definition later.
229 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
230 that ``linkonce`` linkage does not actually allow the optimizer to
231 inline the body of this function into callers because it doesn't
232 know if this definition of the function is the definitive definition
233 within the program or whether it will be overridden by a stronger
234 definition. To enable inlining and other optimizations, use
235 "``linkonce_odr``" linkage.
237 "``weak``" linkage has the same merging semantics as ``linkonce``
238 linkage, except that unreferenced globals with ``weak`` linkage may
239 not be discarded. This is used for globals that are declared "weak"
242 "``common``" linkage is most similar to "``weak``" linkage, but they
243 are used for tentative definitions in C, such as "``int X;``" at
244 global scope. Symbols with "``common``" linkage are merged in the
245 same way as ``weak symbols``, and they may not be deleted if
246 unreferenced. ``common`` symbols may not have an explicit section,
247 must have a zero initializer, and may not be marked
248 ':ref:`constant <globalvars>`'. Functions and aliases may not have
251 .. _linkage_appending:
254 "``appending``" linkage may only be applied to global variables of
255 pointer to array type. When two global variables with appending
256 linkage are linked together, the two global arrays are appended
257 together. This is the LLVM, typesafe, equivalent of having the
258 system linker append together "sections" with identical names when
261 The semantics of this linkage follow the ELF object file model: the
262 symbol is weak until linked, if not linked, the symbol becomes null
263 instead of being an undefined reference.
264 ``linkonce_odr``, ``weak_odr``
265 Some languages allow differing globals to be merged, such as two
266 functions with different semantics. Other languages, such as
267 ``C++``, ensure that only equivalent globals are ever merged (the
268 "one definition rule" --- "ODR"). Such languages can use the
269 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
270 global will only be merged with equivalent globals. These linkage
271 types are otherwise the same as their non-``odr`` versions.
273 If none of the above identifiers are used, the global is externally
274 visible, meaning that it participates in linkage and can be used to
275 resolve external symbol references.
277 It is illegal for a function *declaration* to have any linkage type
278 other than ``external`` or ``extern_weak``.
285 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
286 :ref:`invokes <i_invoke>` can all have an optional calling convention
287 specified for the call. The calling convention of any pair of dynamic
288 caller/callee must match, or the behavior of the program is undefined.
289 The following calling conventions are supported by LLVM, and more may be
292 "``ccc``" - The C calling convention
293 This calling convention (the default if no other calling convention
294 is specified) matches the target C calling conventions. This calling
295 convention supports varargs function calls and tolerates some
296 mismatch in the declared prototype and implemented declaration of
297 the function (as does normal C).
298 "``fastcc``" - The fast calling convention
299 This calling convention attempts to make calls as fast as possible
300 (e.g. by passing things in registers). This calling convention
301 allows the target to use whatever tricks it wants to produce fast
302 code for the target, without having to conform to an externally
303 specified ABI (Application Binary Interface). `Tail calls can only
304 be optimized when this, the GHC or the HiPE convention is
305 used. <CodeGenerator.html#id80>`_ This calling convention does not
306 support varargs and requires the prototype of all callees to exactly
307 match the prototype of the function definition.
308 "``coldcc``" - The cold calling convention
309 This calling convention attempts to make code in the caller as
310 efficient as possible under the assumption that the call is not
311 commonly executed. As such, these calls often preserve all registers
312 so that the call does not break any live ranges in the caller side.
313 This calling convention does not support varargs and requires the
314 prototype of all callees to exactly match the prototype of the
315 function definition. Furthermore the inliner doesn't consider such function
317 "``cc 10``" - GHC convention
318 This calling convention has been implemented specifically for use by
319 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
320 It passes everything in registers, going to extremes to achieve this
321 by disabling callee save registers. This calling convention should
322 not be used lightly but only for specific situations such as an
323 alternative to the *register pinning* performance technique often
324 used when implementing functional programming languages. At the
325 moment only X86 supports this convention and it has the following
328 - On *X86-32* only supports up to 4 bit type parameters. No
329 floating point types are supported.
330 - On *X86-64* only supports up to 10 bit type parameters and 6
331 floating point parameters.
333 This calling convention supports `tail call
334 optimization <CodeGenerator.html#id80>`_ but requires both the
335 caller and callee are using it.
336 "``cc 11``" - The HiPE calling convention
337 This calling convention has been implemented specifically for use by
338 the `High-Performance Erlang
339 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
340 native code compiler of the `Ericsson's Open Source Erlang/OTP
341 system <http://www.erlang.org/download.shtml>`_. It uses more
342 registers for argument passing than the ordinary C calling
343 convention and defines no callee-saved registers. The calling
344 convention properly supports `tail call
345 optimization <CodeGenerator.html#id80>`_ but requires that both the
346 caller and the callee use it. It uses a *register pinning*
347 mechanism, similar to GHC's convention, for keeping frequently
348 accessed runtime components pinned to specific hardware registers.
349 At the moment only X86 supports this convention (both 32 and 64
351 "``webkit_jscc``" - WebKit's JavaScript calling convention
352 This calling convention has been implemented for `WebKit FTL JIT
353 <https://trac.webkit.org/wiki/FTLJIT>`_. It passes arguments on the
354 stack right to left (as cdecl does), and returns a value in the
355 platform's customary return register.
356 "``anyregcc``" - Dynamic calling convention for code patching
357 This is a special convention that supports patching an arbitrary code
358 sequence in place of a call site. This convention forces the call
359 arguments into registers but allows them to be dynamcially
360 allocated. This can currently only be used with calls to
361 llvm.experimental.patchpoint because only this intrinsic records
362 the location of its arguments in a side table. See :doc:`StackMaps`.
363 "``preserve_mostcc``" - The `PreserveMost` calling convention
364 This calling convention attempts to make the code in the caller as little
365 intrusive as possible. This calling convention behaves identical to the `C`
366 calling convention on how arguments and return values are passed, but it
367 uses a different set of caller/callee-saved registers. This alleviates the
368 burden of saving and recovering a large register set before and after the
369 call in the caller. If the arguments are passed in callee-saved registers,
370 then they will be preserved by the callee across the call. This doesn't
371 apply for values returned in callee-saved registers.
373 - On X86-64 the callee preserves all general purpose registers, except for
374 R11. R11 can be used as a scratch register. Floating-point registers
375 (XMMs/YMMs) are not preserved and need to be saved by the caller.
377 The idea behind this convention is to support calls to runtime functions
378 that have a hot path and a cold path. The hot path is usually a small piece
379 of code that doesn't many registers. The cold path might need to call out to
380 another function and therefore only needs to preserve the caller-saved
381 registers, which haven't already been saved by the caller. The
382 `PreserveMost` calling convention is very similar to the `cold` calling
383 convention in terms of caller/callee-saved registers, but they are used for
384 different types of function calls. `coldcc` is for function calls that are
385 rarely executed, whereas `preserve_mostcc` function calls are intended to be
386 on the hot path and definitely executed a lot. Furthermore `preserve_mostcc`
387 doesn't prevent the inliner from inlining the function call.
389 This calling convention will be used by a future version of the ObjectiveC
390 runtime and should therefore still be considered experimental at this time.
391 Although this convention was created to optimize certain runtime calls to
392 the ObjectiveC runtime, it is not limited to this runtime and might be used
393 by other runtimes in the future too. The current implementation only
394 supports X86-64, but the intention is to support more architectures in the
396 "``preserve_allcc``" - The `PreserveAll` calling convention
397 This calling convention attempts to make the code in the caller even less
398 intrusive than the `PreserveMost` calling convention. This calling
399 convention also behaves identical to the `C` calling convention on how
400 arguments and return values are passed, but it uses a different set of
401 caller/callee-saved registers. This removes the burden of saving and
402 recovering a large register set before and after the call in the caller. If
403 the arguments are passed in callee-saved registers, then they will be
404 preserved by the callee across the call. This doesn't apply for values
405 returned in callee-saved registers.
407 - On X86-64 the callee preserves all general purpose registers, except for
408 R11. R11 can be used as a scratch register. Furthermore it also preserves
409 all floating-point registers (XMMs/YMMs).
411 The idea behind this convention is to support calls to runtime functions
412 that don't need to call out to any other functions.
414 This calling convention, like the `PreserveMost` calling convention, will be
415 used by a future version of the ObjectiveC runtime and should be considered
416 experimental at this time.
417 "``cc <n>``" - Numbered convention
418 Any calling convention may be specified by number, allowing
419 target-specific calling conventions to be used. Target specific
420 calling conventions start at 64.
422 More calling conventions can be added/defined on an as-needed basis, to
423 support Pascal conventions or any other well-known target-independent
426 .. _visibilitystyles:
431 All Global Variables and Functions have one of the following visibility
434 "``default``" - Default style
435 On targets that use the ELF object file format, default visibility
436 means that the declaration is visible to other modules and, in
437 shared libraries, means that the declared entity may be overridden.
438 On Darwin, default visibility means that the declaration is visible
439 to other modules. Default visibility corresponds to "external
440 linkage" in the language.
441 "``hidden``" - Hidden style
442 Two declarations of an object with hidden visibility refer to the
443 same object if they are in the same shared object. Usually, hidden
444 visibility indicates that the symbol will not be placed into the
445 dynamic symbol table, so no other module (executable or shared
446 library) can reference it directly.
447 "``protected``" - Protected style
448 On ELF, protected visibility indicates that the symbol will be
449 placed in the dynamic symbol table, but that references within the
450 defining module will bind to the local symbol. That is, the symbol
451 cannot be overridden by another module.
458 All Global Variables, Functions and Aliases can have one of the following
462 "``dllimport``" causes the compiler to reference a function or variable via
463 a global pointer to a pointer that is set up by the DLL exporting the
464 symbol. On Microsoft Windows targets, the pointer name is formed by
465 combining ``__imp_`` and the function or variable name.
467 "``dllexport``" causes the compiler to provide a global pointer to a pointer
468 in a DLL, so that it can be referenced with the ``dllimport`` attribute. On
469 Microsoft Windows targets, the pointer name is formed by combining
470 ``__imp_`` and the function or variable name. Since this storage class
471 exists for defining a dll interface, the compiler, assembler and linker know
472 it is externally referenced and must refrain from deleting the symbol.
477 LLVM IR allows you to specify name aliases for certain types. This can
478 make it easier to read the IR and make the IR more condensed
479 (particularly when recursive types are involved). An example of a name
484 %mytype = type { %mytype*, i32 }
486 You may give a name to any :ref:`type <typesystem>` except
487 ":ref:`void <t_void>`". Type name aliases may be used anywhere a type is
488 expected with the syntax "%mytype".
490 Note that type names are aliases for the structural type that they
491 indicate, and that you can therefore specify multiple names for the same
492 type. This often leads to confusing behavior when dumping out a .ll
493 file. Since LLVM IR uses structural typing, the name is not part of the
494 type. When printing out LLVM IR, the printer will pick *one name* to
495 render all types of a particular shape. This means that if you have code
496 where two different source types end up having the same LLVM type, that
497 the dumper will sometimes print the "wrong" or unexpected type. This is
498 an important design point and isn't going to change.
505 Global variables define regions of memory allocated at compilation time
508 Global variables definitions must be initialized, may have an explicit section
509 to be placed in, and may have an optional explicit alignment specified.
511 Global variables in other translation units can also be declared, in which
512 case they don't have an initializer.
514 A variable may be defined as ``thread_local``, which means that it will
515 not be shared by threads (each thread will have a separated copy of the
516 variable). Not all targets support thread-local variables. Optionally, a
517 TLS model may be specified:
520 For variables that are only used within the current shared library.
522 For variables in modules that will not be loaded dynamically.
524 For variables defined in the executable and only used within it.
526 The models correspond to the ELF TLS models; see `ELF Handling For
527 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
528 more information on under which circumstances the different models may
529 be used. The target may choose a different TLS model if the specified
530 model is not supported, or if a better choice of model can be made.
532 A variable may be defined as a global ``constant``, which indicates that
533 the contents of the variable will **never** be modified (enabling better
534 optimization, allowing the global data to be placed in the read-only
535 section of an executable, etc). Note that variables that need runtime
536 initialization cannot be marked ``constant`` as there is a store to the
539 LLVM explicitly allows *declarations* of global variables to be marked
540 constant, even if the final definition of the global is not. This
541 capability can be used to enable slightly better optimization of the
542 program, but requires the language definition to guarantee that
543 optimizations based on the 'constantness' are valid for the translation
544 units that do not include the definition.
546 As SSA values, global variables define pointer values that are in scope
547 (i.e. they dominate) all basic blocks in the program. Global variables
548 always define a pointer to their "content" type because they describe a
549 region of memory, and all memory objects in LLVM are accessed through
552 Global variables can be marked with ``unnamed_addr`` which indicates
553 that the address is not significant, only the content. Constants marked
554 like this can be merged with other constants if they have the same
555 initializer. Note that a constant with significant address *can* be
556 merged with a ``unnamed_addr`` constant, the result being a constant
557 whose address is significant.
559 A global variable may be declared to reside in a target-specific
560 numbered address space. For targets that support them, address spaces
561 may affect how optimizations are performed and/or what target
562 instructions are used to access the variable. The default address space
563 is zero. The address space qualifier must precede any other attributes.
565 LLVM allows an explicit section to be specified for globals. If the
566 target supports it, it will emit globals to the section specified.
568 By default, global initializers are optimized by assuming that global
569 variables defined within the module are not modified from their
570 initial values before the start of the global initializer. This is
571 true even for variables potentially accessible from outside the
572 module, including those with external linkage or appearing in
573 ``@llvm.used`` or dllexported variables. This assumption may be suppressed
574 by marking the variable with ``externally_initialized``.
576 An explicit alignment may be specified for a global, which must be a
577 power of 2. If not present, or if the alignment is set to zero, the
578 alignment of the global is set by the target to whatever it feels
579 convenient. If an explicit alignment is specified, the global is forced
580 to have exactly that alignment. Targets and optimizers are not allowed
581 to over-align the global if the global has an assigned section. In this
582 case, the extra alignment could be observable: for example, code could
583 assume that the globals are densely packed in their section and try to
584 iterate over them as an array, alignment padding would break this
587 Globals can also have a :ref:`DLL storage class <dllstorageclass>`.
591 [@<GlobalVarName> =] [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal]
592 [AddrSpace] [unnamed_addr] [ExternallyInitialized]
593 <global | constant> <Type>
594 [, section "name"] [, align <Alignment>]
596 For example, the following defines a global in a numbered address space
597 with an initializer, section, and alignment:
601 @G = addrspace(5) constant float 1.0, section "foo", align 4
603 The following example just declares a global variable
607 @G = external global i32
609 The following example defines a thread-local global with the
610 ``initialexec`` TLS model:
614 @G = thread_local(initialexec) global i32 0, align 4
616 .. _functionstructure:
621 LLVM function definitions consist of the "``define``" keyword, an
622 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
623 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
624 an optional :ref:`calling convention <callingconv>`,
625 an optional ``unnamed_addr`` attribute, a return type, an optional
626 :ref:`parameter attribute <paramattrs>` for the return type, a function
627 name, a (possibly empty) argument list (each with optional :ref:`parameter
628 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
629 an optional section, an optional alignment, an optional :ref:`garbage
630 collector name <gc>`, an optional :ref:`prefix <prefixdata>`, an opening
631 curly brace, a list of basic blocks, and a closing curly brace.
633 LLVM function declarations consist of the "``declare``" keyword, an
634 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
635 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
636 an optional :ref:`calling convention <callingconv>`,
637 an optional ``unnamed_addr`` attribute, a return type, an optional
638 :ref:`parameter attribute <paramattrs>` for the return type, a function
639 name, a possibly empty list of arguments, an optional alignment, an optional
640 :ref:`garbage collector name <gc>` and an optional :ref:`prefix <prefixdata>`.
642 A function definition contains a list of basic blocks, forming the CFG (Control
643 Flow Graph) for the function. Each basic block may optionally start with a label
644 (giving the basic block a symbol table entry), contains a list of instructions,
645 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
646 function return). If an explicit label is not provided, a block is assigned an
647 implicit numbered label, using the next value from the same counter as used for
648 unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
649 entry block does not have an explicit label, it will be assigned label "%0",
650 then the first unnamed temporary in that block will be "%1", etc.
652 The first basic block in a function is special in two ways: it is
653 immediately executed on entrance to the function, and it is not allowed
654 to have predecessor basic blocks (i.e. there can not be any branches to
655 the entry block of a function). Because the block can have no
656 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
658 LLVM allows an explicit section to be specified for functions. If the
659 target supports it, it will emit functions to the section specified.
661 An explicit alignment may be specified for a function. If not present,
662 or if the alignment is set to zero, the alignment of the function is set
663 by the target to whatever it feels convenient. If an explicit alignment
664 is specified, the function is forced to have at least that much
665 alignment. All alignments must be a power of 2.
667 If the ``unnamed_addr`` attribute is given, the address is know to not
668 be significant and two identical functions can be merged.
672 define [linkage] [visibility] [DLLStorageClass]
674 <ResultType> @<FunctionName> ([argument list])
675 [fn Attrs] [section "name"] [align N]
676 [gc] [prefix Constant] { ... }
683 Aliases act as "second name" for the aliasee value (which can be either
684 function, global variable, another alias or bitcast of global value).
685 Aliases may have an optional :ref:`linkage type <linkage>`, an optional
686 :ref:`visibility style <visibility>`, and an optional :ref:`DLL storage class
691 @<Name> = [Visibility] [DLLStorageClass] alias [Linkage] <AliaseeTy> @<Aliasee>
693 The linkage must be one of ``private``, ``linker_private``,
694 ``linker_private_weak``, ``internal``, ``linkonce``, ``weak``,
695 ``linkonce_odr``, ``weak_odr``, ``external``. Note that some system linkers
696 might not correctly handle dropping a weak symbol that is aliased by a non-weak
699 .. _namedmetadatastructure:
704 Named metadata is a collection of metadata. :ref:`Metadata
705 nodes <metadata>` (but not metadata strings) are the only valid
706 operands for a named metadata.
710 ; Some unnamed metadata nodes, which are referenced by the named metadata.
711 !0 = metadata !{metadata !"zero"}
712 !1 = metadata !{metadata !"one"}
713 !2 = metadata !{metadata !"two"}
715 !name = !{!0, !1, !2}
722 The return type and each parameter of a function type may have a set of
723 *parameter attributes* associated with them. Parameter attributes are
724 used to communicate additional information about the result or
725 parameters of a function. Parameter attributes are considered to be part
726 of the function, not of the function type, so functions with different
727 parameter attributes can have the same function type.
729 Parameter attributes are simple keywords that follow the type specified.
730 If multiple parameter attributes are needed, they are space separated.
735 declare i32 @printf(i8* noalias nocapture, ...)
736 declare i32 @atoi(i8 zeroext)
737 declare signext i8 @returns_signed_char()
739 Note that any attributes for the function result (``nounwind``,
740 ``readonly``) come immediately after the argument list.
742 Currently, only the following parameter attributes are defined:
745 This indicates to the code generator that the parameter or return
746 value should be zero-extended to the extent required by the target's
747 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
748 the caller (for a parameter) or the callee (for a return value).
750 This indicates to the code generator that the parameter or return
751 value should be sign-extended to the extent required by the target's
752 ABI (which is usually 32-bits) by the caller (for a parameter) or
753 the callee (for a return value).
755 This indicates that this parameter or return value should be treated
756 in a special target-dependent fashion during while emitting code for
757 a function call or return (usually, by putting it in a register as
758 opposed to memory, though some targets use it to distinguish between
759 two different kinds of registers). Use of this attribute is
762 This indicates that the pointer parameter should really be passed by
763 value to the function. The attribute implies that a hidden copy of
764 the pointee is made between the caller and the callee, so the callee
765 is unable to modify the value in the caller. This attribute is only
766 valid on LLVM pointer arguments. It is generally used to pass
767 structs and arrays by value, but is also valid on pointers to
768 scalars. The copy is considered to belong to the caller not the
769 callee (for example, ``readonly`` functions should not write to
770 ``byval`` parameters). This is not a valid attribute for return
773 The byval attribute also supports specifying an alignment with the
774 align attribute. It indicates the alignment of the stack slot to
775 form and the known alignment of the pointer specified to the call
776 site. If the alignment is not specified, then the code generator
777 makes a target-specific assumption.
783 .. Warning:: This feature is unstable and not fully implemented.
785 The ``inalloca`` argument attribute allows the caller to take the
786 address of outgoing stack arguments. An ``inalloca`` argument must
787 be a pointer to stack memory produced by an ``alloca`` instruction.
788 The alloca, or argument allocation, must also be tagged with the
789 inalloca keyword. Only the past argument may have the ``inalloca``
790 attribute, and that argument is guaranteed to be passed in memory.
792 An argument allocation may be used by a call at most once because
793 the call may deallocate it. The ``inalloca`` attribute cannot be
794 used in conjunction with other attributes that affect argument
795 storage, like ``inreg``, ``nest``, ``sret``, or ``byval``. The
796 ``inalloca`` attribute also disables LLVM's implicit lowering of
797 large aggregate return values, which means that frontend authors
798 must lower them with ``sret`` pointers.
800 When the call site is reached, the argument allocation must have
801 been the most recent stack allocation that is still live, or the
802 results are undefined. It is possible to allocate additional stack
803 space after an argument allocation and before its call site, but it
804 must be cleared off with :ref:`llvm.stackrestore
807 See :doc:`InAlloca` for more information on how to use this
811 This indicates that the pointer parameter specifies the address of a
812 structure that is the return value of the function in the source
813 program. This pointer must be guaranteed by the caller to be valid:
814 loads and stores to the structure may be assumed by the callee
815 not to trap and to be properly aligned. This may only be applied to
816 the first parameter. This is not a valid attribute for return
819 This indicates that pointer values :ref:`based <pointeraliasing>` on
820 the argument or return value do not alias pointer values which are
821 not *based* on it, ignoring certain "irrelevant" dependencies. For a
822 call to the parent function, dependencies between memory references
823 from before or after the call and from those during the call are
824 "irrelevant" to the ``noalias`` keyword for the arguments and return
825 value used in that call. The caller shares the responsibility with
826 the callee for ensuring that these requirements are met. For further
827 details, please see the discussion of the NoAlias response in `alias
828 analysis <AliasAnalysis.html#MustMayNo>`_.
830 Note that this definition of ``noalias`` is intentionally similar
831 to the definition of ``restrict`` in C99 for function arguments,
832 though it is slightly weaker.
834 For function return values, C99's ``restrict`` is not meaningful,
835 while LLVM's ``noalias`` is.
837 This indicates that the callee does not make any copies of the
838 pointer that outlive the callee itself. This is not a valid
839 attribute for return values.
844 This indicates that the pointer parameter can be excised using the
845 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
846 attribute for return values and can only be applied to one parameter.
849 This indicates that the function always returns the argument as its return
850 value. This is an optimization hint to the code generator when generating
851 the caller, allowing tail call optimization and omission of register saves
852 and restores in some cases; it is not checked or enforced when generating
853 the callee. The parameter and the function return type must be valid
854 operands for the :ref:`bitcast instruction <i_bitcast>`. This is not a
855 valid attribute for return values and can only be applied to one parameter.
859 Garbage Collector Names
860 -----------------------
862 Each function may specify a garbage collector name, which is simply a
867 define void @f() gc "name" { ... }
869 The compiler declares the supported values of *name*. Specifying a
870 collector which will cause the compiler to alter its output in order to
871 support the named garbage collection algorithm.
878 Prefix data is data associated with a function which the code generator
879 will emit immediately before the function body. The purpose of this feature
880 is to allow frontends to associate language-specific runtime metadata with
881 specific functions and make it available through the function pointer while
882 still allowing the function pointer to be called. To access the data for a
883 given function, a program may bitcast the function pointer to a pointer to
884 the constant's type. This implies that the IR symbol points to the start
887 To maintain the semantics of ordinary function calls, the prefix data must
888 have a particular format. Specifically, it must begin with a sequence of
889 bytes which decode to a sequence of machine instructions, valid for the
890 module's target, which transfer control to the point immediately succeeding
891 the prefix data, without performing any other visible action. This allows
892 the inliner and other passes to reason about the semantics of the function
893 definition without needing to reason about the prefix data. Obviously this
894 makes the format of the prefix data highly target dependent.
896 Prefix data is laid out as if it were an initializer for a global variable
897 of the prefix data's type. No padding is automatically placed between the
898 prefix data and the function body. If padding is required, it must be part
901 A trivial example of valid prefix data for the x86 architecture is ``i8 144``,
902 which encodes the ``nop`` instruction:
906 define void @f() prefix i8 144 { ... }
908 Generally prefix data can be formed by encoding a relative branch instruction
909 which skips the metadata, as in this example of valid prefix data for the
910 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
914 %0 = type <{ i8, i8, i8* }>
916 define void @f() prefix %0 <{ i8 235, i8 8, i8* @md}> { ... }
918 A function may have prefix data but no body. This has similar semantics
919 to the ``available_externally`` linkage in that the data may be used by the
920 optimizers but will not be emitted in the object file.
927 Attribute groups are groups of attributes that are referenced by objects within
928 the IR. They are important for keeping ``.ll`` files readable, because a lot of
929 functions will use the same set of attributes. In the degenerative case of a
930 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
931 group will capture the important command line flags used to build that file.
933 An attribute group is a module-level object. To use an attribute group, an
934 object references the attribute group's ID (e.g. ``#37``). An object may refer
935 to more than one attribute group. In that situation, the attributes from the
936 different groups are merged.
938 Here is an example of attribute groups for a function that should always be
939 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
943 ; Target-independent attributes:
944 attributes #0 = { alwaysinline alignstack=4 }
946 ; Target-dependent attributes:
947 attributes #1 = { "no-sse" }
949 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
950 define void @f() #0 #1 { ... }
957 Function attributes are set to communicate additional information about
958 a function. Function attributes are considered to be part of the
959 function, not of the function type, so functions with different function
960 attributes can have the same function type.
962 Function attributes are simple keywords that follow the type specified.
963 If multiple attributes are needed, they are space separated. For
968 define void @f() noinline { ... }
969 define void @f() alwaysinline { ... }
970 define void @f() alwaysinline optsize { ... }
971 define void @f() optsize { ... }
974 This attribute indicates that, when emitting the prologue and
975 epilogue, the backend should forcibly align the stack pointer.
976 Specify the desired alignment, which must be a power of two, in
979 This attribute indicates that the inliner should attempt to inline
980 this function into callers whenever possible, ignoring any active
981 inlining size threshold for this caller.
983 This indicates that the callee function at a call site should be
984 recognized as a built-in function, even though the function's declaration
985 uses the ``nobuiltin`` attribute. This is only valid at call sites for
986 direct calls to functions which are declared with the ``nobuiltin``
989 This attribute indicates that this function is rarely called. When
990 computing edge weights, basic blocks post-dominated by a cold
991 function call are also considered to be cold; and, thus, given low
994 This attribute indicates that the source code contained a hint that
995 inlining this function is desirable (such as the "inline" keyword in
996 C/C++). It is just a hint; it imposes no requirements on the
999 This attribute suggests that optimization passes and code generator
1000 passes make choices that keep the code size of this function as small
1001 as possible and perform optimizations that may sacrifice runtime
1002 performance in order to minimize the size of the generated code.
1004 This attribute disables prologue / epilogue emission for the
1005 function. This can have very system-specific consequences.
1007 This indicates that the callee function at a call site is not recognized as
1008 a built-in function. LLVM will retain the original call and not replace it
1009 with equivalent code based on the semantics of the built-in function, unless
1010 the call site uses the ``builtin`` attribute. This is valid at call sites
1011 and on function declarations and definitions.
1013 This attribute indicates that calls to the function cannot be
1014 duplicated. A call to a ``noduplicate`` function may be moved
1015 within its parent function, but may not be duplicated within
1016 its parent function.
1018 A function containing a ``noduplicate`` call may still
1019 be an inlining candidate, provided that the call is not
1020 duplicated by inlining. That implies that the function has
1021 internal linkage and only has one call site, so the original
1022 call is dead after inlining.
1024 This attributes disables implicit floating point instructions.
1026 This attribute indicates that the inliner should never inline this
1027 function in any situation. This attribute may not be used together
1028 with the ``alwaysinline`` attribute.
1030 This attribute suppresses lazy symbol binding for the function. This
1031 may make calls to the function faster, at the cost of extra program
1032 startup time if the function is not called during program startup.
1034 This attribute indicates that the code generator should not use a
1035 red zone, even if the target-specific ABI normally permits it.
1037 This function attribute indicates that the function never returns
1038 normally. This produces undefined behavior at runtime if the
1039 function ever does dynamically return.
1041 This function attribute indicates that the function never returns
1042 with an unwind or exceptional control flow. If the function does
1043 unwind, its runtime behavior is undefined.
1045 This function attribute indicates that the function is not optimized
1046 by any optimization or code generator passes with the
1047 exception of interprocedural optimization passes.
1048 This attribute cannot be used together with the ``alwaysinline``
1049 attribute; this attribute is also incompatible
1050 with the ``minsize`` attribute and the ``optsize`` attribute.
1052 This attribute requires the ``noinline`` attribute to be specified on
1053 the function as well, so the function is never inlined into any caller.
1054 Only functions with the ``alwaysinline`` attribute are valid
1055 candidates for inlining into the body of this function.
1057 This attribute suggests that optimization passes and code generator
1058 passes make choices that keep the code size of this function low,
1059 and otherwise do optimizations specifically to reduce code size as
1060 long as they do not significantly impact runtime performance.
1062 On a function, this attribute indicates that the function computes its
1063 result (or decides to unwind an exception) based strictly on its arguments,
1064 without dereferencing any pointer arguments or otherwise accessing
1065 any mutable state (e.g. memory, control registers, etc) visible to
1066 caller functions. It does not write through any pointer arguments
1067 (including ``byval`` arguments) and never changes any state visible
1068 to callers. This means that it cannot unwind exceptions by calling
1069 the ``C++`` exception throwing methods.
1071 On an argument, this attribute indicates that the function does not
1072 dereference that pointer argument, even though it may read or write the
1073 memory that the pointer points to if accessed through other pointers.
1075 On a function, this attribute indicates that the function does not write
1076 through any pointer arguments (including ``byval`` arguments) or otherwise
1077 modify any state (e.g. memory, control registers, etc) visible to
1078 caller functions. It may dereference pointer arguments and read
1079 state that may be set in the caller. A readonly function always
1080 returns the same value (or unwinds an exception identically) when
1081 called with the same set of arguments and global state. It cannot
1082 unwind an exception by calling the ``C++`` exception throwing
1085 On an argument, this attribute indicates that the function does not write
1086 through this pointer argument, even though it may write to the memory that
1087 the pointer points to.
1089 This attribute indicates that this function can return twice. The C
1090 ``setjmp`` is an example of such a function. The compiler disables
1091 some optimizations (like tail calls) in the caller of these
1093 ``sanitize_address``
1094 This attribute indicates that AddressSanitizer checks
1095 (dynamic address safety analysis) are enabled for this function.
1097 This attribute indicates that MemorySanitizer checks (dynamic detection
1098 of accesses to uninitialized memory) are enabled for this function.
1100 This attribute indicates that ThreadSanitizer checks
1101 (dynamic thread safety analysis) are enabled for this function.
1103 This attribute indicates that the function should emit a stack
1104 smashing protector. It is in the form of a "canary" --- a random value
1105 placed on the stack before the local variables that's checked upon
1106 return from the function to see if it has been overwritten. A
1107 heuristic is used to determine if a function needs stack protectors
1108 or not. The heuristic used will enable protectors for functions with:
1110 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1111 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1112 - Calls to alloca() with variable sizes or constant sizes greater than
1113 ``ssp-buffer-size``.
1115 Variables that are identified as requiring a protector will be arranged
1116 on the stack such that they are adjacent to the stack protector guard.
1118 If a function that has an ``ssp`` attribute is inlined into a
1119 function that doesn't have an ``ssp`` attribute, then the resulting
1120 function will have an ``ssp`` attribute.
1122 This attribute indicates that the function should *always* emit a
1123 stack smashing protector. This overrides the ``ssp`` function
1126 Variables that are identified as requiring a protector will be arranged
1127 on the stack such that they are adjacent to the stack protector guard.
1128 The specific layout rules are:
1130 #. Large arrays and structures containing large arrays
1131 (``>= ssp-buffer-size``) are closest to the stack protector.
1132 #. Small arrays and structures containing small arrays
1133 (``< ssp-buffer-size``) are 2nd closest to the protector.
1134 #. Variables that have had their address taken are 3rd closest to the
1137 If a function that has an ``sspreq`` attribute is inlined into a
1138 function that doesn't have an ``sspreq`` attribute or which has an
1139 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1140 an ``sspreq`` attribute.
1142 This attribute indicates that the function should emit a stack smashing
1143 protector. This attribute causes a strong heuristic to be used when
1144 determining if a function needs stack protectors. The strong heuristic
1145 will enable protectors for functions with:
1147 - Arrays of any size and type
1148 - Aggregates containing an array of any size and type.
1149 - Calls to alloca().
1150 - Local variables that have had their address taken.
1152 Variables that are identified as requiring a protector will be arranged
1153 on the stack such that they are adjacent to the stack protector guard.
1154 The specific layout rules are:
1156 #. Large arrays and structures containing large arrays
1157 (``>= ssp-buffer-size``) are closest to the stack protector.
1158 #. Small arrays and structures containing small arrays
1159 (``< ssp-buffer-size``) are 2nd closest to the protector.
1160 #. Variables that have had their address taken are 3rd closest to the
1163 This overrides the ``ssp`` function attribute.
1165 If a function that has an ``sspstrong`` attribute is inlined into a
1166 function that doesn't have an ``sspstrong`` attribute, then the
1167 resulting function will have an ``sspstrong`` attribute.
1169 This attribute indicates that the ABI being targeted requires that
1170 an unwind table entry be produce for this function even if we can
1171 show that no exceptions passes by it. This is normally the case for
1172 the ELF x86-64 abi, but it can be disabled for some compilation
1177 Module-Level Inline Assembly
1178 ----------------------------
1180 Modules may contain "module-level inline asm" blocks, which corresponds
1181 to the GCC "file scope inline asm" blocks. These blocks are internally
1182 concatenated by LLVM and treated as a single unit, but may be separated
1183 in the ``.ll`` file if desired. The syntax is very simple:
1185 .. code-block:: llvm
1187 module asm "inline asm code goes here"
1188 module asm "more can go here"
1190 The strings can contain any character by escaping non-printable
1191 characters. The escape sequence used is simply "\\xx" where "xx" is the
1192 two digit hex code for the number.
1194 The inline asm code is simply printed to the machine code .s file when
1195 assembly code is generated.
1197 .. _langref_datalayout:
1202 A module may specify a target specific data layout string that specifies
1203 how data is to be laid out in memory. The syntax for the data layout is
1206 .. code-block:: llvm
1208 target datalayout = "layout specification"
1210 The *layout specification* consists of a list of specifications
1211 separated by the minus sign character ('-'). Each specification starts
1212 with a letter and may include other information after the letter to
1213 define some aspect of the data layout. The specifications accepted are
1217 Specifies that the target lays out data in big-endian form. That is,
1218 the bits with the most significance have the lowest address
1221 Specifies that the target lays out data in little-endian form. That
1222 is, the bits with the least significance have the lowest address
1225 Specifies the natural alignment of the stack in bits. Alignment
1226 promotion of stack variables is limited to the natural stack
1227 alignment to avoid dynamic stack realignment. The stack alignment
1228 must be a multiple of 8-bits. If omitted, the natural stack
1229 alignment defaults to "unspecified", which does not prevent any
1230 alignment promotions.
1231 ``p[n]:<size>:<abi>:<pref>``
1232 This specifies the *size* of a pointer and its ``<abi>`` and
1233 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1234 bits. The address space, ``n`` is optional, and if not specified,
1235 denotes the default address space 0. The value of ``n`` must be
1236 in the range [1,2^23).
1237 ``i<size>:<abi>:<pref>``
1238 This specifies the alignment for an integer type of a given bit
1239 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1240 ``v<size>:<abi>:<pref>``
1241 This specifies the alignment for a vector type of a given bit
1243 ``f<size>:<abi>:<pref>``
1244 This specifies the alignment for a floating point type of a given bit
1245 ``<size>``. Only values of ``<size>`` that are supported by the target
1246 will work. 32 (float) and 64 (double) are supported on all targets; 80
1247 or 128 (different flavors of long double) are also supported on some
1250 This specifies the alignment for an object of aggregate type.
1252 If present, specifies that llvm names are mangled in the output. The
1255 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1256 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1257 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1258 symbols get a ``_`` prefix.
1259 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1260 functions also get a suffix based on the frame size.
1261 ``n<size1>:<size2>:<size3>...``
1262 This specifies a set of native integer widths for the target CPU in
1263 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1264 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1265 this set are considered to support most general arithmetic operations
1268 On every specification that takes a ``<abi>:<pref>``, specifying the
1269 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1270 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1272 When constructing the data layout for a given target, LLVM starts with a
1273 default set of specifications which are then (possibly) overridden by
1274 the specifications in the ``datalayout`` keyword. The default
1275 specifications are given in this list:
1277 - ``E`` - big endian
1278 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1279 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1280 same as the default address space.
1281 - ``S0`` - natural stack alignment is unspecified
1282 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1283 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1284 - ``i16:16:16`` - i16 is 16-bit aligned
1285 - ``i32:32:32`` - i32 is 32-bit aligned
1286 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1287 alignment of 64-bits
1288 - ``f16:16:16`` - half is 16-bit aligned
1289 - ``f32:32:32`` - float is 32-bit aligned
1290 - ``f64:64:64`` - double is 64-bit aligned
1291 - ``f128:128:128`` - quad is 128-bit aligned
1292 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1293 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1294 - ``a:0:64`` - aggregates are 64-bit aligned
1296 When LLVM is determining the alignment for a given type, it uses the
1299 #. If the type sought is an exact match for one of the specifications,
1300 that specification is used.
1301 #. If no match is found, and the type sought is an integer type, then
1302 the smallest integer type that is larger than the bitwidth of the
1303 sought type is used. If none of the specifications are larger than
1304 the bitwidth then the largest integer type is used. For example,
1305 given the default specifications above, the i7 type will use the
1306 alignment of i8 (next largest) while both i65 and i256 will use the
1307 alignment of i64 (largest specified).
1308 #. If no match is found, and the type sought is a vector type, then the
1309 largest vector type that is smaller than the sought vector type will
1310 be used as a fall back. This happens because <128 x double> can be
1311 implemented in terms of 64 <2 x double>, for example.
1313 The function of the data layout string may not be what you expect.
1314 Notably, this is not a specification from the frontend of what alignment
1315 the code generator should use.
1317 Instead, if specified, the target data layout is required to match what
1318 the ultimate *code generator* expects. This string is used by the
1319 mid-level optimizers to improve code, and this only works if it matches
1320 what the ultimate code generator uses. If you would like to generate IR
1321 that does not embed this target-specific detail into the IR, then you
1322 don't have to specify the string. This will disable some optimizations
1323 that require precise layout information, but this also prevents those
1324 optimizations from introducing target specificity into the IR.
1331 A module may specify a target triple string that describes the target
1332 host. The syntax for the target triple is simply:
1334 .. code-block:: llvm
1336 target triple = "x86_64-apple-macosx10.7.0"
1338 The *target triple* string consists of a series of identifiers delimited
1339 by the minus sign character ('-'). The canonical forms are:
1343 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1344 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1346 This information is passed along to the backend so that it generates
1347 code for the proper architecture. It's possible to override this on the
1348 command line with the ``-mtriple`` command line option.
1350 .. _pointeraliasing:
1352 Pointer Aliasing Rules
1353 ----------------------
1355 Any memory access must be done through a pointer value associated with
1356 an address range of the memory access, otherwise the behavior is
1357 undefined. Pointer values are associated with address ranges according
1358 to the following rules:
1360 - A pointer value is associated with the addresses associated with any
1361 value it is *based* on.
1362 - An address of a global variable is associated with the address range
1363 of the variable's storage.
1364 - The result value of an allocation instruction is associated with the
1365 address range of the allocated storage.
1366 - A null pointer in the default address-space is associated with no
1368 - An integer constant other than zero or a pointer value returned from
1369 a function not defined within LLVM may be associated with address
1370 ranges allocated through mechanisms other than those provided by
1371 LLVM. Such ranges shall not overlap with any ranges of addresses
1372 allocated by mechanisms provided by LLVM.
1374 A pointer value is *based* on another pointer value according to the
1377 - A pointer value formed from a ``getelementptr`` operation is *based*
1378 on the first operand of the ``getelementptr``.
1379 - The result value of a ``bitcast`` is *based* on the operand of the
1381 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1382 values that contribute (directly or indirectly) to the computation of
1383 the pointer's value.
1384 - The "*based* on" relationship is transitive.
1386 Note that this definition of *"based"* is intentionally similar to the
1387 definition of *"based"* in C99, though it is slightly weaker.
1389 LLVM IR does not associate types with memory. The result type of a
1390 ``load`` merely indicates the size and alignment of the memory from
1391 which to load, as well as the interpretation of the value. The first
1392 operand type of a ``store`` similarly only indicates the size and
1393 alignment of the store.
1395 Consequently, type-based alias analysis, aka TBAA, aka
1396 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1397 :ref:`Metadata <metadata>` may be used to encode additional information
1398 which specialized optimization passes may use to implement type-based
1403 Volatile Memory Accesses
1404 ------------------------
1406 Certain memory accesses, such as :ref:`load <i_load>`'s,
1407 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1408 marked ``volatile``. The optimizers must not change the number of
1409 volatile operations or change their order of execution relative to other
1410 volatile operations. The optimizers *may* change the order of volatile
1411 operations relative to non-volatile operations. This is not Java's
1412 "volatile" and has no cross-thread synchronization behavior.
1414 IR-level volatile loads and stores cannot safely be optimized into
1415 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1416 flagged volatile. Likewise, the backend should never split or merge
1417 target-legal volatile load/store instructions.
1419 .. admonition:: Rationale
1421 Platforms may rely on volatile loads and stores of natively supported
1422 data width to be executed as single instruction. For example, in C
1423 this holds for an l-value of volatile primitive type with native
1424 hardware support, but not necessarily for aggregate types. The
1425 frontend upholds these expectations, which are intentionally
1426 unspecified in the IR. The rules above ensure that IR transformation
1427 do not violate the frontend's contract with the language.
1431 Memory Model for Concurrent Operations
1432 --------------------------------------
1434 The LLVM IR does not define any way to start parallel threads of
1435 execution or to register signal handlers. Nonetheless, there are
1436 platform-specific ways to create them, and we define LLVM IR's behavior
1437 in their presence. This model is inspired by the C++0x memory model.
1439 For a more informal introduction to this model, see the :doc:`Atomics`.
1441 We define a *happens-before* partial order as the least partial order
1444 - Is a superset of single-thread program order, and
1445 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1446 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1447 techniques, like pthread locks, thread creation, thread joining,
1448 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1449 Constraints <ordering>`).
1451 Note that program order does not introduce *happens-before* edges
1452 between a thread and signals executing inside that thread.
1454 Every (defined) read operation (load instructions, memcpy, atomic
1455 loads/read-modify-writes, etc.) R reads a series of bytes written by
1456 (defined) write operations (store instructions, atomic
1457 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1458 section, initialized globals are considered to have a write of the
1459 initializer which is atomic and happens before any other read or write
1460 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1461 may see any write to the same byte, except:
1463 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1464 write\ :sub:`2` happens before R\ :sub:`byte`, then
1465 R\ :sub:`byte` does not see write\ :sub:`1`.
1466 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1467 R\ :sub:`byte` does not see write\ :sub:`3`.
1469 Given that definition, R\ :sub:`byte` is defined as follows:
1471 - If R is volatile, the result is target-dependent. (Volatile is
1472 supposed to give guarantees which can support ``sig_atomic_t`` in
1473 C/C++, and may be used for accesses to addresses which do not behave
1474 like normal memory. It does not generally provide cross-thread
1476 - Otherwise, if there is no write to the same byte that happens before
1477 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1478 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1479 R\ :sub:`byte` returns the value written by that write.
1480 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1481 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1482 Memory Ordering Constraints <ordering>` section for additional
1483 constraints on how the choice is made.
1484 - Otherwise R\ :sub:`byte` returns ``undef``.
1486 R returns the value composed of the series of bytes it read. This
1487 implies that some bytes within the value may be ``undef`` **without**
1488 the entire value being ``undef``. Note that this only defines the
1489 semantics of the operation; it doesn't mean that targets will emit more
1490 than one instruction to read the series of bytes.
1492 Note that in cases where none of the atomic intrinsics are used, this
1493 model places only one restriction on IR transformations on top of what
1494 is required for single-threaded execution: introducing a store to a byte
1495 which might not otherwise be stored is not allowed in general.
1496 (Specifically, in the case where another thread might write to and read
1497 from an address, introducing a store can change a load that may see
1498 exactly one write into a load that may see multiple writes.)
1502 Atomic Memory Ordering Constraints
1503 ----------------------------------
1505 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1506 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1507 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1508 an ordering parameter that determines which other atomic instructions on
1509 the same address they *synchronize with*. These semantics are borrowed
1510 from Java and C++0x, but are somewhat more colloquial. If these
1511 descriptions aren't precise enough, check those specs (see spec
1512 references in the :doc:`atomics guide <Atomics>`).
1513 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1514 differently since they don't take an address. See that instruction's
1515 documentation for details.
1517 For a simpler introduction to the ordering constraints, see the
1521 The set of values that can be read is governed by the happens-before
1522 partial order. A value cannot be read unless some operation wrote
1523 it. This is intended to provide a guarantee strong enough to model
1524 Java's non-volatile shared variables. This ordering cannot be
1525 specified for read-modify-write operations; it is not strong enough
1526 to make them atomic in any interesting way.
1528 In addition to the guarantees of ``unordered``, there is a single
1529 total order for modifications by ``monotonic`` operations on each
1530 address. All modification orders must be compatible with the
1531 happens-before order. There is no guarantee that the modification
1532 orders can be combined to a global total order for the whole program
1533 (and this often will not be possible). The read in an atomic
1534 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1535 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1536 order immediately before the value it writes. If one atomic read
1537 happens before another atomic read of the same address, the later
1538 read must see the same value or a later value in the address's
1539 modification order. This disallows reordering of ``monotonic`` (or
1540 stronger) operations on the same address. If an address is written
1541 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1542 read that address repeatedly, the other threads must eventually see
1543 the write. This corresponds to the C++0x/C1x
1544 ``memory_order_relaxed``.
1546 In addition to the guarantees of ``monotonic``, a
1547 *synchronizes-with* edge may be formed with a ``release`` operation.
1548 This is intended to model C++'s ``memory_order_acquire``.
1550 In addition to the guarantees of ``monotonic``, if this operation
1551 writes a value which is subsequently read by an ``acquire``
1552 operation, it *synchronizes-with* that operation. (This isn't a
1553 complete description; see the C++0x definition of a release
1554 sequence.) This corresponds to the C++0x/C1x
1555 ``memory_order_release``.
1556 ``acq_rel`` (acquire+release)
1557 Acts as both an ``acquire`` and ``release`` operation on its
1558 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1559 ``seq_cst`` (sequentially consistent)
1560 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1561 operation which only reads, ``release`` for an operation which only
1562 writes), there is a global total order on all
1563 sequentially-consistent operations on all addresses, which is
1564 consistent with the *happens-before* partial order and with the
1565 modification orders of all the affected addresses. Each
1566 sequentially-consistent read sees the last preceding write to the
1567 same address in this global order. This corresponds to the C++0x/C1x
1568 ``memory_order_seq_cst`` and Java volatile.
1572 If an atomic operation is marked ``singlethread``, it only *synchronizes
1573 with* or participates in modification and seq\_cst total orderings with
1574 other operations running in the same thread (for example, in signal
1582 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1583 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1584 :ref:`frem <i_frem>`) have the following flags that can set to enable
1585 otherwise unsafe floating point operations
1588 No NaNs - Allow optimizations to assume the arguments and result are not
1589 NaN. Such optimizations are required to retain defined behavior over
1590 NaNs, but the value of the result is undefined.
1593 No Infs - Allow optimizations to assume the arguments and result are not
1594 +/-Inf. Such optimizations are required to retain defined behavior over
1595 +/-Inf, but the value of the result is undefined.
1598 No Signed Zeros - Allow optimizations to treat the sign of a zero
1599 argument or result as insignificant.
1602 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1603 argument rather than perform division.
1606 Fast - Allow algebraically equivalent transformations that may
1607 dramatically change results in floating point (e.g. reassociate). This
1608 flag implies all the others.
1615 The LLVM type system is one of the most important features of the
1616 intermediate representation. Being typed enables a number of
1617 optimizations to be performed on the intermediate representation
1618 directly, without having to do extra analyses on the side before the
1619 transformation. A strong type system makes it easier to read the
1620 generated code and enables novel analyses and transformations that are
1621 not feasible to perform on normal three address code representations.
1631 The void type does not represent any value and has no size.
1649 The function type can be thought of as a function signature. It consists of a
1650 return type and a list of formal parameter types. The return type of a function
1651 type is a void type or first class type --- except for :ref:`label <t_label>`
1652 and :ref:`metadata <t_metadata>` types.
1658 <returntype> (<parameter list>)
1660 ...where '``<parameter list>``' is a comma-separated list of type
1661 specifiers. Optionally, the parameter list may include a type ``...``, which
1662 indicates that the function takes a variable number of arguments. Variable
1663 argument functions can access their arguments with the :ref:`variable argument
1664 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
1665 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
1669 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1670 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1671 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1672 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1673 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1674 | ``i32 (i8*, ...)`` | A vararg function that takes at least one :ref:`pointer <t_pointer>` to ``i8`` (char in C), which returns an integer. This is the signature for ``printf`` in LLVM. |
1675 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1676 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1677 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1684 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1685 Values of these types are the only ones which can be produced by
1693 These are the types that are valid in registers from CodeGen's perspective.
1702 The integer type is a very simple type that simply specifies an
1703 arbitrary bit width for the integer type desired. Any bit width from 1
1704 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1712 The number of bits the integer will occupy is specified by the ``N``
1718 +----------------+------------------------------------------------+
1719 | ``i1`` | a single-bit integer. |
1720 +----------------+------------------------------------------------+
1721 | ``i32`` | a 32-bit integer. |
1722 +----------------+------------------------------------------------+
1723 | ``i1942652`` | a really big integer of over 1 million bits. |
1724 +----------------+------------------------------------------------+
1728 Floating Point Types
1729 """"""""""""""""""""
1738 - 16-bit floating point value
1741 - 32-bit floating point value
1744 - 64-bit floating point value
1747 - 128-bit floating point value (112-bit mantissa)
1750 - 80-bit floating point value (X87)
1753 - 128-bit floating point value (two 64-bits)
1762 The x86mmx type represents a value held in an MMX register on an x86
1763 machine. The operations allowed on it are quite limited: parameters and
1764 return values, load and store, and bitcast. User-specified MMX
1765 instructions are represented as intrinsic or asm calls with arguments
1766 and/or results of this type. There are no arrays, vectors or constants
1783 The pointer type is used to specify memory locations. Pointers are
1784 commonly used to reference objects in memory.
1786 Pointer types may have an optional address space attribute defining the
1787 numbered address space where the pointed-to object resides. The default
1788 address space is number zero. The semantics of non-zero address spaces
1789 are target-specific.
1791 Note that LLVM does not permit pointers to void (``void*``) nor does it
1792 permit pointers to labels (``label*``). Use ``i8*`` instead.
1802 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1803 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
1804 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1805 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
1806 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1807 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
1808 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1817 A vector type is a simple derived type that represents a vector of
1818 elements. Vector types are used when multiple primitive data are
1819 operated in parallel using a single instruction (SIMD). A vector type
1820 requires a size (number of elements) and an underlying primitive data
1821 type. Vector types are considered :ref:`first class <t_firstclass>`.
1827 < <# elements> x <elementtype> >
1829 The number of elements is a constant integer value larger than 0;
1830 elementtype may be any integer or floating point type, or a pointer to
1831 these types. Vectors of size zero are not allowed.
1835 +-------------------+--------------------------------------------------+
1836 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
1837 +-------------------+--------------------------------------------------+
1838 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
1839 +-------------------+--------------------------------------------------+
1840 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
1841 +-------------------+--------------------------------------------------+
1842 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
1843 +-------------------+--------------------------------------------------+
1852 The label type represents code labels.
1867 The metadata type represents embedded metadata. No derived types may be
1868 created from metadata except for :ref:`function <t_function>` arguments.
1881 Aggregate Types are a subset of derived types that can contain multiple
1882 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
1883 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
1893 The array type is a very simple derived type that arranges elements
1894 sequentially in memory. The array type requires a size (number of
1895 elements) and an underlying data type.
1901 [<# elements> x <elementtype>]
1903 The number of elements is a constant integer value; ``elementtype`` may
1904 be any type with a size.
1908 +------------------+--------------------------------------+
1909 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
1910 +------------------+--------------------------------------+
1911 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
1912 +------------------+--------------------------------------+
1913 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
1914 +------------------+--------------------------------------+
1916 Here are some examples of multidimensional arrays:
1918 +-----------------------------+----------------------------------------------------------+
1919 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
1920 +-----------------------------+----------------------------------------------------------+
1921 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
1922 +-----------------------------+----------------------------------------------------------+
1923 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
1924 +-----------------------------+----------------------------------------------------------+
1926 There is no restriction on indexing beyond the end of the array implied
1927 by a static type (though there are restrictions on indexing beyond the
1928 bounds of an allocated object in some cases). This means that
1929 single-dimension 'variable sized array' addressing can be implemented in
1930 LLVM with a zero length array type. An implementation of 'pascal style
1931 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
1941 The structure type is used to represent a collection of data members
1942 together in memory. The elements of a structure may be any type that has
1945 Structures in memory are accessed using '``load``' and '``store``' by
1946 getting a pointer to a field with the '``getelementptr``' instruction.
1947 Structures in registers are accessed using the '``extractvalue``' and
1948 '``insertvalue``' instructions.
1950 Structures may optionally be "packed" structures, which indicate that
1951 the alignment of the struct is one byte, and that there is no padding
1952 between the elements. In non-packed structs, padding between field types
1953 is inserted as defined by the DataLayout string in the module, which is
1954 required to match what the underlying code generator expects.
1956 Structures can either be "literal" or "identified". A literal structure
1957 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
1958 identified types are always defined at the top level with a name.
1959 Literal types are uniqued by their contents and can never be recursive
1960 or opaque since there is no way to write one. Identified types can be
1961 recursive, can be opaqued, and are never uniqued.
1967 %T1 = type { <type list> } ; Identified normal struct type
1968 %T2 = type <{ <type list> }> ; Identified packed struct type
1972 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1973 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
1974 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1975 | ``{ float, i32 (i32) * }`` | A pair, where the first element is a ``float`` and the second element is a :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32``, returning an ``i32``. |
1976 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1977 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
1978 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1982 Opaque Structure Types
1983 """"""""""""""""""""""
1987 Opaque structure types are used to represent named structure types that
1988 do not have a body specified. This corresponds (for example) to the C
1989 notion of a forward declared structure.
2000 +--------------+-------------------+
2001 | ``opaque`` | An opaque type. |
2002 +--------------+-------------------+
2007 LLVM has several different basic types of constants. This section
2008 describes them all and their syntax.
2013 **Boolean constants**
2014 The two strings '``true``' and '``false``' are both valid constants
2016 **Integer constants**
2017 Standard integers (such as '4') are constants of the
2018 :ref:`integer <t_integer>` type. Negative numbers may be used with
2020 **Floating point constants**
2021 Floating point constants use standard decimal notation (e.g.
2022 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2023 hexadecimal notation (see below). The assembler requires the exact
2024 decimal value of a floating-point constant. For example, the
2025 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2026 decimal in binary. Floating point constants must have a :ref:`floating
2027 point <t_floating>` type.
2028 **Null pointer constants**
2029 The identifier '``null``' is recognized as a null pointer constant
2030 and must be of :ref:`pointer type <t_pointer>`.
2032 The one non-intuitive notation for constants is the hexadecimal form of
2033 floating point constants. For example, the form
2034 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2035 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2036 constants are required (and the only time that they are generated by the
2037 disassembler) is when a floating point constant must be emitted but it
2038 cannot be represented as a decimal floating point number in a reasonable
2039 number of digits. For example, NaN's, infinities, and other special
2040 values are represented in their IEEE hexadecimal format so that assembly
2041 and disassembly do not cause any bits to change in the constants.
2043 When using the hexadecimal form, constants of types half, float, and
2044 double are represented using the 16-digit form shown above (which
2045 matches the IEEE754 representation for double); half and float values
2046 must, however, be exactly representable as IEEE 754 half and single
2047 precision, respectively. Hexadecimal format is always used for long
2048 double, and there are three forms of long double. The 80-bit format used
2049 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2050 128-bit format used by PowerPC (two adjacent doubles) is represented by
2051 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2052 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2053 will only work if they match the long double format on your target.
2054 The IEEE 16-bit format (half precision) is represented by ``0xH``
2055 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2056 (sign bit at the left).
2058 There are no constants of type x86mmx.
2060 .. _complexconstants:
2065 Complex constants are a (potentially recursive) combination of simple
2066 constants and smaller complex constants.
2068 **Structure constants**
2069 Structure constants are represented with notation similar to
2070 structure type definitions (a comma separated list of elements,
2071 surrounded by braces (``{}``)). For example:
2072 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2073 "``@G = external global i32``". Structure constants must have
2074 :ref:`structure type <t_struct>`, and the number and types of elements
2075 must match those specified by the type.
2077 Array constants are represented with notation similar to array type
2078 definitions (a comma separated list of elements, surrounded by
2079 square brackets (``[]``)). For example:
2080 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2081 :ref:`array type <t_array>`, and the number and types of elements must
2082 match those specified by the type.
2083 **Vector constants**
2084 Vector constants are represented with notation similar to vector
2085 type definitions (a comma separated list of elements, surrounded by
2086 less-than/greater-than's (``<>``)). For example:
2087 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2088 must have :ref:`vector type <t_vector>`, and the number and types of
2089 elements must match those specified by the type.
2090 **Zero initialization**
2091 The string '``zeroinitializer``' can be used to zero initialize a
2092 value to zero of *any* type, including scalar and
2093 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2094 having to print large zero initializers (e.g. for large arrays) and
2095 is always exactly equivalent to using explicit zero initializers.
2097 A metadata node is a structure-like constant with :ref:`metadata
2098 type <t_metadata>`. For example:
2099 "``metadata !{ i32 0, metadata !"test" }``". Unlike other
2100 constants that are meant to be interpreted as part of the
2101 instruction stream, metadata is a place to attach additional
2102 information such as debug info.
2104 Global Variable and Function Addresses
2105 --------------------------------------
2107 The addresses of :ref:`global variables <globalvars>` and
2108 :ref:`functions <functionstructure>` are always implicitly valid
2109 (link-time) constants. These constants are explicitly referenced when
2110 the :ref:`identifier for the global <identifiers>` is used and always have
2111 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2114 .. code-block:: llvm
2118 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2125 The string '``undef``' can be used anywhere a constant is expected, and
2126 indicates that the user of the value may receive an unspecified
2127 bit-pattern. Undefined values may be of any type (other than '``label``'
2128 or '``void``') and be used anywhere a constant is permitted.
2130 Undefined values are useful because they indicate to the compiler that
2131 the program is well defined no matter what value is used. This gives the
2132 compiler more freedom to optimize. Here are some examples of
2133 (potentially surprising) transformations that are valid (in pseudo IR):
2135 .. code-block:: llvm
2145 This is safe because all of the output bits are affected by the undef
2146 bits. Any output bit can have a zero or one depending on the input bits.
2148 .. code-block:: llvm
2159 These logical operations have bits that are not always affected by the
2160 input. For example, if ``%X`` has a zero bit, then the output of the
2161 '``and``' operation will always be a zero for that bit, no matter what
2162 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2163 optimize or assume that the result of the '``and``' is '``undef``'.
2164 However, it is safe to assume that all bits of the '``undef``' could be
2165 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2166 all the bits of the '``undef``' operand to the '``or``' could be set,
2167 allowing the '``or``' to be folded to -1.
2169 .. code-block:: llvm
2171 %A = select undef, %X, %Y
2172 %B = select undef, 42, %Y
2173 %C = select %X, %Y, undef
2183 This set of examples shows that undefined '``select``' (and conditional
2184 branch) conditions can go *either way*, but they have to come from one
2185 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2186 both known to have a clear low bit, then ``%A`` would have to have a
2187 cleared low bit. However, in the ``%C`` example, the optimizer is
2188 allowed to assume that the '``undef``' operand could be the same as
2189 ``%Y``, allowing the whole '``select``' to be eliminated.
2191 .. code-block:: llvm
2193 %A = xor undef, undef
2210 This example points out that two '``undef``' operands are not
2211 necessarily the same. This can be surprising to people (and also matches
2212 C semantics) where they assume that "``X^X``" is always zero, even if
2213 ``X`` is undefined. This isn't true for a number of reasons, but the
2214 short answer is that an '``undef``' "variable" can arbitrarily change
2215 its value over its "live range". This is true because the variable
2216 doesn't actually *have a live range*. Instead, the value is logically
2217 read from arbitrary registers that happen to be around when needed, so
2218 the value is not necessarily consistent over time. In fact, ``%A`` and
2219 ``%C`` need to have the same semantics or the core LLVM "replace all
2220 uses with" concept would not hold.
2222 .. code-block:: llvm
2230 These examples show the crucial difference between an *undefined value*
2231 and *undefined behavior*. An undefined value (like '``undef``') is
2232 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2233 operation can be constant folded to '``undef``', because the '``undef``'
2234 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2235 However, in the second example, we can make a more aggressive
2236 assumption: because the ``undef`` is allowed to be an arbitrary value,
2237 we are allowed to assume that it could be zero. Since a divide by zero
2238 has *undefined behavior*, we are allowed to assume that the operation
2239 does not execute at all. This allows us to delete the divide and all
2240 code after it. Because the undefined operation "can't happen", the
2241 optimizer can assume that it occurs in dead code.
2243 .. code-block:: llvm
2245 a: store undef -> %X
2246 b: store %X -> undef
2251 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2252 value can be assumed to not have any effect; we can assume that the
2253 value is overwritten with bits that happen to match what was already
2254 there. However, a store *to* an undefined location could clobber
2255 arbitrary memory, therefore, it has undefined behavior.
2262 Poison values are similar to :ref:`undef values <undefvalues>`, however
2263 they also represent the fact that an instruction or constant expression
2264 which cannot evoke side effects has nevertheless detected a condition
2265 which results in undefined behavior.
2267 There is currently no way of representing a poison value in the IR; they
2268 only exist when produced by operations such as :ref:`add <i_add>` with
2271 Poison value behavior is defined in terms of value *dependence*:
2273 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2274 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2275 their dynamic predecessor basic block.
2276 - Function arguments depend on the corresponding actual argument values
2277 in the dynamic callers of their functions.
2278 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2279 instructions that dynamically transfer control back to them.
2280 - :ref:`Invoke <i_invoke>` instructions depend on the
2281 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2282 call instructions that dynamically transfer control back to them.
2283 - Non-volatile loads and stores depend on the most recent stores to all
2284 of the referenced memory addresses, following the order in the IR
2285 (including loads and stores implied by intrinsics such as
2286 :ref:`@llvm.memcpy <int_memcpy>`.)
2287 - An instruction with externally visible side effects depends on the
2288 most recent preceding instruction with externally visible side
2289 effects, following the order in the IR. (This includes :ref:`volatile
2290 operations <volatile>`.)
2291 - An instruction *control-depends* on a :ref:`terminator
2292 instruction <terminators>` if the terminator instruction has
2293 multiple successors and the instruction is always executed when
2294 control transfers to one of the successors, and may not be executed
2295 when control is transferred to another.
2296 - Additionally, an instruction also *control-depends* on a terminator
2297 instruction if the set of instructions it otherwise depends on would
2298 be different if the terminator had transferred control to a different
2300 - Dependence is transitive.
2302 Poison Values have the same behavior as :ref:`undef values <undefvalues>`,
2303 with the additional affect that any instruction which has a *dependence*
2304 on a poison value has undefined behavior.
2306 Here are some examples:
2308 .. code-block:: llvm
2311 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2312 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2313 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
2314 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2316 store i32 %poison, i32* @g ; Poison value stored to memory.
2317 %poison2 = load i32* @g ; Poison value loaded back from memory.
2319 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2321 %narrowaddr = bitcast i32* @g to i16*
2322 %wideaddr = bitcast i32* @g to i64*
2323 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2324 %poison4 = load i64* %wideaddr ; Returns a poison value.
2326 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2327 br i1 %cmp, label %true, label %end ; Branch to either destination.
2330 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2331 ; it has undefined behavior.
2335 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2336 ; Both edges into this PHI are
2337 ; control-dependent on %cmp, so this
2338 ; always results in a poison value.
2340 store volatile i32 0, i32* @g ; This would depend on the store in %true
2341 ; if %cmp is true, or the store in %entry
2342 ; otherwise, so this is undefined behavior.
2344 br i1 %cmp, label %second_true, label %second_end
2345 ; The same branch again, but this time the
2346 ; true block doesn't have side effects.
2353 store volatile i32 0, i32* @g ; This time, the instruction always depends
2354 ; on the store in %end. Also, it is
2355 ; control-equivalent to %end, so this is
2356 ; well-defined (ignoring earlier undefined
2357 ; behavior in this example).
2361 Addresses of Basic Blocks
2362 -------------------------
2364 ``blockaddress(@function, %block)``
2366 The '``blockaddress``' constant computes the address of the specified
2367 basic block in the specified function, and always has an ``i8*`` type.
2368 Taking the address of the entry block is illegal.
2370 This value only has defined behavior when used as an operand to the
2371 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2372 against null. Pointer equality tests between labels addresses results in
2373 undefined behavior --- though, again, comparison against null is ok, and
2374 no label is equal to the null pointer. This may be passed around as an
2375 opaque pointer sized value as long as the bits are not inspected. This
2376 allows ``ptrtoint`` and arithmetic to be performed on these values so
2377 long as the original value is reconstituted before the ``indirectbr``
2380 Finally, some targets may provide defined semantics when using the value
2381 as the operand to an inline assembly, but that is target specific.
2385 Constant Expressions
2386 --------------------
2388 Constant expressions are used to allow expressions involving other
2389 constants to be used as constants. Constant expressions may be of any
2390 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2391 that does not have side effects (e.g. load and call are not supported).
2392 The following is the syntax for constant expressions:
2394 ``trunc (CST to TYPE)``
2395 Truncate a constant to another type. The bit size of CST must be
2396 larger than the bit size of TYPE. Both types must be integers.
2397 ``zext (CST to TYPE)``
2398 Zero extend a constant to another type. The bit size of CST must be
2399 smaller than the bit size of TYPE. Both types must be integers.
2400 ``sext (CST to TYPE)``
2401 Sign extend a constant to another type. The bit size of CST must be
2402 smaller than the bit size of TYPE. Both types must be integers.
2403 ``fptrunc (CST to TYPE)``
2404 Truncate a floating point constant to another floating point type.
2405 The size of CST must be larger than the size of TYPE. Both types
2406 must be floating point.
2407 ``fpext (CST to TYPE)``
2408 Floating point extend a constant to another type. The size of CST
2409 must be smaller or equal to the size of TYPE. Both types must be
2411 ``fptoui (CST to TYPE)``
2412 Convert a floating point constant to the corresponding unsigned
2413 integer constant. TYPE must be a scalar or vector integer type. CST
2414 must be of scalar or vector floating point type. Both CST and TYPE
2415 must be scalars, or vectors of the same number of elements. If the
2416 value won't fit in the integer type, the results are undefined.
2417 ``fptosi (CST to TYPE)``
2418 Convert a floating point constant to the corresponding signed
2419 integer constant. TYPE must be a scalar or vector integer type. CST
2420 must be of scalar or vector floating point type. Both CST and TYPE
2421 must be scalars, or vectors of the same number of elements. If the
2422 value won't fit in the integer type, the results are undefined.
2423 ``uitofp (CST to TYPE)``
2424 Convert an unsigned integer constant to the corresponding floating
2425 point constant. TYPE must be a scalar or vector floating point type.
2426 CST must be of scalar or vector integer type. Both CST and TYPE must
2427 be scalars, or vectors of the same number of elements. If the value
2428 won't fit in the floating point type, the results are undefined.
2429 ``sitofp (CST to TYPE)``
2430 Convert a signed integer constant to the corresponding floating
2431 point constant. TYPE must be a scalar or vector floating point type.
2432 CST must be of scalar or vector integer type. Both CST and TYPE must
2433 be scalars, or vectors of the same number of elements. If the value
2434 won't fit in the floating point type, the results are undefined.
2435 ``ptrtoint (CST to TYPE)``
2436 Convert a pointer typed constant to the corresponding integer
2437 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2438 pointer type. The ``CST`` value is zero extended, truncated, or
2439 unchanged to make it fit in ``TYPE``.
2440 ``inttoptr (CST to TYPE)``
2441 Convert an integer constant to a pointer constant. TYPE must be a
2442 pointer type. CST must be of integer type. The CST value is zero
2443 extended, truncated, or unchanged to make it fit in a pointer size.
2444 This one is *really* dangerous!
2445 ``bitcast (CST to TYPE)``
2446 Convert a constant, CST, to another TYPE. The constraints of the
2447 operands are the same as those for the :ref:`bitcast
2448 instruction <i_bitcast>`.
2449 ``addrspacecast (CST to TYPE)``
2450 Convert a constant pointer or constant vector of pointer, CST, to another
2451 TYPE in a different address space. The constraints of the operands are the
2452 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2453 ``getelementptr (CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)``
2454 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2455 constants. As with the :ref:`getelementptr <i_getelementptr>`
2456 instruction, the index list may have zero or more indexes, which are
2457 required to make sense for the type of "CSTPTR".
2458 ``select (COND, VAL1, VAL2)``
2459 Perform the :ref:`select operation <i_select>` on constants.
2460 ``icmp COND (VAL1, VAL2)``
2461 Performs the :ref:`icmp operation <i_icmp>` on constants.
2462 ``fcmp COND (VAL1, VAL2)``
2463 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2464 ``extractelement (VAL, IDX)``
2465 Perform the :ref:`extractelement operation <i_extractelement>` on
2467 ``insertelement (VAL, ELT, IDX)``
2468 Perform the :ref:`insertelement operation <i_insertelement>` on
2470 ``shufflevector (VEC1, VEC2, IDXMASK)``
2471 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2473 ``extractvalue (VAL, IDX0, IDX1, ...)``
2474 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2475 constants. The index list is interpreted in a similar manner as
2476 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2477 least one index value must be specified.
2478 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2479 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2480 The index list is interpreted in a similar manner as indices in a
2481 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2482 value must be specified.
2483 ``OPCODE (LHS, RHS)``
2484 Perform the specified operation of the LHS and RHS constants. OPCODE
2485 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2486 binary <bitwiseops>` operations. The constraints on operands are
2487 the same as those for the corresponding instruction (e.g. no bitwise
2488 operations on floating point values are allowed).
2495 Inline Assembler Expressions
2496 ----------------------------
2498 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2499 Inline Assembly <moduleasm>`) through the use of a special value. This
2500 value represents the inline assembler as a string (containing the
2501 instructions to emit), a list of operand constraints (stored as a
2502 string), a flag that indicates whether or not the inline asm expression
2503 has side effects, and a flag indicating whether the function containing
2504 the asm needs to align its stack conservatively. An example inline
2505 assembler expression is:
2507 .. code-block:: llvm
2509 i32 (i32) asm "bswap $0", "=r,r"
2511 Inline assembler expressions may **only** be used as the callee operand
2512 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2513 Thus, typically we have:
2515 .. code-block:: llvm
2517 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2519 Inline asms with side effects not visible in the constraint list must be
2520 marked as having side effects. This is done through the use of the
2521 '``sideeffect``' keyword, like so:
2523 .. code-block:: llvm
2525 call void asm sideeffect "eieio", ""()
2527 In some cases inline asms will contain code that will not work unless
2528 the stack is aligned in some way, such as calls or SSE instructions on
2529 x86, yet will not contain code that does that alignment within the asm.
2530 The compiler should make conservative assumptions about what the asm
2531 might contain and should generate its usual stack alignment code in the
2532 prologue if the '``alignstack``' keyword is present:
2534 .. code-block:: llvm
2536 call void asm alignstack "eieio", ""()
2538 Inline asms also support using non-standard assembly dialects. The
2539 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2540 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2541 the only supported dialects. An example is:
2543 .. code-block:: llvm
2545 call void asm inteldialect "eieio", ""()
2547 If multiple keywords appear the '``sideeffect``' keyword must come
2548 first, the '``alignstack``' keyword second and the '``inteldialect``'
2554 The call instructions that wrap inline asm nodes may have a
2555 "``!srcloc``" MDNode attached to it that contains a list of constant
2556 integers. If present, the code generator will use the integer as the
2557 location cookie value when report errors through the ``LLVMContext``
2558 error reporting mechanisms. This allows a front-end to correlate backend
2559 errors that occur with inline asm back to the source code that produced
2562 .. code-block:: llvm
2564 call void asm sideeffect "something bad", ""(), !srcloc !42
2566 !42 = !{ i32 1234567 }
2568 It is up to the front-end to make sense of the magic numbers it places
2569 in the IR. If the MDNode contains multiple constants, the code generator
2570 will use the one that corresponds to the line of the asm that the error
2575 Metadata Nodes and Metadata Strings
2576 -----------------------------------
2578 LLVM IR allows metadata to be attached to instructions in the program
2579 that can convey extra information about the code to the optimizers and
2580 code generator. One example application of metadata is source-level
2581 debug information. There are two metadata primitives: strings and nodes.
2582 All metadata has the ``metadata`` type and is identified in syntax by a
2583 preceding exclamation point ('``!``').
2585 A metadata string is a string surrounded by double quotes. It can
2586 contain any character by escaping non-printable characters with
2587 "``\xx``" where "``xx``" is the two digit hex code. For example:
2590 Metadata nodes are represented with notation similar to structure
2591 constants (a comma separated list of elements, surrounded by braces and
2592 preceded by an exclamation point). Metadata nodes can have any values as
2593 their operand. For example:
2595 .. code-block:: llvm
2597 !{ metadata !"test\00", i32 10}
2599 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2600 metadata nodes, which can be looked up in the module symbol table. For
2603 .. code-block:: llvm
2605 !foo = metadata !{!4, !3}
2607 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2608 function is using two metadata arguments:
2610 .. code-block:: llvm
2612 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2614 Metadata can be attached with an instruction. Here metadata ``!21`` is
2615 attached to the ``add`` instruction using the ``!dbg`` identifier:
2617 .. code-block:: llvm
2619 %indvar.next = add i64 %indvar, 1, !dbg !21
2621 More information about specific metadata nodes recognized by the
2622 optimizers and code generator is found below.
2627 In LLVM IR, memory does not have types, so LLVM's own type system is not
2628 suitable for doing TBAA. Instead, metadata is added to the IR to
2629 describe a type system of a higher level language. This can be used to
2630 implement typical C/C++ TBAA, but it can also be used to implement
2631 custom alias analysis behavior for other languages.
2633 The current metadata format is very simple. TBAA metadata nodes have up
2634 to three fields, e.g.:
2636 .. code-block:: llvm
2638 !0 = metadata !{ metadata !"an example type tree" }
2639 !1 = metadata !{ metadata !"int", metadata !0 }
2640 !2 = metadata !{ metadata !"float", metadata !0 }
2641 !3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2643 The first field is an identity field. It can be any value, usually a
2644 metadata string, which uniquely identifies the type. The most important
2645 name in the tree is the name of the root node. Two trees with different
2646 root node names are entirely disjoint, even if they have leaves with
2649 The second field identifies the type's parent node in the tree, or is
2650 null or omitted for a root node. A type is considered to alias all of
2651 its descendants and all of its ancestors in the tree. Also, a type is
2652 considered to alias all types in other trees, so that bitcode produced
2653 from multiple front-ends is handled conservatively.
2655 If the third field is present, it's an integer which if equal to 1
2656 indicates that the type is "constant" (meaning
2657 ``pointsToConstantMemory`` should return true; see `other useful
2658 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
2660 '``tbaa.struct``' Metadata
2661 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2663 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
2664 aggregate assignment operations in C and similar languages, however it
2665 is defined to copy a contiguous region of memory, which is more than
2666 strictly necessary for aggregate types which contain holes due to
2667 padding. Also, it doesn't contain any TBAA information about the fields
2670 ``!tbaa.struct`` metadata can describe which memory subregions in a
2671 memcpy are padding and what the TBAA tags of the struct are.
2673 The current metadata format is very simple. ``!tbaa.struct`` metadata
2674 nodes are a list of operands which are in conceptual groups of three.
2675 For each group of three, the first operand gives the byte offset of a
2676 field in bytes, the second gives its size in bytes, and the third gives
2679 .. code-block:: llvm
2681 !4 = metadata !{ i64 0, i64 4, metadata !1, i64 8, i64 4, metadata !2 }
2683 This describes a struct with two fields. The first is at offset 0 bytes
2684 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
2685 and has size 4 bytes and has tbaa tag !2.
2687 Note that the fields need not be contiguous. In this example, there is a
2688 4 byte gap between the two fields. This gap represents padding which
2689 does not carry useful data and need not be preserved.
2691 '``fpmath``' Metadata
2692 ^^^^^^^^^^^^^^^^^^^^^
2694 ``fpmath`` metadata may be attached to any instruction of floating point
2695 type. It can be used to express the maximum acceptable error in the
2696 result of that instruction, in ULPs, thus potentially allowing the
2697 compiler to use a more efficient but less accurate method of computing
2698 it. ULP is defined as follows:
2700 If ``x`` is a real number that lies between two finite consecutive
2701 floating-point numbers ``a`` and ``b``, without being equal to one
2702 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
2703 distance between the two non-equal finite floating-point numbers
2704 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
2706 The metadata node shall consist of a single positive floating point
2707 number representing the maximum relative error, for example:
2709 .. code-block:: llvm
2711 !0 = metadata !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
2713 '``range``' Metadata
2714 ^^^^^^^^^^^^^^^^^^^^
2716 ``range`` metadata may be attached only to loads of integer types. It
2717 expresses the possible ranges the loaded value is in. The ranges are
2718 represented with a flattened list of integers. The loaded value is known
2719 to be in the union of the ranges defined by each consecutive pair. Each
2720 pair has the following properties:
2722 - The type must match the type loaded by the instruction.
2723 - The pair ``a,b`` represents the range ``[a,b)``.
2724 - Both ``a`` and ``b`` are constants.
2725 - The range is allowed to wrap.
2726 - The range should not represent the full or empty set. That is,
2729 In addition, the pairs must be in signed order of the lower bound and
2730 they must be non-contiguous.
2734 .. code-block:: llvm
2736 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
2737 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
2738 %c = load i8* %z, align 1, !range !2 ; Can only be 0, 1, 3, 4 or 5
2739 %d = load i8* %z, align 1, !range !3 ; Can only be -2, -1, 3, 4 or 5
2741 !0 = metadata !{ i8 0, i8 2 }
2742 !1 = metadata !{ i8 255, i8 2 }
2743 !2 = metadata !{ i8 0, i8 2, i8 3, i8 6 }
2744 !3 = metadata !{ i8 -2, i8 0, i8 3, i8 6 }
2749 It is sometimes useful to attach information to loop constructs. Currently,
2750 loop metadata is implemented as metadata attached to the branch instruction
2751 in the loop latch block. This type of metadata refer to a metadata node that is
2752 guaranteed to be separate for each loop. The loop identifier metadata is
2753 specified with the name ``llvm.loop``.
2755 The loop identifier metadata is implemented using a metadata that refers to
2756 itself to avoid merging it with any other identifier metadata, e.g.,
2757 during module linkage or function inlining. That is, each loop should refer
2758 to their own identification metadata even if they reside in separate functions.
2759 The following example contains loop identifier metadata for two separate loop
2762 .. code-block:: llvm
2764 !0 = metadata !{ metadata !0 }
2765 !1 = metadata !{ metadata !1 }
2767 The loop identifier metadata can be used to specify additional per-loop
2768 metadata. Any operands after the first operand can be treated as user-defined
2769 metadata. For example the ``llvm.vectorizer.unroll`` metadata is understood
2770 by the loop vectorizer to indicate how many times to unroll the loop:
2772 .. code-block:: llvm
2774 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
2776 !0 = metadata !{ metadata !0, metadata !1 }
2777 !1 = metadata !{ metadata !"llvm.vectorizer.unroll", i32 2 }
2782 Metadata types used to annotate memory accesses with information helpful
2783 for optimizations are prefixed with ``llvm.mem``.
2785 '``llvm.mem.parallel_loop_access``' Metadata
2786 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2788 For a loop to be parallel, in addition to using
2789 the ``llvm.loop`` metadata to mark the loop latch branch instruction,
2790 also all of the memory accessing instructions in the loop body need to be
2791 marked with the ``llvm.mem.parallel_loop_access`` metadata. If there
2792 is at least one memory accessing instruction not marked with the metadata,
2793 the loop must be considered a sequential loop. This causes parallel loops to be
2794 converted to sequential loops due to optimization passes that are unaware of
2795 the parallel semantics and that insert new memory instructions to the loop
2798 Example of a loop that is considered parallel due to its correct use of
2799 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
2800 metadata types that refer to the same loop identifier metadata.
2802 .. code-block:: llvm
2806 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2808 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2810 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
2814 !0 = metadata !{ metadata !0 }
2816 It is also possible to have nested parallel loops. In that case the
2817 memory accesses refer to a list of loop identifier metadata nodes instead of
2818 the loop identifier metadata node directly:
2820 .. code-block:: llvm
2827 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2829 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2831 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
2835 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2837 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2839 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
2841 outer.for.end: ; preds = %for.body
2843 !0 = metadata !{ metadata !1, metadata !2 } ; a list of loop identifiers
2844 !1 = metadata !{ metadata !1 } ; an identifier for the inner loop
2845 !2 = metadata !{ metadata !2 } ; an identifier for the outer loop
2847 '``llvm.vectorizer``'
2848 ^^^^^^^^^^^^^^^^^^^^^
2850 Metadata prefixed with ``llvm.vectorizer`` is used to control per-loop
2851 vectorization parameters such as vectorization factor and unroll factor.
2853 ``llvm.vectorizer`` metadata should be used in conjunction with ``llvm.loop``
2854 loop identification metadata.
2856 '``llvm.vectorizer.unroll``' Metadata
2857 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2859 This metadata instructs the loop vectorizer to unroll the specified
2860 loop exactly ``N`` times.
2862 The first operand is the string ``llvm.vectorizer.unroll`` and the second
2863 operand is an integer specifying the unroll factor. For example:
2865 .. code-block:: llvm
2867 !0 = metadata !{ metadata !"llvm.vectorizer.unroll", i32 4 }
2869 Note that setting ``llvm.vectorizer.unroll`` to 1 disables unrolling of the
2872 If ``llvm.vectorizer.unroll`` is set to 0 then the amount of unrolling will be
2873 determined automatically.
2875 '``llvm.vectorizer.width``' Metadata
2876 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2878 This metadata sets the target width of the vectorizer to ``N``. Without
2879 this metadata, the vectorizer will choose a width automatically.
2880 Regardless of this metadata, the vectorizer will only vectorize loops if
2881 it believes it is valid to do so.
2883 The first operand is the string ``llvm.vectorizer.width`` and the second
2884 operand is an integer specifying the width. For example:
2886 .. code-block:: llvm
2888 !0 = metadata !{ metadata !"llvm.vectorizer.width", i32 4 }
2890 Note that setting ``llvm.vectorizer.width`` to 1 disables vectorization of the
2893 If ``llvm.vectorizer.width`` is set to 0 then the width will be determined
2896 Module Flags Metadata
2897 =====================
2899 Information about the module as a whole is difficult to convey to LLVM's
2900 subsystems. The LLVM IR isn't sufficient to transmit this information.
2901 The ``llvm.module.flags`` named metadata exists in order to facilitate
2902 this. These flags are in the form of key / value pairs --- much like a
2903 dictionary --- making it easy for any subsystem who cares about a flag to
2906 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
2907 Each triplet has the following form:
2909 - The first element is a *behavior* flag, which specifies the behavior
2910 when two (or more) modules are merged together, and it encounters two
2911 (or more) metadata with the same ID. The supported behaviors are
2913 - The second element is a metadata string that is a unique ID for the
2914 metadata. Each module may only have one flag entry for each unique ID (not
2915 including entries with the **Require** behavior).
2916 - The third element is the value of the flag.
2918 When two (or more) modules are merged together, the resulting
2919 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
2920 each unique metadata ID string, there will be exactly one entry in the merged
2921 modules ``llvm.module.flags`` metadata table, and the value for that entry will
2922 be determined by the merge behavior flag, as described below. The only exception
2923 is that entries with the *Require* behavior are always preserved.
2925 The following behaviors are supported:
2936 Emits an error if two values disagree, otherwise the resulting value
2937 is that of the operands.
2941 Emits a warning if two values disagree. The result value will be the
2942 operand for the flag from the first module being linked.
2946 Adds a requirement that another module flag be present and have a
2947 specified value after linking is performed. The value must be a
2948 metadata pair, where the first element of the pair is the ID of the
2949 module flag to be restricted, and the second element of the pair is
2950 the value the module flag should be restricted to. This behavior can
2951 be used to restrict the allowable results (via triggering of an
2952 error) of linking IDs with the **Override** behavior.
2956 Uses the specified value, regardless of the behavior or value of the
2957 other module. If both modules specify **Override**, but the values
2958 differ, an error will be emitted.
2962 Appends the two values, which are required to be metadata nodes.
2966 Appends the two values, which are required to be metadata
2967 nodes. However, duplicate entries in the second list are dropped
2968 during the append operation.
2970 It is an error for a particular unique flag ID to have multiple behaviors,
2971 except in the case of **Require** (which adds restrictions on another metadata
2972 value) or **Override**.
2974 An example of module flags:
2976 .. code-block:: llvm
2978 !0 = metadata !{ i32 1, metadata !"foo", i32 1 }
2979 !1 = metadata !{ i32 4, metadata !"bar", i32 37 }
2980 !2 = metadata !{ i32 2, metadata !"qux", i32 42 }
2981 !3 = metadata !{ i32 3, metadata !"qux",
2983 metadata !"foo", i32 1
2986 !llvm.module.flags = !{ !0, !1, !2, !3 }
2988 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
2989 if two or more ``!"foo"`` flags are seen is to emit an error if their
2990 values are not equal.
2992 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
2993 behavior if two or more ``!"bar"`` flags are seen is to use the value
2996 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
2997 behavior if two or more ``!"qux"`` flags are seen is to emit a
2998 warning if their values are not equal.
3000 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
3004 metadata !{ metadata !"foo", i32 1 }
3006 The behavior is to emit an error if the ``llvm.module.flags`` does not
3007 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
3010 Objective-C Garbage Collection Module Flags Metadata
3011 ----------------------------------------------------
3013 On the Mach-O platform, Objective-C stores metadata about garbage
3014 collection in a special section called "image info". The metadata
3015 consists of a version number and a bitmask specifying what types of
3016 garbage collection are supported (if any) by the file. If two or more
3017 modules are linked together their garbage collection metadata needs to
3018 be merged rather than appended together.
3020 The Objective-C garbage collection module flags metadata consists of the
3021 following key-value pairs:
3030 * - ``Objective-C Version``
3031 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
3033 * - ``Objective-C Image Info Version``
3034 - **[Required]** --- The version of the image info section. Currently
3037 * - ``Objective-C Image Info Section``
3038 - **[Required]** --- The section to place the metadata. Valid values are
3039 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
3040 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
3041 Objective-C ABI version 2.
3043 * - ``Objective-C Garbage Collection``
3044 - **[Required]** --- Specifies whether garbage collection is supported or
3045 not. Valid values are 0, for no garbage collection, and 2, for garbage
3046 collection supported.
3048 * - ``Objective-C GC Only``
3049 - **[Optional]** --- Specifies that only garbage collection is supported.
3050 If present, its value must be 6. This flag requires that the
3051 ``Objective-C Garbage Collection`` flag have the value 2.
3053 Some important flag interactions:
3055 - If a module with ``Objective-C Garbage Collection`` set to 0 is
3056 merged with a module with ``Objective-C Garbage Collection`` set to
3057 2, then the resulting module has the
3058 ``Objective-C Garbage Collection`` flag set to 0.
3059 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
3060 merged with a module with ``Objective-C GC Only`` set to 6.
3062 Automatic Linker Flags Module Flags Metadata
3063 --------------------------------------------
3065 Some targets support embedding flags to the linker inside individual object
3066 files. Typically this is used in conjunction with language extensions which
3067 allow source files to explicitly declare the libraries they depend on, and have
3068 these automatically be transmitted to the linker via object files.
3070 These flags are encoded in the IR using metadata in the module flags section,
3071 using the ``Linker Options`` key. The merge behavior for this flag is required
3072 to be ``AppendUnique``, and the value for the key is expected to be a metadata
3073 node which should be a list of other metadata nodes, each of which should be a
3074 list of metadata strings defining linker options.
3076 For example, the following metadata section specifies two separate sets of
3077 linker options, presumably to link against ``libz`` and the ``Cocoa``
3080 !0 = metadata !{ i32 6, metadata !"Linker Options",
3082 metadata !{ metadata !"-lz" },
3083 metadata !{ metadata !"-framework", metadata !"Cocoa" } } }
3084 !llvm.module.flags = !{ !0 }
3086 The metadata encoding as lists of lists of options, as opposed to a collapsed
3087 list of options, is chosen so that the IR encoding can use multiple option
3088 strings to specify e.g., a single library, while still having that specifier be
3089 preserved as an atomic element that can be recognized by a target specific
3090 assembly writer or object file emitter.
3092 Each individual option is required to be either a valid option for the target's
3093 linker, or an option that is reserved by the target specific assembly writer or
3094 object file emitter. No other aspect of these options is defined by the IR.
3096 .. _intrinsicglobalvariables:
3098 Intrinsic Global Variables
3099 ==========================
3101 LLVM has a number of "magic" global variables that contain data that
3102 affect code generation or other IR semantics. These are documented here.
3103 All globals of this sort should have a section specified as
3104 "``llvm.metadata``". This section and all globals that start with
3105 "``llvm.``" are reserved for use by LLVM.
3109 The '``llvm.used``' Global Variable
3110 -----------------------------------
3112 The ``@llvm.used`` global is an array which has
3113 :ref:`appending linkage <linkage_appending>`. This array contains a list of
3114 pointers to named global variables, functions and aliases which may optionally
3115 have a pointer cast formed of bitcast or getelementptr. For example, a legal
3118 .. code-block:: llvm
3123 @llvm.used = appending global [2 x i8*] [
3125 i8* bitcast (i32* @Y to i8*)
3126 ], section "llvm.metadata"
3128 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
3129 and linker are required to treat the symbol as if there is a reference to the
3130 symbol that it cannot see (which is why they have to be named). For example, if
3131 a variable has internal linkage and no references other than that from the
3132 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
3133 references from inline asms and other things the compiler cannot "see", and
3134 corresponds to "``attribute((used))``" in GNU C.
3136 On some targets, the code generator must emit a directive to the
3137 assembler or object file to prevent the assembler and linker from
3138 molesting the symbol.
3140 .. _gv_llvmcompilerused:
3142 The '``llvm.compiler.used``' Global Variable
3143 --------------------------------------------
3145 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
3146 directive, except that it only prevents the compiler from touching the
3147 symbol. On targets that support it, this allows an intelligent linker to
3148 optimize references to the symbol without being impeded as it would be
3151 This is a rare construct that should only be used in rare circumstances,
3152 and should not be exposed to source languages.
3154 .. _gv_llvmglobalctors:
3156 The '``llvm.global_ctors``' Global Variable
3157 -------------------------------------------
3159 .. code-block:: llvm
3161 %0 = type { i32, void ()* }
3162 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
3164 The ``@llvm.global_ctors`` array contains a list of constructor
3165 functions and associated priorities. The functions referenced by this
3166 array will be called in ascending order of priority (i.e. lowest first)
3167 when the module is loaded. The order of functions with the same priority
3170 .. _llvmglobaldtors:
3172 The '``llvm.global_dtors``' Global Variable
3173 -------------------------------------------
3175 .. code-block:: llvm
3177 %0 = type { i32, void ()* }
3178 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
3180 The ``@llvm.global_dtors`` array contains a list of destructor functions
3181 and associated priorities. The functions referenced by this array will
3182 be called in descending order of priority (i.e. highest first) when the
3183 module is loaded. The order of functions with the same priority is not
3186 Instruction Reference
3187 =====================
3189 The LLVM instruction set consists of several different classifications
3190 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
3191 instructions <binaryops>`, :ref:`bitwise binary
3192 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
3193 :ref:`other instructions <otherops>`.
3197 Terminator Instructions
3198 -----------------------
3200 As mentioned :ref:`previously <functionstructure>`, every basic block in a
3201 program ends with a "Terminator" instruction, which indicates which
3202 block should be executed after the current block is finished. These
3203 terminator instructions typically yield a '``void``' value: they produce
3204 control flow, not values (the one exception being the
3205 ':ref:`invoke <i_invoke>`' instruction).
3207 The terminator instructions are: ':ref:`ret <i_ret>`',
3208 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
3209 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
3210 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
3214 '``ret``' Instruction
3215 ^^^^^^^^^^^^^^^^^^^^^
3222 ret <type> <value> ; Return a value from a non-void function
3223 ret void ; Return from void function
3228 The '``ret``' instruction is used to return control flow (and optionally
3229 a value) from a function back to the caller.
3231 There are two forms of the '``ret``' instruction: one that returns a
3232 value and then causes control flow, and one that just causes control
3238 The '``ret``' instruction optionally accepts a single argument, the
3239 return value. The type of the return value must be a ':ref:`first
3240 class <t_firstclass>`' type.
3242 A function is not :ref:`well formed <wellformed>` if it it has a non-void
3243 return type and contains a '``ret``' instruction with no return value or
3244 a return value with a type that does not match its type, or if it has a
3245 void return type and contains a '``ret``' instruction with a return
3251 When the '``ret``' instruction is executed, control flow returns back to
3252 the calling function's context. If the caller is a
3253 ":ref:`call <i_call>`" instruction, execution continues at the
3254 instruction after the call. If the caller was an
3255 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
3256 beginning of the "normal" destination block. If the instruction returns
3257 a value, that value shall set the call or invoke instruction's return
3263 .. code-block:: llvm
3265 ret i32 5 ; Return an integer value of 5
3266 ret void ; Return from a void function
3267 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
3271 '``br``' Instruction
3272 ^^^^^^^^^^^^^^^^^^^^
3279 br i1 <cond>, label <iftrue>, label <iffalse>
3280 br label <dest> ; Unconditional branch
3285 The '``br``' instruction is used to cause control flow to transfer to a
3286 different basic block in the current function. There are two forms of
3287 this instruction, corresponding to a conditional branch and an
3288 unconditional branch.
3293 The conditional branch form of the '``br``' instruction takes a single
3294 '``i1``' value and two '``label``' values. The unconditional form of the
3295 '``br``' instruction takes a single '``label``' value as a target.
3300 Upon execution of a conditional '``br``' instruction, the '``i1``'
3301 argument is evaluated. If the value is ``true``, control flows to the
3302 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
3303 to the '``iffalse``' ``label`` argument.
3308 .. code-block:: llvm
3311 %cond = icmp eq i32 %a, %b
3312 br i1 %cond, label %IfEqual, label %IfUnequal
3320 '``switch``' Instruction
3321 ^^^^^^^^^^^^^^^^^^^^^^^^
3328 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
3333 The '``switch``' instruction is used to transfer control flow to one of
3334 several different places. It is a generalization of the '``br``'
3335 instruction, allowing a branch to occur to one of many possible
3341 The '``switch``' instruction uses three parameters: an integer
3342 comparison value '``value``', a default '``label``' destination, and an
3343 array of pairs of comparison value constants and '``label``'s. The table
3344 is not allowed to contain duplicate constant entries.
3349 The ``switch`` instruction specifies a table of values and destinations.
3350 When the '``switch``' instruction is executed, this table is searched
3351 for the given value. If the value is found, control flow is transferred
3352 to the corresponding destination; otherwise, control flow is transferred
3353 to the default destination.
3358 Depending on properties of the target machine and the particular
3359 ``switch`` instruction, this instruction may be code generated in
3360 different ways. For example, it could be generated as a series of
3361 chained conditional branches or with a lookup table.
3366 .. code-block:: llvm
3368 ; Emulate a conditional br instruction
3369 %Val = zext i1 %value to i32
3370 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
3372 ; Emulate an unconditional br instruction
3373 switch i32 0, label %dest [ ]
3375 ; Implement a jump table:
3376 switch i32 %val, label %otherwise [ i32 0, label %onzero
3378 i32 2, label %ontwo ]
3382 '``indirectbr``' Instruction
3383 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3390 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
3395 The '``indirectbr``' instruction implements an indirect branch to a
3396 label within the current function, whose address is specified by
3397 "``address``". Address must be derived from a
3398 :ref:`blockaddress <blockaddress>` constant.
3403 The '``address``' argument is the address of the label to jump to. The
3404 rest of the arguments indicate the full set of possible destinations
3405 that the address may point to. Blocks are allowed to occur multiple
3406 times in the destination list, though this isn't particularly useful.
3408 This destination list is required so that dataflow analysis has an
3409 accurate understanding of the CFG.
3414 Control transfers to the block specified in the address argument. All
3415 possible destination blocks must be listed in the label list, otherwise
3416 this instruction has undefined behavior. This implies that jumps to
3417 labels defined in other functions have undefined behavior as well.
3422 This is typically implemented with a jump through a register.
3427 .. code-block:: llvm
3429 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
3433 '``invoke``' Instruction
3434 ^^^^^^^^^^^^^^^^^^^^^^^^
3441 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
3442 to label <normal label> unwind label <exception label>
3447 The '``invoke``' instruction causes control to transfer to a specified
3448 function, with the possibility of control flow transfer to either the
3449 '``normal``' label or the '``exception``' label. If the callee function
3450 returns with the "``ret``" instruction, control flow will return to the
3451 "normal" label. If the callee (or any indirect callees) returns via the
3452 ":ref:`resume <i_resume>`" instruction or other exception handling
3453 mechanism, control is interrupted and continued at the dynamically
3454 nearest "exception" label.
3456 The '``exception``' label is a `landing
3457 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
3458 '``exception``' label is required to have the
3459 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
3460 information about the behavior of the program after unwinding happens,
3461 as its first non-PHI instruction. The restrictions on the
3462 "``landingpad``" instruction's tightly couples it to the "``invoke``"
3463 instruction, so that the important information contained within the
3464 "``landingpad``" instruction can't be lost through normal code motion.
3469 This instruction requires several arguments:
3471 #. The optional "cconv" marker indicates which :ref:`calling
3472 convention <callingconv>` the call should use. If none is
3473 specified, the call defaults to using C calling conventions.
3474 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
3475 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
3477 #. '``ptr to function ty``': shall be the signature of the pointer to
3478 function value being invoked. In most cases, this is a direct
3479 function invocation, but indirect ``invoke``'s are just as possible,
3480 branching off an arbitrary pointer to function value.
3481 #. '``function ptr val``': An LLVM value containing a pointer to a
3482 function to be invoked.
3483 #. '``function args``': argument list whose types match the function
3484 signature argument types and parameter attributes. All arguments must
3485 be of :ref:`first class <t_firstclass>` type. If the function signature
3486 indicates the function accepts a variable number of arguments, the
3487 extra arguments can be specified.
3488 #. '``normal label``': the label reached when the called function
3489 executes a '``ret``' instruction.
3490 #. '``exception label``': the label reached when a callee returns via
3491 the :ref:`resume <i_resume>` instruction or other exception handling
3493 #. The optional :ref:`function attributes <fnattrs>` list. Only
3494 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
3495 attributes are valid here.
3500 This instruction is designed to operate as a standard '``call``'
3501 instruction in most regards. The primary difference is that it
3502 establishes an association with a label, which is used by the runtime
3503 library to unwind the stack.
3505 This instruction is used in languages with destructors to ensure that
3506 proper cleanup is performed in the case of either a ``longjmp`` or a
3507 thrown exception. Additionally, this is important for implementation of
3508 '``catch``' clauses in high-level languages that support them.
3510 For the purposes of the SSA form, the definition of the value returned
3511 by the '``invoke``' instruction is deemed to occur on the edge from the
3512 current block to the "normal" label. If the callee unwinds then no
3513 return value is available.
3518 .. code-block:: llvm
3520 %retval = invoke i32 @Test(i32 15) to label %Continue
3521 unwind label %TestCleanup ; {i32}:retval set
3522 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
3523 unwind label %TestCleanup ; {i32}:retval set
3527 '``resume``' Instruction
3528 ^^^^^^^^^^^^^^^^^^^^^^^^
3535 resume <type> <value>
3540 The '``resume``' instruction is a terminator instruction that has no
3546 The '``resume``' instruction requires one argument, which must have the
3547 same type as the result of any '``landingpad``' instruction in the same
3553 The '``resume``' instruction resumes propagation of an existing
3554 (in-flight) exception whose unwinding was interrupted with a
3555 :ref:`landingpad <i_landingpad>` instruction.
3560 .. code-block:: llvm
3562 resume { i8*, i32 } %exn
3566 '``unreachable``' Instruction
3567 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3579 The '``unreachable``' instruction has no defined semantics. This
3580 instruction is used to inform the optimizer that a particular portion of
3581 the code is not reachable. This can be used to indicate that the code
3582 after a no-return function cannot be reached, and other facts.
3587 The '``unreachable``' instruction has no defined semantics.
3594 Binary operators are used to do most of the computation in a program.
3595 They require two operands of the same type, execute an operation on
3596 them, and produce a single value. The operands might represent multiple
3597 data, as is the case with the :ref:`vector <t_vector>` data type. The
3598 result value has the same type as its operands.
3600 There are several different binary operators:
3604 '``add``' Instruction
3605 ^^^^^^^^^^^^^^^^^^^^^
3612 <result> = add <ty> <op1>, <op2> ; yields {ty}:result
3613 <result> = add nuw <ty> <op1>, <op2> ; yields {ty}:result
3614 <result> = add nsw <ty> <op1>, <op2> ; yields {ty}:result
3615 <result> = add nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3620 The '``add``' instruction returns the sum of its two operands.
3625 The two arguments to the '``add``' instruction must be
3626 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3627 arguments must have identical types.
3632 The value produced is the integer sum of the two operands.
3634 If the sum has unsigned overflow, the result returned is the
3635 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3638 Because LLVM integers use a two's complement representation, this
3639 instruction is appropriate for both signed and unsigned integers.
3641 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3642 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3643 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
3644 unsigned and/or signed overflow, respectively, occurs.
3649 .. code-block:: llvm
3651 <result> = add i32 4, %var ; yields {i32}:result = 4 + %var
3655 '``fadd``' Instruction
3656 ^^^^^^^^^^^^^^^^^^^^^^
3663 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3668 The '``fadd``' instruction returns the sum of its two operands.
3673 The two arguments to the '``fadd``' instruction must be :ref:`floating
3674 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3675 Both arguments must have identical types.
3680 The value produced is the floating point sum of the two operands. This
3681 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
3682 which are optimization hints to enable otherwise unsafe floating point
3688 .. code-block:: llvm
3690 <result> = fadd float 4.0, %var ; yields {float}:result = 4.0 + %var
3692 '``sub``' Instruction
3693 ^^^^^^^^^^^^^^^^^^^^^
3700 <result> = sub <ty> <op1>, <op2> ; yields {ty}:result
3701 <result> = sub nuw <ty> <op1>, <op2> ; yields {ty}:result
3702 <result> = sub nsw <ty> <op1>, <op2> ; yields {ty}:result
3703 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3708 The '``sub``' instruction returns the difference of its two operands.
3710 Note that the '``sub``' instruction is used to represent the '``neg``'
3711 instruction present in most other intermediate representations.
3716 The two arguments to the '``sub``' instruction must be
3717 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3718 arguments must have identical types.
3723 The value produced is the integer difference of the two operands.
3725 If the difference has unsigned overflow, the result returned is the
3726 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3729 Because LLVM integers use a two's complement representation, this
3730 instruction is appropriate for both signed and unsigned integers.
3732 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3733 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3734 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
3735 unsigned and/or signed overflow, respectively, occurs.
3740 .. code-block:: llvm
3742 <result> = sub i32 4, %var ; yields {i32}:result = 4 - %var
3743 <result> = sub i32 0, %val ; yields {i32}:result = -%var
3747 '``fsub``' Instruction
3748 ^^^^^^^^^^^^^^^^^^^^^^
3755 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3760 The '``fsub``' instruction returns the difference of its two operands.
3762 Note that the '``fsub``' instruction is used to represent the '``fneg``'
3763 instruction present in most other intermediate representations.
3768 The two arguments to the '``fsub``' instruction must be :ref:`floating
3769 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3770 Both arguments must have identical types.
3775 The value produced is the floating point difference of the two operands.
3776 This instruction can also take any number of :ref:`fast-math
3777 flags <fastmath>`, which are optimization hints to enable otherwise
3778 unsafe floating point optimizations:
3783 .. code-block:: llvm
3785 <result> = fsub float 4.0, %var ; yields {float}:result = 4.0 - %var
3786 <result> = fsub float -0.0, %val ; yields {float}:result = -%var
3788 '``mul``' Instruction
3789 ^^^^^^^^^^^^^^^^^^^^^
3796 <result> = mul <ty> <op1>, <op2> ; yields {ty}:result
3797 <result> = mul nuw <ty> <op1>, <op2> ; yields {ty}:result
3798 <result> = mul nsw <ty> <op1>, <op2> ; yields {ty}:result
3799 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3804 The '``mul``' instruction returns the product of its two operands.
3809 The two arguments to the '``mul``' instruction must be
3810 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3811 arguments must have identical types.
3816 The value produced is the integer product of the two operands.
3818 If the result of the multiplication has unsigned overflow, the result
3819 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
3820 bit width of the result.
3822 Because LLVM integers use a two's complement representation, and the
3823 result is the same width as the operands, this instruction returns the
3824 correct result for both signed and unsigned integers. If a full product
3825 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
3826 sign-extended or zero-extended as appropriate to the width of the full
3829 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3830 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3831 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
3832 unsigned and/or signed overflow, respectively, occurs.
3837 .. code-block:: llvm
3839 <result> = mul i32 4, %var ; yields {i32}:result = 4 * %var
3843 '``fmul``' Instruction
3844 ^^^^^^^^^^^^^^^^^^^^^^
3851 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3856 The '``fmul``' instruction returns the product of its two operands.
3861 The two arguments to the '``fmul``' instruction must be :ref:`floating
3862 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3863 Both arguments must have identical types.
3868 The value produced is the floating point product of the two operands.
3869 This instruction can also take any number of :ref:`fast-math
3870 flags <fastmath>`, which are optimization hints to enable otherwise
3871 unsafe floating point optimizations:
3876 .. code-block:: llvm
3878 <result> = fmul float 4.0, %var ; yields {float}:result = 4.0 * %var
3880 '``udiv``' Instruction
3881 ^^^^^^^^^^^^^^^^^^^^^^
3888 <result> = udiv <ty> <op1>, <op2> ; yields {ty}:result
3889 <result> = udiv exact <ty> <op1>, <op2> ; yields {ty}:result
3894 The '``udiv``' instruction returns the quotient of its two operands.
3899 The two arguments to the '``udiv``' instruction must be
3900 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3901 arguments must have identical types.
3906 The value produced is the unsigned integer quotient of the two operands.
3908 Note that unsigned integer division and signed integer division are
3909 distinct operations; for signed integer division, use '``sdiv``'.
3911 Division by zero leads to undefined behavior.
3913 If the ``exact`` keyword is present, the result value of the ``udiv`` is
3914 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
3915 such, "((a udiv exact b) mul b) == a").
3920 .. code-block:: llvm
3922 <result> = udiv i32 4, %var ; yields {i32}:result = 4 / %var
3924 '``sdiv``' Instruction
3925 ^^^^^^^^^^^^^^^^^^^^^^
3932 <result> = sdiv <ty> <op1>, <op2> ; yields {ty}:result
3933 <result> = sdiv exact <ty> <op1>, <op2> ; yields {ty}:result
3938 The '``sdiv``' instruction returns the quotient of its two operands.
3943 The two arguments to the '``sdiv``' instruction must be
3944 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3945 arguments must have identical types.
3950 The value produced is the signed integer quotient of the two operands
3951 rounded towards zero.
3953 Note that signed integer division and unsigned integer division are
3954 distinct operations; for unsigned integer division, use '``udiv``'.
3956 Division by zero leads to undefined behavior. Overflow also leads to
3957 undefined behavior; this is a rare case, but can occur, for example, by
3958 doing a 32-bit division of -2147483648 by -1.
3960 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
3961 a :ref:`poison value <poisonvalues>` if the result would be rounded.
3966 .. code-block:: llvm
3968 <result> = sdiv i32 4, %var ; yields {i32}:result = 4 / %var
3972 '``fdiv``' Instruction
3973 ^^^^^^^^^^^^^^^^^^^^^^
3980 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3985 The '``fdiv``' instruction returns the quotient of its two operands.
3990 The two arguments to the '``fdiv``' instruction must be :ref:`floating
3991 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3992 Both arguments must have identical types.
3997 The value produced is the floating point quotient of the two operands.
3998 This instruction can also take any number of :ref:`fast-math
3999 flags <fastmath>`, which are optimization hints to enable otherwise
4000 unsafe floating point optimizations:
4005 .. code-block:: llvm
4007 <result> = fdiv float 4.0, %var ; yields {float}:result = 4.0 / %var
4009 '``urem``' Instruction
4010 ^^^^^^^^^^^^^^^^^^^^^^
4017 <result> = urem <ty> <op1>, <op2> ; yields {ty}:result
4022 The '``urem``' instruction returns the remainder from the unsigned
4023 division of its two arguments.
4028 The two arguments to the '``urem``' instruction must be
4029 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4030 arguments must have identical types.
4035 This instruction returns the unsigned integer *remainder* of a division.
4036 This instruction always performs an unsigned division to get the
4039 Note that unsigned integer remainder and signed integer remainder are
4040 distinct operations; for signed integer remainder, use '``srem``'.
4042 Taking the remainder of a division by zero leads to undefined behavior.
4047 .. code-block:: llvm
4049 <result> = urem i32 4, %var ; yields {i32}:result = 4 % %var
4051 '``srem``' Instruction
4052 ^^^^^^^^^^^^^^^^^^^^^^
4059 <result> = srem <ty> <op1>, <op2> ; yields {ty}:result
4064 The '``srem``' instruction returns the remainder from the signed
4065 division of its two operands. This instruction can also take
4066 :ref:`vector <t_vector>` versions of the values in which case the elements
4072 The two arguments to the '``srem``' instruction must be
4073 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4074 arguments must have identical types.
4079 This instruction returns the *remainder* of a division (where the result
4080 is either zero or has the same sign as the dividend, ``op1``), not the
4081 *modulo* operator (where the result is either zero or has the same sign
4082 as the divisor, ``op2``) of a value. For more information about the
4083 difference, see `The Math
4084 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
4085 table of how this is implemented in various languages, please see
4087 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
4089 Note that signed integer remainder and unsigned integer remainder are
4090 distinct operations; for unsigned integer remainder, use '``urem``'.
4092 Taking the remainder of a division by zero leads to undefined behavior.
4093 Overflow also leads to undefined behavior; this is a rare case, but can
4094 occur, for example, by taking the remainder of a 32-bit division of
4095 -2147483648 by -1. (The remainder doesn't actually overflow, but this
4096 rule lets srem be implemented using instructions that return both the
4097 result of the division and the remainder.)
4102 .. code-block:: llvm
4104 <result> = srem i32 4, %var ; yields {i32}:result = 4 % %var
4108 '``frem``' Instruction
4109 ^^^^^^^^^^^^^^^^^^^^^^
4116 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
4121 The '``frem``' instruction returns the remainder from the division of
4127 The two arguments to the '``frem``' instruction must be :ref:`floating
4128 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4129 Both arguments must have identical types.
4134 This instruction returns the *remainder* of a division. The remainder
4135 has the same sign as the dividend. This instruction can also take any
4136 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
4137 to enable otherwise unsafe floating point optimizations:
4142 .. code-block:: llvm
4144 <result> = frem float 4.0, %var ; yields {float}:result = 4.0 % %var
4148 Bitwise Binary Operations
4149 -------------------------
4151 Bitwise binary operators are used to do various forms of bit-twiddling
4152 in a program. They are generally very efficient instructions and can
4153 commonly be strength reduced from other instructions. They require two
4154 operands of the same type, execute an operation on them, and produce a
4155 single value. The resulting value is the same type as its operands.
4157 '``shl``' Instruction
4158 ^^^^^^^^^^^^^^^^^^^^^
4165 <result> = shl <ty> <op1>, <op2> ; yields {ty}:result
4166 <result> = shl nuw <ty> <op1>, <op2> ; yields {ty}:result
4167 <result> = shl nsw <ty> <op1>, <op2> ; yields {ty}:result
4168 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
4173 The '``shl``' instruction returns the first operand shifted to the left
4174 a specified number of bits.
4179 Both arguments to the '``shl``' instruction must be the same
4180 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4181 '``op2``' is treated as an unsigned value.
4186 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
4187 where ``n`` is the width of the result. If ``op2`` is (statically or
4188 dynamically) negative or equal to or larger than the number of bits in
4189 ``op1``, the result is undefined. If the arguments are vectors, each
4190 vector element of ``op1`` is shifted by the corresponding shift amount
4193 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
4194 value <poisonvalues>` if it shifts out any non-zero bits. If the
4195 ``nsw`` keyword is present, then the shift produces a :ref:`poison
4196 value <poisonvalues>` if it shifts out any bits that disagree with the
4197 resultant sign bit. As such, NUW/NSW have the same semantics as they
4198 would if the shift were expressed as a mul instruction with the same
4199 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
4204 .. code-block:: llvm
4206 <result> = shl i32 4, %var ; yields {i32}: 4 << %var
4207 <result> = shl i32 4, 2 ; yields {i32}: 16
4208 <result> = shl i32 1, 10 ; yields {i32}: 1024
4209 <result> = shl i32 1, 32 ; undefined
4210 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
4212 '``lshr``' Instruction
4213 ^^^^^^^^^^^^^^^^^^^^^^
4220 <result> = lshr <ty> <op1>, <op2> ; yields {ty}:result
4221 <result> = lshr exact <ty> <op1>, <op2> ; yields {ty}:result
4226 The '``lshr``' instruction (logical shift right) returns the first
4227 operand shifted to the right a specified number of bits with zero fill.
4232 Both arguments to the '``lshr``' instruction must be the same
4233 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4234 '``op2``' is treated as an unsigned value.
4239 This instruction always performs a logical shift right operation. The
4240 most significant bits of the result will be filled with zero bits after
4241 the shift. If ``op2`` is (statically or dynamically) equal to or larger
4242 than the number of bits in ``op1``, the result is undefined. If the
4243 arguments are vectors, each vector element of ``op1`` is shifted by the
4244 corresponding shift amount in ``op2``.
4246 If the ``exact`` keyword is present, the result value of the ``lshr`` is
4247 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4253 .. code-block:: llvm
4255 <result> = lshr i32 4, 1 ; yields {i32}:result = 2
4256 <result> = lshr i32 4, 2 ; yields {i32}:result = 1
4257 <result> = lshr i8 4, 3 ; yields {i8}:result = 0
4258 <result> = lshr i8 -2, 1 ; yields {i8}:result = 0x7F
4259 <result> = lshr i32 1, 32 ; undefined
4260 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
4262 '``ashr``' Instruction
4263 ^^^^^^^^^^^^^^^^^^^^^^
4270 <result> = ashr <ty> <op1>, <op2> ; yields {ty}:result
4271 <result> = ashr exact <ty> <op1>, <op2> ; yields {ty}:result
4276 The '``ashr``' instruction (arithmetic shift right) returns the first
4277 operand shifted to the right a specified number of bits with sign
4283 Both arguments to the '``ashr``' instruction must be the same
4284 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4285 '``op2``' is treated as an unsigned value.
4290 This instruction always performs an arithmetic shift right operation,
4291 The most significant bits of the result will be filled with the sign bit
4292 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
4293 than the number of bits in ``op1``, the result is undefined. If the
4294 arguments are vectors, each vector element of ``op1`` is shifted by the
4295 corresponding shift amount in ``op2``.
4297 If the ``exact`` keyword is present, the result value of the ``ashr`` is
4298 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4304 .. code-block:: llvm
4306 <result> = ashr i32 4, 1 ; yields {i32}:result = 2
4307 <result> = ashr i32 4, 2 ; yields {i32}:result = 1
4308 <result> = ashr i8 4, 3 ; yields {i8}:result = 0
4309 <result> = ashr i8 -2, 1 ; yields {i8}:result = -1
4310 <result> = ashr i32 1, 32 ; undefined
4311 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
4313 '``and``' Instruction
4314 ^^^^^^^^^^^^^^^^^^^^^
4321 <result> = and <ty> <op1>, <op2> ; yields {ty}:result
4326 The '``and``' instruction returns the bitwise logical and of its two
4332 The two arguments to the '``and``' instruction must be
4333 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4334 arguments must have identical types.
4339 The truth table used for the '``and``' instruction is:
4356 .. code-block:: llvm
4358 <result> = and i32 4, %var ; yields {i32}:result = 4 & %var
4359 <result> = and i32 15, 40 ; yields {i32}:result = 8
4360 <result> = and i32 4, 8 ; yields {i32}:result = 0
4362 '``or``' Instruction
4363 ^^^^^^^^^^^^^^^^^^^^
4370 <result> = or <ty> <op1>, <op2> ; yields {ty}:result
4375 The '``or``' instruction returns the bitwise logical inclusive or of its
4381 The two arguments to the '``or``' instruction must be
4382 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4383 arguments must have identical types.
4388 The truth table used for the '``or``' instruction is:
4407 <result> = or i32 4, %var ; yields {i32}:result = 4 | %var
4408 <result> = or i32 15, 40 ; yields {i32}:result = 47
4409 <result> = or i32 4, 8 ; yields {i32}:result = 12
4411 '``xor``' Instruction
4412 ^^^^^^^^^^^^^^^^^^^^^
4419 <result> = xor <ty> <op1>, <op2> ; yields {ty}:result
4424 The '``xor``' instruction returns the bitwise logical exclusive or of
4425 its two operands. The ``xor`` is used to implement the "one's
4426 complement" operation, which is the "~" operator in C.
4431 The two arguments to the '``xor``' instruction must be
4432 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4433 arguments must have identical types.
4438 The truth table used for the '``xor``' instruction is:
4455 .. code-block:: llvm
4457 <result> = xor i32 4, %var ; yields {i32}:result = 4 ^ %var
4458 <result> = xor i32 15, 40 ; yields {i32}:result = 39
4459 <result> = xor i32 4, 8 ; yields {i32}:result = 12
4460 <result> = xor i32 %V, -1 ; yields {i32}:result = ~%V
4465 LLVM supports several instructions to represent vector operations in a
4466 target-independent manner. These instructions cover the element-access
4467 and vector-specific operations needed to process vectors effectively.
4468 While LLVM does directly support these vector operations, many
4469 sophisticated algorithms will want to use target-specific intrinsics to
4470 take full advantage of a specific target.
4472 .. _i_extractelement:
4474 '``extractelement``' Instruction
4475 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4482 <result> = extractelement <n x <ty>> <val>, i32 <idx> ; yields <ty>
4487 The '``extractelement``' instruction extracts a single scalar element
4488 from a vector at a specified index.
4493 The first operand of an '``extractelement``' instruction is a value of
4494 :ref:`vector <t_vector>` type. The second operand is an index indicating
4495 the position from which to extract the element. The index may be a
4501 The result is a scalar of the same type as the element type of ``val``.
4502 Its value is the value at position ``idx`` of ``val``. If ``idx``
4503 exceeds the length of ``val``, the results are undefined.
4508 .. code-block:: llvm
4510 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
4512 .. _i_insertelement:
4514 '``insertelement``' Instruction
4515 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4522 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, i32 <idx> ; yields <n x <ty>>
4527 The '``insertelement``' instruction inserts a scalar element into a
4528 vector at a specified index.
4533 The first operand of an '``insertelement``' instruction is a value of
4534 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
4535 type must equal the element type of the first operand. The third operand
4536 is an index indicating the position at which to insert the value. The
4537 index may be a variable.
4542 The result is a vector of the same type as ``val``. Its element values
4543 are those of ``val`` except at position ``idx``, where it gets the value
4544 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
4550 .. code-block:: llvm
4552 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
4554 .. _i_shufflevector:
4556 '``shufflevector``' Instruction
4557 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4564 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
4569 The '``shufflevector``' instruction constructs a permutation of elements
4570 from two input vectors, returning a vector with the same element type as
4571 the input and length that is the same as the shuffle mask.
4576 The first two operands of a '``shufflevector``' instruction are vectors
4577 with the same type. The third argument is a shuffle mask whose element
4578 type is always 'i32'. The result of the instruction is a vector whose
4579 length is the same as the shuffle mask and whose element type is the
4580 same as the element type of the first two operands.
4582 The shuffle mask operand is required to be a constant vector with either
4583 constant integer or undef values.
4588 The elements of the two input vectors are numbered from left to right
4589 across both of the vectors. The shuffle mask operand specifies, for each
4590 element of the result vector, which element of the two input vectors the
4591 result element gets. The element selector may be undef (meaning "don't
4592 care") and the second operand may be undef if performing a shuffle from
4598 .. code-block:: llvm
4600 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4601 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
4602 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
4603 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
4604 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
4605 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
4606 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4607 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
4609 Aggregate Operations
4610 --------------------
4612 LLVM supports several instructions for working with
4613 :ref:`aggregate <t_aggregate>` values.
4617 '``extractvalue``' Instruction
4618 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4625 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
4630 The '``extractvalue``' instruction extracts the value of a member field
4631 from an :ref:`aggregate <t_aggregate>` value.
4636 The first operand of an '``extractvalue``' instruction is a value of
4637 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
4638 constant indices to specify which value to extract in a similar manner
4639 as indices in a '``getelementptr``' instruction.
4641 The major differences to ``getelementptr`` indexing are:
4643 - Since the value being indexed is not a pointer, the first index is
4644 omitted and assumed to be zero.
4645 - At least one index must be specified.
4646 - Not only struct indices but also array indices must be in bounds.
4651 The result is the value at the position in the aggregate specified by
4657 .. code-block:: llvm
4659 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
4663 '``insertvalue``' Instruction
4664 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4671 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
4676 The '``insertvalue``' instruction inserts a value into a member field in
4677 an :ref:`aggregate <t_aggregate>` value.
4682 The first operand of an '``insertvalue``' instruction is a value of
4683 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
4684 a first-class value to insert. The following operands are constant
4685 indices indicating the position at which to insert the value in a
4686 similar manner as indices in a '``extractvalue``' instruction. The value
4687 to insert must have the same type as the value identified by the
4693 The result is an aggregate of the same type as ``val``. Its value is
4694 that of ``val`` except that the value at the position specified by the
4695 indices is that of ``elt``.
4700 .. code-block:: llvm
4702 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
4703 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
4704 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 ; yields {i32 1, float %val}
4708 Memory Access and Addressing Operations
4709 ---------------------------------------
4711 A key design point of an SSA-based representation is how it represents
4712 memory. In LLVM, no memory locations are in SSA form, which makes things
4713 very simple. This section describes how to read, write, and allocate
4718 '``alloca``' Instruction
4719 ^^^^^^^^^^^^^^^^^^^^^^^^
4726 <result> = alloca <type>[, inalloca][, <ty> <NumElements>][, align <alignment>] ; yields {type*}:result
4731 The '``alloca``' instruction allocates memory on the stack frame of the
4732 currently executing function, to be automatically released when this
4733 function returns to its caller. The object is always allocated in the
4734 generic address space (address space zero).
4739 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
4740 bytes of memory on the runtime stack, returning a pointer of the
4741 appropriate type to the program. If "NumElements" is specified, it is
4742 the number of elements allocated, otherwise "NumElements" is defaulted
4743 to be one. If a constant alignment is specified, the value result of the
4744 allocation is guaranteed to be aligned to at least that boundary. If not
4745 specified, or if zero, the target can choose to align the allocation on
4746 any convenient boundary compatible with the type.
4748 '``type``' may be any sized type.
4753 Memory is allocated; a pointer is returned. The operation is undefined
4754 if there is insufficient stack space for the allocation. '``alloca``'d
4755 memory is automatically released when the function returns. The
4756 '``alloca``' instruction is commonly used to represent automatic
4757 variables that must have an address available. When the function returns
4758 (either with the ``ret`` or ``resume`` instructions), the memory is
4759 reclaimed. Allocating zero bytes is legal, but the result is undefined.
4760 The order in which memory is allocated (ie., which way the stack grows)
4766 .. code-block:: llvm
4768 %ptr = alloca i32 ; yields {i32*}:ptr
4769 %ptr = alloca i32, i32 4 ; yields {i32*}:ptr
4770 %ptr = alloca i32, i32 4, align 1024 ; yields {i32*}:ptr
4771 %ptr = alloca i32, align 1024 ; yields {i32*}:ptr
4775 '``load``' Instruction
4776 ^^^^^^^^^^^^^^^^^^^^^^
4783 <result> = load [volatile] <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>]
4784 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
4785 !<index> = !{ i32 1 }
4790 The '``load``' instruction is used to read from memory.
4795 The argument to the ``load`` instruction specifies the memory address
4796 from which to load. The pointer must point to a :ref:`first
4797 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
4798 then the optimizer is not allowed to modify the number or order of
4799 execution of this ``load`` with other :ref:`volatile
4800 operations <volatile>`.
4802 If the ``load`` is marked as ``atomic``, it takes an extra
4803 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
4804 ``release`` and ``acq_rel`` orderings are not valid on ``load``
4805 instructions. Atomic loads produce :ref:`defined <memmodel>` results
4806 when they may see multiple atomic stores. The type of the pointee must
4807 be an integer type whose bit width is a power of two greater than or
4808 equal to eight and less than or equal to a target-specific size limit.
4809 ``align`` must be explicitly specified on atomic loads, and the load has
4810 undefined behavior if the alignment is not set to a value which is at
4811 least the size in bytes of the pointee. ``!nontemporal`` does not have
4812 any defined semantics for atomic loads.
4814 The optional constant ``align`` argument specifies the alignment of the
4815 operation (that is, the alignment of the memory address). A value of 0
4816 or an omitted ``align`` argument means that the operation has the ABI
4817 alignment for the target. It is the responsibility of the code emitter
4818 to ensure that the alignment information is correct. Overestimating the
4819 alignment results in undefined behavior. Underestimating the alignment
4820 may produce less efficient code. An alignment of 1 is always safe.
4822 The optional ``!nontemporal`` metadata must reference a single
4823 metadata name ``<index>`` corresponding to a metadata node with one
4824 ``i32`` entry of value 1. The existence of the ``!nontemporal``
4825 metadata on the instruction tells the optimizer and code generator
4826 that this load is not expected to be reused in the cache. The code
4827 generator may select special instructions to save cache bandwidth, such
4828 as the ``MOVNT`` instruction on x86.
4830 The optional ``!invariant.load`` metadata must reference a single
4831 metadata name ``<index>`` corresponding to a metadata node with no
4832 entries. The existence of the ``!invariant.load`` metadata on the
4833 instruction tells the optimizer and code generator that this load
4834 address points to memory which does not change value during program
4835 execution. The optimizer may then move this load around, for example, by
4836 hoisting it out of loops using loop invariant code motion.
4841 The location of memory pointed to is loaded. If the value being loaded
4842 is of scalar type then the number of bytes read does not exceed the
4843 minimum number of bytes needed to hold all bits of the type. For
4844 example, loading an ``i24`` reads at most three bytes. When loading a
4845 value of a type like ``i20`` with a size that is not an integral number
4846 of bytes, the result is undefined if the value was not originally
4847 written using a store of the same type.
4852 .. code-block:: llvm
4854 %ptr = alloca i32 ; yields {i32*}:ptr
4855 store i32 3, i32* %ptr ; yields {void}
4856 %val = load i32* %ptr ; yields {i32}:val = i32 3
4860 '``store``' Instruction
4861 ^^^^^^^^^^^^^^^^^^^^^^^
4868 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields {void}
4869 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields {void}
4874 The '``store``' instruction is used to write to memory.
4879 There are two arguments to the ``store`` instruction: a value to store
4880 and an address at which to store it. The type of the ``<pointer>``
4881 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
4882 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
4883 then the optimizer is not allowed to modify the number or order of
4884 execution of this ``store`` with other :ref:`volatile
4885 operations <volatile>`.
4887 If the ``store`` is marked as ``atomic``, it takes an extra
4888 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
4889 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
4890 instructions. Atomic loads produce :ref:`defined <memmodel>` results
4891 when they may see multiple atomic stores. The type of the pointee must
4892 be an integer type whose bit width is a power of two greater than or
4893 equal to eight and less than or equal to a target-specific size limit.
4894 ``align`` must be explicitly specified on atomic stores, and the store
4895 has undefined behavior if the alignment is not set to a value which is
4896 at least the size in bytes of the pointee. ``!nontemporal`` does not
4897 have any defined semantics for atomic stores.
4899 The optional constant ``align`` argument specifies the alignment of the
4900 operation (that is, the alignment of the memory address). A value of 0
4901 or an omitted ``align`` argument means that the operation has the ABI
4902 alignment for the target. It is the responsibility of the code emitter
4903 to ensure that the alignment information is correct. Overestimating the
4904 alignment results in undefined behavior. Underestimating the
4905 alignment may produce less efficient code. An alignment of 1 is always
4908 The optional ``!nontemporal`` metadata must reference a single metadata
4909 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
4910 value 1. The existence of the ``!nontemporal`` metadata on the instruction
4911 tells the optimizer and code generator that this load is not expected to
4912 be reused in the cache. The code generator may select special
4913 instructions to save cache bandwidth, such as the MOVNT instruction on
4919 The contents of memory are updated to contain ``<value>`` at the
4920 location specified by the ``<pointer>`` operand. If ``<value>`` is
4921 of scalar type then the number of bytes written does not exceed the
4922 minimum number of bytes needed to hold all bits of the type. For
4923 example, storing an ``i24`` writes at most three bytes. When writing a
4924 value of a type like ``i20`` with a size that is not an integral number
4925 of bytes, it is unspecified what happens to the extra bits that do not
4926 belong to the type, but they will typically be overwritten.
4931 .. code-block:: llvm
4933 %ptr = alloca i32 ; yields {i32*}:ptr
4934 store i32 3, i32* %ptr ; yields {void}
4935 %val = load i32* %ptr ; yields {i32}:val = i32 3
4939 '``fence``' Instruction
4940 ^^^^^^^^^^^^^^^^^^^^^^^
4947 fence [singlethread] <ordering> ; yields {void}
4952 The '``fence``' instruction is used to introduce happens-before edges
4958 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
4959 defines what *synchronizes-with* edges they add. They can only be given
4960 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
4965 A fence A which has (at least) ``release`` ordering semantics
4966 *synchronizes with* a fence B with (at least) ``acquire`` ordering
4967 semantics if and only if there exist atomic operations X and Y, both
4968 operating on some atomic object M, such that A is sequenced before X, X
4969 modifies M (either directly or through some side effect of a sequence
4970 headed by X), Y is sequenced before B, and Y observes M. This provides a
4971 *happens-before* dependency between A and B. Rather than an explicit
4972 ``fence``, one (but not both) of the atomic operations X or Y might
4973 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
4974 still *synchronize-with* the explicit ``fence`` and establish the
4975 *happens-before* edge.
4977 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
4978 ``acquire`` and ``release`` semantics specified above, participates in
4979 the global program order of other ``seq_cst`` operations and/or fences.
4981 The optional ":ref:`singlethread <singlethread>`" argument specifies
4982 that the fence only synchronizes with other fences in the same thread.
4983 (This is useful for interacting with signal handlers.)
4988 .. code-block:: llvm
4990 fence acquire ; yields {void}
4991 fence singlethread seq_cst ; yields {void}
4995 '``cmpxchg``' Instruction
4996 ^^^^^^^^^^^^^^^^^^^^^^^^^
5003 cmpxchg [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <ordering> ; yields {ty}
5008 The '``cmpxchg``' instruction is used to atomically modify memory. It
5009 loads a value in memory and compares it to a given value. If they are
5010 equal, it stores a new value into the memory.
5015 There are three arguments to the '``cmpxchg``' instruction: an address
5016 to operate on, a value to compare to the value currently be at that
5017 address, and a new value to place at that address if the compared values
5018 are equal. The type of '<cmp>' must be an integer type whose bit width
5019 is a power of two greater than or equal to eight and less than or equal
5020 to a target-specific size limit. '<cmp>' and '<new>' must have the same
5021 type, and the type of '<pointer>' must be a pointer to that type. If the
5022 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
5023 to modify the number or order of execution of this ``cmpxchg`` with
5024 other :ref:`volatile operations <volatile>`.
5026 The :ref:`ordering <ordering>` argument specifies how this ``cmpxchg``
5027 synchronizes with other atomic operations.
5029 The optional "``singlethread``" argument declares that the ``cmpxchg``
5030 is only atomic with respect to code (usually signal handlers) running in
5031 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
5032 respect to all other code in the system.
5034 The pointer passed into cmpxchg must have alignment greater than or
5035 equal to the size in memory of the operand.
5040 The contents of memory at the location specified by the '``<pointer>``'
5041 operand is read and compared to '``<cmp>``'; if the read value is the
5042 equal, '``<new>``' is written. The original value at the location is
5045 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose
5046 of identifying release sequences. A failed ``cmpxchg`` is equivalent to an
5047 atomic load with an ordering parameter determined by dropping any
5048 ``release`` part of the ``cmpxchg``'s ordering.
5053 .. code-block:: llvm
5056 %orig = atomic load i32* %ptr unordered ; yields {i32}
5060 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
5061 %squared = mul i32 %cmp, %cmp
5062 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared ; yields {i32}
5063 %success = icmp eq i32 %cmp, %old
5064 br i1 %success, label %done, label %loop
5071 '``atomicrmw``' Instruction
5072 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
5079 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields {ty}
5084 The '``atomicrmw``' instruction is used to atomically modify memory.
5089 There are three arguments to the '``atomicrmw``' instruction: an
5090 operation to apply, an address whose value to modify, an argument to the
5091 operation. The operation must be one of the following keywords:
5105 The type of '<value>' must be an integer type whose bit width is a power
5106 of two greater than or equal to eight and less than or equal to a
5107 target-specific size limit. The type of the '``<pointer>``' operand must
5108 be a pointer to that type. If the ``atomicrmw`` is marked as
5109 ``volatile``, then the optimizer is not allowed to modify the number or
5110 order of execution of this ``atomicrmw`` with other :ref:`volatile
5111 operations <volatile>`.
5116 The contents of memory at the location specified by the '``<pointer>``'
5117 operand are atomically read, modified, and written back. The original
5118 value at the location is returned. The modification is specified by the
5121 - xchg: ``*ptr = val``
5122 - add: ``*ptr = *ptr + val``
5123 - sub: ``*ptr = *ptr - val``
5124 - and: ``*ptr = *ptr & val``
5125 - nand: ``*ptr = ~(*ptr & val)``
5126 - or: ``*ptr = *ptr | val``
5127 - xor: ``*ptr = *ptr ^ val``
5128 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
5129 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
5130 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
5132 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
5138 .. code-block:: llvm
5140 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields {i32}
5142 .. _i_getelementptr:
5144 '``getelementptr``' Instruction
5145 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5152 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
5153 <result> = getelementptr inbounds <pty>* <ptrval>{, <ty> <idx>}*
5154 <result> = getelementptr <ptr vector> ptrval, <vector index type> idx
5159 The '``getelementptr``' instruction is used to get the address of a
5160 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
5161 address calculation only and does not access memory.
5166 The first argument is always a pointer or a vector of pointers, and
5167 forms the basis of the calculation. The remaining arguments are indices
5168 that indicate which of the elements of the aggregate object are indexed.
5169 The interpretation of each index is dependent on the type being indexed
5170 into. The first index always indexes the pointer value given as the
5171 first argument, the second index indexes a value of the type pointed to
5172 (not necessarily the value directly pointed to, since the first index
5173 can be non-zero), etc. The first type indexed into must be a pointer
5174 value, subsequent types can be arrays, vectors, and structs. Note that
5175 subsequent types being indexed into can never be pointers, since that
5176 would require loading the pointer before continuing calculation.
5178 The type of each index argument depends on the type it is indexing into.
5179 When indexing into a (optionally packed) structure, only ``i32`` integer
5180 **constants** are allowed (when using a vector of indices they must all
5181 be the **same** ``i32`` integer constant). When indexing into an array,
5182 pointer or vector, integers of any width are allowed, and they are not
5183 required to be constant. These integers are treated as signed values
5186 For example, let's consider a C code fragment and how it gets compiled
5202 int *foo(struct ST *s) {
5203 return &s[1].Z.B[5][13];
5206 The LLVM code generated by Clang is:
5208 .. code-block:: llvm
5210 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
5211 %struct.ST = type { i32, double, %struct.RT }
5213 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
5215 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5222 In the example above, the first index is indexing into the
5223 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
5224 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
5225 indexes into the third element of the structure, yielding a
5226 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
5227 structure. The third index indexes into the second element of the
5228 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
5229 dimensions of the array are subscripted into, yielding an '``i32``'
5230 type. The '``getelementptr``' instruction returns a pointer to this
5231 element, thus computing a value of '``i32*``' type.
5233 Note that it is perfectly legal to index partially through a structure,
5234 returning a pointer to an inner element. Because of this, the LLVM code
5235 for the given testcase is equivalent to:
5237 .. code-block:: llvm
5239 define i32* @foo(%struct.ST* %s) {
5240 %t1 = getelementptr %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
5241 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
5242 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
5243 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
5244 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
5248 If the ``inbounds`` keyword is present, the result value of the
5249 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
5250 pointer is not an *in bounds* address of an allocated object, or if any
5251 of the addresses that would be formed by successive addition of the
5252 offsets implied by the indices to the base address with infinitely
5253 precise signed arithmetic are not an *in bounds* address of that
5254 allocated object. The *in bounds* addresses for an allocated object are
5255 all the addresses that point into the object, plus the address one byte
5256 past the end. In cases where the base is a vector of pointers the
5257 ``inbounds`` keyword applies to each of the computations element-wise.
5259 If the ``inbounds`` keyword is not present, the offsets are added to the
5260 base address with silently-wrapping two's complement arithmetic. If the
5261 offsets have a different width from the pointer, they are sign-extended
5262 or truncated to the width of the pointer. The result value of the
5263 ``getelementptr`` may be outside the object pointed to by the base
5264 pointer. The result value may not necessarily be used to access memory
5265 though, even if it happens to point into allocated storage. See the
5266 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
5269 The getelementptr instruction is often confusing. For some more insight
5270 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
5275 .. code-block:: llvm
5277 ; yields [12 x i8]*:aptr
5278 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5280 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
5282 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
5284 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
5286 In cases where the pointer argument is a vector of pointers, each index
5287 must be a vector with the same number of elements. For example:
5289 .. code-block:: llvm
5291 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5293 Conversion Operations
5294 ---------------------
5296 The instructions in this category are the conversion instructions
5297 (casting) which all take a single operand and a type. They perform
5298 various bit conversions on the operand.
5300 '``trunc .. to``' Instruction
5301 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5308 <result> = trunc <ty> <value> to <ty2> ; yields ty2
5313 The '``trunc``' instruction truncates its operand to the type ``ty2``.
5318 The '``trunc``' instruction takes a value to trunc, and a type to trunc
5319 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
5320 of the same number of integers. The bit size of the ``value`` must be
5321 larger than the bit size of the destination type, ``ty2``. Equal sized
5322 types are not allowed.
5327 The '``trunc``' instruction truncates the high order bits in ``value``
5328 and converts the remaining bits to ``ty2``. Since the source size must
5329 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
5330 It will always truncate bits.
5335 .. code-block:: llvm
5337 %X = trunc i32 257 to i8 ; yields i8:1
5338 %Y = trunc i32 123 to i1 ; yields i1:true
5339 %Z = trunc i32 122 to i1 ; yields i1:false
5340 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
5342 '``zext .. to``' Instruction
5343 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5350 <result> = zext <ty> <value> to <ty2> ; yields ty2
5355 The '``zext``' instruction zero extends its operand to type ``ty2``.
5360 The '``zext``' instruction takes a value to cast, and a type to cast it
5361 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5362 the same number of integers. The bit size of the ``value`` must be
5363 smaller than the bit size of the destination type, ``ty2``.
5368 The ``zext`` fills the high order bits of the ``value`` with zero bits
5369 until it reaches the size of the destination type, ``ty2``.
5371 When zero extending from i1, the result will always be either 0 or 1.
5376 .. code-block:: llvm
5378 %X = zext i32 257 to i64 ; yields i64:257
5379 %Y = zext i1 true to i32 ; yields i32:1
5380 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5382 '``sext .. to``' Instruction
5383 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5390 <result> = sext <ty> <value> to <ty2> ; yields ty2
5395 The '``sext``' sign extends ``value`` to the type ``ty2``.
5400 The '``sext``' instruction takes a value to cast, and a type to cast it
5401 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5402 the same number of integers. The bit size of the ``value`` must be
5403 smaller than the bit size of the destination type, ``ty2``.
5408 The '``sext``' instruction performs a sign extension by copying the sign
5409 bit (highest order bit) of the ``value`` until it reaches the bit size
5410 of the type ``ty2``.
5412 When sign extending from i1, the extension always results in -1 or 0.
5417 .. code-block:: llvm
5419 %X = sext i8 -1 to i16 ; yields i16 :65535
5420 %Y = sext i1 true to i32 ; yields i32:-1
5421 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5423 '``fptrunc .. to``' Instruction
5424 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5431 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
5436 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
5441 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
5442 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
5443 The size of ``value`` must be larger than the size of ``ty2``. This
5444 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
5449 The '``fptrunc``' instruction truncates a ``value`` from a larger
5450 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
5451 point <t_floating>` type. If the value cannot fit within the
5452 destination type, ``ty2``, then the results are undefined.
5457 .. code-block:: llvm
5459 %X = fptrunc double 123.0 to float ; yields float:123.0
5460 %Y = fptrunc double 1.0E+300 to float ; yields undefined
5462 '``fpext .. to``' Instruction
5463 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5470 <result> = fpext <ty> <value> to <ty2> ; yields ty2
5475 The '``fpext``' extends a floating point ``value`` to a larger floating
5481 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
5482 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
5483 to. The source type must be smaller than the destination type.
5488 The '``fpext``' instruction extends the ``value`` from a smaller
5489 :ref:`floating point <t_floating>` type to a larger :ref:`floating
5490 point <t_floating>` type. The ``fpext`` cannot be used to make a
5491 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
5492 *no-op cast* for a floating point cast.
5497 .. code-block:: llvm
5499 %X = fpext float 3.125 to double ; yields double:3.125000e+00
5500 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
5502 '``fptoui .. to``' Instruction
5503 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5510 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
5515 The '``fptoui``' converts a floating point ``value`` to its unsigned
5516 integer equivalent of type ``ty2``.
5521 The '``fptoui``' instruction takes a value to cast, which must be a
5522 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5523 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5524 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5525 type with the same number of elements as ``ty``
5530 The '``fptoui``' instruction converts its :ref:`floating
5531 point <t_floating>` operand into the nearest (rounding towards zero)
5532 unsigned integer value. If the value cannot fit in ``ty2``, the results
5538 .. code-block:: llvm
5540 %X = fptoui double 123.0 to i32 ; yields i32:123
5541 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
5542 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
5544 '``fptosi .. to``' Instruction
5545 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5552 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
5557 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
5558 ``value`` to type ``ty2``.
5563 The '``fptosi``' instruction takes a value to cast, which must be a
5564 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5565 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5566 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5567 type with the same number of elements as ``ty``
5572 The '``fptosi``' instruction converts its :ref:`floating
5573 point <t_floating>` operand into the nearest (rounding towards zero)
5574 signed integer value. If the value cannot fit in ``ty2``, the results
5580 .. code-block:: llvm
5582 %X = fptosi double -123.0 to i32 ; yields i32:-123
5583 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
5584 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
5586 '``uitofp .. to``' Instruction
5587 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5594 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
5599 The '``uitofp``' instruction regards ``value`` as an unsigned integer
5600 and converts that value to the ``ty2`` type.
5605 The '``uitofp``' instruction takes a value to cast, which must be a
5606 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5607 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5608 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5609 type with the same number of elements as ``ty``
5614 The '``uitofp``' instruction interprets its operand as an unsigned
5615 integer quantity and converts it to the corresponding floating point
5616 value. If the value cannot fit in the floating point value, the results
5622 .. code-block:: llvm
5624 %X = uitofp i32 257 to float ; yields float:257.0
5625 %Y = uitofp i8 -1 to double ; yields double:255.0
5627 '``sitofp .. to``' Instruction
5628 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5635 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
5640 The '``sitofp``' instruction regards ``value`` as a signed integer and
5641 converts that value to the ``ty2`` type.
5646 The '``sitofp``' instruction takes a value to cast, which must be a
5647 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5648 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5649 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5650 type with the same number of elements as ``ty``
5655 The '``sitofp``' instruction interprets its operand as a signed integer
5656 quantity and converts it to the corresponding floating point value. If
5657 the value cannot fit in the floating point value, the results are
5663 .. code-block:: llvm
5665 %X = sitofp i32 257 to float ; yields float:257.0
5666 %Y = sitofp i8 -1 to double ; yields double:-1.0
5670 '``ptrtoint .. to``' Instruction
5671 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5678 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
5683 The '``ptrtoint``' instruction converts the pointer or a vector of
5684 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
5689 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
5690 a a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
5691 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
5692 a vector of integers type.
5697 The '``ptrtoint``' instruction converts ``value`` to integer type
5698 ``ty2`` by interpreting the pointer value as an integer and either
5699 truncating or zero extending that value to the size of the integer type.
5700 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
5701 ``value`` is larger than ``ty2`` then a truncation is done. If they are
5702 the same size, then nothing is done (*no-op cast*) other than a type
5708 .. code-block:: llvm
5710 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
5711 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
5712 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
5716 '``inttoptr .. to``' Instruction
5717 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5724 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
5729 The '``inttoptr``' instruction converts an integer ``value`` to a
5730 pointer type, ``ty2``.
5735 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
5736 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
5742 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
5743 applying either a zero extension or a truncation depending on the size
5744 of the integer ``value``. If ``value`` is larger than the size of a
5745 pointer then a truncation is done. If ``value`` is smaller than the size
5746 of a pointer then a zero extension is done. If they are the same size,
5747 nothing is done (*no-op cast*).
5752 .. code-block:: llvm
5754 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
5755 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
5756 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
5757 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
5761 '``bitcast .. to``' Instruction
5762 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5769 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
5774 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
5780 The '``bitcast``' instruction takes a value to cast, which must be a
5781 non-aggregate first class value, and a type to cast it to, which must
5782 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
5783 bit sizes of ``value`` and the destination type, ``ty2``, must be
5784 identical. If the source type is a pointer, the destination type must
5785 also be a pointer of the same size. This instruction supports bitwise
5786 conversion of vectors to integers and to vectors of other types (as
5787 long as they have the same size).
5792 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
5793 is always a *no-op cast* because no bits change with this
5794 conversion. The conversion is done as if the ``value`` had been stored
5795 to memory and read back as type ``ty2``. Pointer (or vector of
5796 pointers) types may only be converted to other pointer (or vector of
5797 pointers) types with the same address space through this instruction.
5798 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
5799 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
5804 .. code-block:: llvm
5806 %X = bitcast i8 255 to i8 ; yields i8 :-1
5807 %Y = bitcast i32* %x to sint* ; yields sint*:%x
5808 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
5809 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
5811 .. _i_addrspacecast:
5813 '``addrspacecast .. to``' Instruction
5814 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5821 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
5826 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
5827 address space ``n`` to type ``pty2`` in address space ``m``.
5832 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
5833 to cast and a pointer type to cast it to, which must have a different
5839 The '``addrspacecast``' instruction converts the pointer value
5840 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
5841 value modification, depending on the target and the address space
5842 pair. Pointer conversions within the same address space must be
5843 performed with the ``bitcast`` instruction. Note that if the address space
5844 conversion is legal then both result and operand refer to the same memory
5850 .. code-block:: llvm
5852 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
5853 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
5854 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
5861 The instructions in this category are the "miscellaneous" instructions,
5862 which defy better classification.
5866 '``icmp``' Instruction
5867 ^^^^^^^^^^^^^^^^^^^^^^
5874 <result> = icmp <cond> <ty> <op1>, <op2> ; yields {i1} or {<N x i1>}:result
5879 The '``icmp``' instruction returns a boolean value or a vector of
5880 boolean values based on comparison of its two integer, integer vector,
5881 pointer, or pointer vector operands.
5886 The '``icmp``' instruction takes three operands. The first operand is
5887 the condition code indicating the kind of comparison to perform. It is
5888 not a value, just a keyword. The possible condition code are:
5891 #. ``ne``: not equal
5892 #. ``ugt``: unsigned greater than
5893 #. ``uge``: unsigned greater or equal
5894 #. ``ult``: unsigned less than
5895 #. ``ule``: unsigned less or equal
5896 #. ``sgt``: signed greater than
5897 #. ``sge``: signed greater or equal
5898 #. ``slt``: signed less than
5899 #. ``sle``: signed less or equal
5901 The remaining two arguments must be :ref:`integer <t_integer>` or
5902 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
5903 must also be identical types.
5908 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
5909 code given as ``cond``. The comparison performed always yields either an
5910 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
5912 #. ``eq``: yields ``true`` if the operands are equal, ``false``
5913 otherwise. No sign interpretation is necessary or performed.
5914 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
5915 otherwise. No sign interpretation is necessary or performed.
5916 #. ``ugt``: interprets the operands as unsigned values and yields
5917 ``true`` if ``op1`` is greater than ``op2``.
5918 #. ``uge``: interprets the operands as unsigned values and yields
5919 ``true`` if ``op1`` is greater than or equal to ``op2``.
5920 #. ``ult``: interprets the operands as unsigned values and yields
5921 ``true`` if ``op1`` is less than ``op2``.
5922 #. ``ule``: interprets the operands as unsigned values and yields
5923 ``true`` if ``op1`` is less than or equal to ``op2``.
5924 #. ``sgt``: interprets the operands as signed values and yields ``true``
5925 if ``op1`` is greater than ``op2``.
5926 #. ``sge``: interprets the operands as signed values and yields ``true``
5927 if ``op1`` is greater than or equal to ``op2``.
5928 #. ``slt``: interprets the operands as signed values and yields ``true``
5929 if ``op1`` is less than ``op2``.
5930 #. ``sle``: interprets the operands as signed values and yields ``true``
5931 if ``op1`` is less than or equal to ``op2``.
5933 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
5934 are compared as if they were integers.
5936 If the operands are integer vectors, then they are compared element by
5937 element. The result is an ``i1`` vector with the same number of elements
5938 as the values being compared. Otherwise, the result is an ``i1``.
5943 .. code-block:: llvm
5945 <result> = icmp eq i32 4, 5 ; yields: result=false
5946 <result> = icmp ne float* %X, %X ; yields: result=false
5947 <result> = icmp ult i16 4, 5 ; yields: result=true
5948 <result> = icmp sgt i16 4, 5 ; yields: result=false
5949 <result> = icmp ule i16 -4, 5 ; yields: result=false
5950 <result> = icmp sge i16 4, 5 ; yields: result=false
5952 Note that the code generator does not yet support vector types with the
5953 ``icmp`` instruction.
5957 '``fcmp``' Instruction
5958 ^^^^^^^^^^^^^^^^^^^^^^
5965 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields {i1} or {<N x i1>}:result
5970 The '``fcmp``' instruction returns a boolean value or vector of boolean
5971 values based on comparison of its operands.
5973 If the operands are floating point scalars, then the result type is a
5974 boolean (:ref:`i1 <t_integer>`).
5976 If the operands are floating point vectors, then the result type is a
5977 vector of boolean with the same number of elements as the operands being
5983 The '``fcmp``' instruction takes three operands. The first operand is
5984 the condition code indicating the kind of comparison to perform. It is
5985 not a value, just a keyword. The possible condition code are:
5987 #. ``false``: no comparison, always returns false
5988 #. ``oeq``: ordered and equal
5989 #. ``ogt``: ordered and greater than
5990 #. ``oge``: ordered and greater than or equal
5991 #. ``olt``: ordered and less than
5992 #. ``ole``: ordered and less than or equal
5993 #. ``one``: ordered and not equal
5994 #. ``ord``: ordered (no nans)
5995 #. ``ueq``: unordered or equal
5996 #. ``ugt``: unordered or greater than
5997 #. ``uge``: unordered or greater than or equal
5998 #. ``ult``: unordered or less than
5999 #. ``ule``: unordered or less than or equal
6000 #. ``une``: unordered or not equal
6001 #. ``uno``: unordered (either nans)
6002 #. ``true``: no comparison, always returns true
6004 *Ordered* means that neither operand is a QNAN while *unordered* means
6005 that either operand may be a QNAN.
6007 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
6008 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
6009 type. They must have identical types.
6014 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
6015 condition code given as ``cond``. If the operands are vectors, then the
6016 vectors are compared element by element. Each comparison performed
6017 always yields an :ref:`i1 <t_integer>` result, as follows:
6019 #. ``false``: always yields ``false``, regardless of operands.
6020 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
6021 is equal to ``op2``.
6022 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
6023 is greater than ``op2``.
6024 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
6025 is greater than or equal to ``op2``.
6026 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
6027 is less than ``op2``.
6028 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
6029 is less than or equal to ``op2``.
6030 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
6031 is not equal to ``op2``.
6032 #. ``ord``: yields ``true`` if both operands are not a QNAN.
6033 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
6035 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
6036 greater than ``op2``.
6037 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
6038 greater than or equal to ``op2``.
6039 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
6041 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
6042 less than or equal to ``op2``.
6043 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
6044 not equal to ``op2``.
6045 #. ``uno``: yields ``true`` if either operand is a QNAN.
6046 #. ``true``: always yields ``true``, regardless of operands.
6051 .. code-block:: llvm
6053 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
6054 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
6055 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
6056 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
6058 Note that the code generator does not yet support vector types with the
6059 ``fcmp`` instruction.
6063 '``phi``' Instruction
6064 ^^^^^^^^^^^^^^^^^^^^^
6071 <result> = phi <ty> [ <val0>, <label0>], ...
6076 The '``phi``' instruction is used to implement the φ node in the SSA
6077 graph representing the function.
6082 The type of the incoming values is specified with the first type field.
6083 After this, the '``phi``' instruction takes a list of pairs as
6084 arguments, with one pair for each predecessor basic block of the current
6085 block. Only values of :ref:`first class <t_firstclass>` type may be used as
6086 the value arguments to the PHI node. Only labels may be used as the
6089 There must be no non-phi instructions between the start of a basic block
6090 and the PHI instructions: i.e. PHI instructions must be first in a basic
6093 For the purposes of the SSA form, the use of each incoming value is
6094 deemed to occur on the edge from the corresponding predecessor block to
6095 the current block (but after any definition of an '``invoke``'
6096 instruction's return value on the same edge).
6101 At runtime, the '``phi``' instruction logically takes on the value
6102 specified by the pair corresponding to the predecessor basic block that
6103 executed just prior to the current block.
6108 .. code-block:: llvm
6110 Loop: ; Infinite loop that counts from 0 on up...
6111 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
6112 %nextindvar = add i32 %indvar, 1
6117 '``select``' Instruction
6118 ^^^^^^^^^^^^^^^^^^^^^^^^
6125 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
6127 selty is either i1 or {<N x i1>}
6132 The '``select``' instruction is used to choose one value based on a
6133 condition, without branching.
6138 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
6139 values indicating the condition, and two values of the same :ref:`first
6140 class <t_firstclass>` type. If the val1/val2 are vectors and the
6141 condition is a scalar, then entire vectors are selected, not individual
6147 If the condition is an i1 and it evaluates to 1, the instruction returns
6148 the first value argument; otherwise, it returns the second value
6151 If the condition is a vector of i1, then the value arguments must be
6152 vectors of the same size, and the selection is done element by element.
6157 .. code-block:: llvm
6159 %X = select i1 true, i8 17, i8 42 ; yields i8:17
6163 '``call``' Instruction
6164 ^^^^^^^^^^^^^^^^^^^^^^
6171 <result> = [tail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
6176 The '``call``' instruction represents a simple function call.
6181 This instruction requires several arguments:
6183 #. The optional "tail" marker indicates that the callee function does
6184 not access any allocas or varargs in the caller. Note that calls may
6185 be marked "tail" even if they do not occur before a
6186 :ref:`ret <i_ret>` instruction. If the "tail" marker is present, the
6187 function call is eligible for tail call optimization, but `might not
6188 in fact be optimized into a jump <CodeGenerator.html#tailcallopt>`_.
6189 The code generator may optimize calls marked "tail" with either 1)
6190 automatic `sibling call
6191 optimization <CodeGenerator.html#sibcallopt>`_ when the caller and
6192 callee have matching signatures, or 2) forced tail call optimization
6193 when the following extra requirements are met:
6195 - Caller and callee both have the calling convention ``fastcc``.
6196 - The call is in tail position (ret immediately follows call and ret
6197 uses value of call or is void).
6198 - Option ``-tailcallopt`` is enabled, or
6199 ``llvm::GuaranteedTailCallOpt`` is ``true``.
6200 - `Platform specific constraints are
6201 met. <CodeGenerator.html#tailcallopt>`_
6203 #. The optional "cconv" marker indicates which :ref:`calling
6204 convention <callingconv>` the call should use. If none is
6205 specified, the call defaults to using C calling conventions. The
6206 calling convention of the call must match the calling convention of
6207 the target function, or else the behavior is undefined.
6208 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
6209 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
6211 #. '``ty``': the type of the call instruction itself which is also the
6212 type of the return value. Functions that return no value are marked
6214 #. '``fnty``': shall be the signature of the pointer to function value
6215 being invoked. The argument types must match the types implied by
6216 this signature. This type can be omitted if the function is not
6217 varargs and if the function type does not return a pointer to a
6219 #. '``fnptrval``': An LLVM value containing a pointer to a function to
6220 be invoked. In most cases, this is a direct function invocation, but
6221 indirect ``call``'s are just as possible, calling an arbitrary pointer
6223 #. '``function args``': argument list whose types match the function
6224 signature argument types and parameter attributes. All arguments must
6225 be of :ref:`first class <t_firstclass>` type. If the function signature
6226 indicates the function accepts a variable number of arguments, the
6227 extra arguments can be specified.
6228 #. The optional :ref:`function attributes <fnattrs>` list. Only
6229 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
6230 attributes are valid here.
6235 The '``call``' instruction is used to cause control flow to transfer to
6236 a specified function, with its incoming arguments bound to the specified
6237 values. Upon a '``ret``' instruction in the called function, control
6238 flow continues with the instruction after the function call, and the
6239 return value of the function is bound to the result argument.
6244 .. code-block:: llvm
6246 %retval = call i32 @test(i32 %argc)
6247 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
6248 %X = tail call i32 @foo() ; yields i32
6249 %Y = tail call fastcc i32 @foo() ; yields i32
6250 call void %foo(i8 97 signext)
6252 %struct.A = type { i32, i8 }
6253 %r = call %struct.A @foo() ; yields { 32, i8 }
6254 %gr = extractvalue %struct.A %r, 0 ; yields i32
6255 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
6256 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
6257 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
6259 llvm treats calls to some functions with names and arguments that match
6260 the standard C99 library as being the C99 library functions, and may
6261 perform optimizations or generate code for them under that assumption.
6262 This is something we'd like to change in the future to provide better
6263 support for freestanding environments and non-C-based languages.
6267 '``va_arg``' Instruction
6268 ^^^^^^^^^^^^^^^^^^^^^^^^
6275 <resultval> = va_arg <va_list*> <arglist>, <argty>
6280 The '``va_arg``' instruction is used to access arguments passed through
6281 the "variable argument" area of a function call. It is used to implement
6282 the ``va_arg`` macro in C.
6287 This instruction takes a ``va_list*`` value and the type of the
6288 argument. It returns a value of the specified argument type and
6289 increments the ``va_list`` to point to the next argument. The actual
6290 type of ``va_list`` is target specific.
6295 The '``va_arg``' instruction loads an argument of the specified type
6296 from the specified ``va_list`` and causes the ``va_list`` to point to
6297 the next argument. For more information, see the variable argument
6298 handling :ref:`Intrinsic Functions <int_varargs>`.
6300 It is legal for this instruction to be called in a function which does
6301 not take a variable number of arguments, for example, the ``vfprintf``
6304 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
6305 function <intrinsics>` because it takes a type as an argument.
6310 See the :ref:`variable argument processing <int_varargs>` section.
6312 Note that the code generator does not yet fully support va\_arg on many
6313 targets. Also, it does not currently support va\_arg with aggregate
6314 types on any target.
6318 '``landingpad``' Instruction
6319 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6326 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
6327 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
6329 <clause> := catch <type> <value>
6330 <clause> := filter <array constant type> <array constant>
6335 The '``landingpad``' instruction is used by `LLVM's exception handling
6336 system <ExceptionHandling.html#overview>`_ to specify that a basic block
6337 is a landing pad --- one where the exception lands, and corresponds to the
6338 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
6339 defines values supplied by the personality function (``pers_fn``) upon
6340 re-entry to the function. The ``resultval`` has the type ``resultty``.
6345 This instruction takes a ``pers_fn`` value. This is the personality
6346 function associated with the unwinding mechanism. The optional
6347 ``cleanup`` flag indicates that the landing pad block is a cleanup.
6349 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
6350 contains the global variable representing the "type" that may be caught
6351 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
6352 clause takes an array constant as its argument. Use
6353 "``[0 x i8**] undef``" for a filter which cannot throw. The
6354 '``landingpad``' instruction must contain *at least* one ``clause`` or
6355 the ``cleanup`` flag.
6360 The '``landingpad``' instruction defines the values which are set by the
6361 personality function (``pers_fn``) upon re-entry to the function, and
6362 therefore the "result type" of the ``landingpad`` instruction. As with
6363 calling conventions, how the personality function results are
6364 represented in LLVM IR is target specific.
6366 The clauses are applied in order from top to bottom. If two
6367 ``landingpad`` instructions are merged together through inlining, the
6368 clauses from the calling function are appended to the list of clauses.
6369 When the call stack is being unwound due to an exception being thrown,
6370 the exception is compared against each ``clause`` in turn. If it doesn't
6371 match any of the clauses, and the ``cleanup`` flag is not set, then
6372 unwinding continues further up the call stack.
6374 The ``landingpad`` instruction has several restrictions:
6376 - A landing pad block is a basic block which is the unwind destination
6377 of an '``invoke``' instruction.
6378 - A landing pad block must have a '``landingpad``' instruction as its
6379 first non-PHI instruction.
6380 - There can be only one '``landingpad``' instruction within the landing
6382 - A basic block that is not a landing pad block may not include a
6383 '``landingpad``' instruction.
6384 - All '``landingpad``' instructions in a function must have the same
6385 personality function.
6390 .. code-block:: llvm
6392 ;; A landing pad which can catch an integer.
6393 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6395 ;; A landing pad that is a cleanup.
6396 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6398 ;; A landing pad which can catch an integer and can only throw a double.
6399 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6401 filter [1 x i8**] [@_ZTId]
6408 LLVM supports the notion of an "intrinsic function". These functions
6409 have well known names and semantics and are required to follow certain
6410 restrictions. Overall, these intrinsics represent an extension mechanism
6411 for the LLVM language that does not require changing all of the
6412 transformations in LLVM when adding to the language (or the bitcode
6413 reader/writer, the parser, etc...).
6415 Intrinsic function names must all start with an "``llvm.``" prefix. This
6416 prefix is reserved in LLVM for intrinsic names; thus, function names may
6417 not begin with this prefix. Intrinsic functions must always be external
6418 functions: you cannot define the body of intrinsic functions. Intrinsic
6419 functions may only be used in call or invoke instructions: it is illegal
6420 to take the address of an intrinsic function. Additionally, because
6421 intrinsic functions are part of the LLVM language, it is required if any
6422 are added that they be documented here.
6424 Some intrinsic functions can be overloaded, i.e., the intrinsic
6425 represents a family of functions that perform the same operation but on
6426 different data types. Because LLVM can represent over 8 million
6427 different integer types, overloading is used commonly to allow an
6428 intrinsic function to operate on any integer type. One or more of the
6429 argument types or the result type can be overloaded to accept any
6430 integer type. Argument types may also be defined as exactly matching a
6431 previous argument's type or the result type. This allows an intrinsic
6432 function which accepts multiple arguments, but needs all of them to be
6433 of the same type, to only be overloaded with respect to a single
6434 argument or the result.
6436 Overloaded intrinsics will have the names of its overloaded argument
6437 types encoded into its function name, each preceded by a period. Only
6438 those types which are overloaded result in a name suffix. Arguments
6439 whose type is matched against another type do not. For example, the
6440 ``llvm.ctpop`` function can take an integer of any width and returns an
6441 integer of exactly the same integer width. This leads to a family of
6442 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
6443 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
6444 overloaded, and only one type suffix is required. Because the argument's
6445 type is matched against the return type, it does not require its own
6448 To learn how to add an intrinsic function, please see the `Extending
6449 LLVM Guide <ExtendingLLVM.html>`_.
6453 Variable Argument Handling Intrinsics
6454 -------------------------------------
6456 Variable argument support is defined in LLVM with the
6457 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
6458 functions. These functions are related to the similarly named macros
6459 defined in the ``<stdarg.h>`` header file.
6461 All of these functions operate on arguments that use a target-specific
6462 value type "``va_list``". The LLVM assembly language reference manual
6463 does not define what this type is, so all transformations should be
6464 prepared to handle these functions regardless of the type used.
6466 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
6467 variable argument handling intrinsic functions are used.
6469 .. code-block:: llvm
6471 define i32 @test(i32 %X, ...) {
6472 ; Initialize variable argument processing
6474 %ap2 = bitcast i8** %ap to i8*
6475 call void @llvm.va_start(i8* %ap2)
6477 ; Read a single integer argument
6478 %tmp = va_arg i8** %ap, i32
6480 ; Demonstrate usage of llvm.va_copy and llvm.va_end
6482 %aq2 = bitcast i8** %aq to i8*
6483 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
6484 call void @llvm.va_end(i8* %aq2)
6486 ; Stop processing of arguments.
6487 call void @llvm.va_end(i8* %ap2)
6491 declare void @llvm.va_start(i8*)
6492 declare void @llvm.va_copy(i8*, i8*)
6493 declare void @llvm.va_end(i8*)
6497 '``llvm.va_start``' Intrinsic
6498 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6505 declare void @llvm.va_start(i8* <arglist>)
6510 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
6511 subsequent use by ``va_arg``.
6516 The argument is a pointer to a ``va_list`` element to initialize.
6521 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
6522 available in C. In a target-dependent way, it initializes the
6523 ``va_list`` element to which the argument points, so that the next call
6524 to ``va_arg`` will produce the first variable argument passed to the
6525 function. Unlike the C ``va_start`` macro, this intrinsic does not need
6526 to know the last argument of the function as the compiler can figure
6529 '``llvm.va_end``' Intrinsic
6530 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6537 declare void @llvm.va_end(i8* <arglist>)
6542 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
6543 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
6548 The argument is a pointer to a ``va_list`` to destroy.
6553 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
6554 available in C. In a target-dependent way, it destroys the ``va_list``
6555 element to which the argument points. Calls to
6556 :ref:`llvm.va_start <int_va_start>` and
6557 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
6562 '``llvm.va_copy``' Intrinsic
6563 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6570 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
6575 The '``llvm.va_copy``' intrinsic copies the current argument position
6576 from the source argument list to the destination argument list.
6581 The first argument is a pointer to a ``va_list`` element to initialize.
6582 The second argument is a pointer to a ``va_list`` element to copy from.
6587 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
6588 available in C. In a target-dependent way, it copies the source
6589 ``va_list`` element into the destination ``va_list`` element. This
6590 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
6591 arbitrarily complex and require, for example, memory allocation.
6593 Accurate Garbage Collection Intrinsics
6594 --------------------------------------
6596 LLVM support for `Accurate Garbage Collection <GarbageCollection.html>`_
6597 (GC) requires the implementation and generation of these intrinsics.
6598 These intrinsics allow identification of :ref:`GC roots on the
6599 stack <int_gcroot>`, as well as garbage collector implementations that
6600 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
6601 Front-ends for type-safe garbage collected languages should generate
6602 these intrinsics to make use of the LLVM garbage collectors. For more
6603 details, see `Accurate Garbage Collection with
6604 LLVM <GarbageCollection.html>`_.
6606 The garbage collection intrinsics only operate on objects in the generic
6607 address space (address space zero).
6611 '``llvm.gcroot``' Intrinsic
6612 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6619 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
6624 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
6625 the code generator, and allows some metadata to be associated with it.
6630 The first argument specifies the address of a stack object that contains
6631 the root pointer. The second pointer (which must be either a constant or
6632 a global value address) contains the meta-data to be associated with the
6638 At runtime, a call to this intrinsic stores a null pointer into the
6639 "ptrloc" location. At compile-time, the code generator generates
6640 information to allow the runtime to find the pointer at GC safe points.
6641 The '``llvm.gcroot``' intrinsic may only be used in a function which
6642 :ref:`specifies a GC algorithm <gc>`.
6646 '``llvm.gcread``' Intrinsic
6647 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6654 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
6659 The '``llvm.gcread``' intrinsic identifies reads of references from heap
6660 locations, allowing garbage collector implementations that require read
6666 The second argument is the address to read from, which should be an
6667 address allocated from the garbage collector. The first object is a
6668 pointer to the start of the referenced object, if needed by the language
6669 runtime (otherwise null).
6674 The '``llvm.gcread``' intrinsic has the same semantics as a load
6675 instruction, but may be replaced with substantially more complex code by
6676 the garbage collector runtime, as needed. The '``llvm.gcread``'
6677 intrinsic may only be used in a function which :ref:`specifies a GC
6682 '``llvm.gcwrite``' Intrinsic
6683 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6690 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
6695 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
6696 locations, allowing garbage collector implementations that require write
6697 barriers (such as generational or reference counting collectors).
6702 The first argument is the reference to store, the second is the start of
6703 the object to store it to, and the third is the address of the field of
6704 Obj to store to. If the runtime does not require a pointer to the
6705 object, Obj may be null.
6710 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
6711 instruction, but may be replaced with substantially more complex code by
6712 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
6713 intrinsic may only be used in a function which :ref:`specifies a GC
6716 Code Generator Intrinsics
6717 -------------------------
6719 These intrinsics are provided by LLVM to expose special features that
6720 may only be implemented with code generator support.
6722 '``llvm.returnaddress``' Intrinsic
6723 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6730 declare i8 *@llvm.returnaddress(i32 <level>)
6735 The '``llvm.returnaddress``' intrinsic attempts to compute a
6736 target-specific value indicating the return address of the current
6737 function or one of its callers.
6742 The argument to this intrinsic indicates which function to return the
6743 address for. Zero indicates the calling function, one indicates its
6744 caller, etc. The argument is **required** to be a constant integer
6750 The '``llvm.returnaddress``' intrinsic either returns a pointer
6751 indicating the return address of the specified call frame, or zero if it
6752 cannot be identified. The value returned by this intrinsic is likely to
6753 be incorrect or 0 for arguments other than zero, so it should only be
6754 used for debugging purposes.
6756 Note that calling this intrinsic does not prevent function inlining or
6757 other aggressive transformations, so the value returned may not be that
6758 of the obvious source-language caller.
6760 '``llvm.frameaddress``' Intrinsic
6761 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6768 declare i8* @llvm.frameaddress(i32 <level>)
6773 The '``llvm.frameaddress``' intrinsic attempts to return the
6774 target-specific frame pointer value for the specified stack frame.
6779 The argument to this intrinsic indicates which function to return the
6780 frame pointer for. Zero indicates the calling function, one indicates
6781 its caller, etc. The argument is **required** to be a constant integer
6787 The '``llvm.frameaddress``' intrinsic either returns a pointer
6788 indicating the frame address of the specified call frame, or zero if it
6789 cannot be identified. The value returned by this intrinsic is likely to
6790 be incorrect or 0 for arguments other than zero, so it should only be
6791 used for debugging purposes.
6793 Note that calling this intrinsic does not prevent function inlining or
6794 other aggressive transformations, so the value returned may not be that
6795 of the obvious source-language caller.
6799 '``llvm.stacksave``' Intrinsic
6800 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6807 declare i8* @llvm.stacksave()
6812 The '``llvm.stacksave``' intrinsic is used to remember the current state
6813 of the function stack, for use with
6814 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
6815 implementing language features like scoped automatic variable sized
6821 This intrinsic returns a opaque pointer value that can be passed to
6822 :ref:`llvm.stackrestore <int_stackrestore>`. When an
6823 ``llvm.stackrestore`` intrinsic is executed with a value saved from
6824 ``llvm.stacksave``, it effectively restores the state of the stack to
6825 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
6826 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
6827 were allocated after the ``llvm.stacksave`` was executed.
6829 .. _int_stackrestore:
6831 '``llvm.stackrestore``' Intrinsic
6832 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6839 declare void @llvm.stackrestore(i8* %ptr)
6844 The '``llvm.stackrestore``' intrinsic is used to restore the state of
6845 the function stack to the state it was in when the corresponding
6846 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
6847 useful for implementing language features like scoped automatic variable
6848 sized arrays in C99.
6853 See the description for :ref:`llvm.stacksave <int_stacksave>`.
6855 '``llvm.prefetch``' Intrinsic
6856 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6863 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
6868 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
6869 insert a prefetch instruction if supported; otherwise, it is a noop.
6870 Prefetches have no effect on the behavior of the program but can change
6871 its performance characteristics.
6876 ``address`` is the address to be prefetched, ``rw`` is the specifier
6877 determining if the fetch should be for a read (0) or write (1), and
6878 ``locality`` is a temporal locality specifier ranging from (0) - no
6879 locality, to (3) - extremely local keep in cache. The ``cache type``
6880 specifies whether the prefetch is performed on the data (1) or
6881 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
6882 arguments must be constant integers.
6887 This intrinsic does not modify the behavior of the program. In
6888 particular, prefetches cannot trap and do not produce a value. On
6889 targets that support this intrinsic, the prefetch can provide hints to
6890 the processor cache for better performance.
6892 '``llvm.pcmarker``' Intrinsic
6893 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6900 declare void @llvm.pcmarker(i32 <id>)
6905 The '``llvm.pcmarker``' intrinsic is a method to export a Program
6906 Counter (PC) in a region of code to simulators and other tools. The
6907 method is target specific, but it is expected that the marker will use
6908 exported symbols to transmit the PC of the marker. The marker makes no
6909 guarantees that it will remain with any specific instruction after
6910 optimizations. It is possible that the presence of a marker will inhibit
6911 optimizations. The intended use is to be inserted after optimizations to
6912 allow correlations of simulation runs.
6917 ``id`` is a numerical id identifying the marker.
6922 This intrinsic does not modify the behavior of the program. Backends
6923 that do not support this intrinsic may ignore it.
6925 '``llvm.readcyclecounter``' Intrinsic
6926 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6933 declare i64 @llvm.readcyclecounter()
6938 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
6939 counter register (or similar low latency, high accuracy clocks) on those
6940 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6941 should map to RPCC. As the backing counters overflow quickly (on the
6942 order of 9 seconds on alpha), this should only be used for small
6948 When directly supported, reading the cycle counter should not modify any
6949 memory. Implementations are allowed to either return a application
6950 specific value or a system wide value. On backends without support, this
6951 is lowered to a constant 0.
6953 Note that runtime support may be conditional on the privilege-level code is
6954 running at and the host platform.
6956 Standard C Library Intrinsics
6957 -----------------------------
6959 LLVM provides intrinsics for a few important standard C library
6960 functions. These intrinsics allow source-language front-ends to pass
6961 information about the alignment of the pointer arguments to the code
6962 generator, providing opportunity for more efficient code generation.
6966 '``llvm.memcpy``' Intrinsic
6967 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6972 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
6973 integer bit width and for different address spaces. Not all targets
6974 support all bit widths however.
6978 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
6979 i32 <len>, i32 <align>, i1 <isvolatile>)
6980 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
6981 i64 <len>, i32 <align>, i1 <isvolatile>)
6986 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
6987 source location to the destination location.
6989 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
6990 intrinsics do not return a value, takes extra alignment/isvolatile
6991 arguments and the pointers can be in specified address spaces.
6996 The first argument is a pointer to the destination, the second is a
6997 pointer to the source. The third argument is an integer argument
6998 specifying the number of bytes to copy, the fourth argument is the
6999 alignment of the source and destination locations, and the fifth is a
7000 boolean indicating a volatile access.
7002 If the call to this intrinsic has an alignment value that is not 0 or 1,
7003 then the caller guarantees that both the source and destination pointers
7004 are aligned to that boundary.
7006 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
7007 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7008 very cleanly specified and it is unwise to depend on it.
7013 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7014 source location to the destination location, which are not allowed to
7015 overlap. It copies "len" bytes of memory over. If the argument is known
7016 to be aligned to some boundary, this can be specified as the fourth
7017 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
7019 '``llvm.memmove``' Intrinsic
7020 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7025 This is an overloaded intrinsic. You can use llvm.memmove on any integer
7026 bit width and for different address space. Not all targets support all
7031 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7032 i32 <len>, i32 <align>, i1 <isvolatile>)
7033 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7034 i64 <len>, i32 <align>, i1 <isvolatile>)
7039 The '``llvm.memmove.*``' intrinsics move a block of memory from the
7040 source location to the destination location. It is similar to the
7041 '``llvm.memcpy``' intrinsic but allows the two memory locations to
7044 Note that, unlike the standard libc function, the ``llvm.memmove.*``
7045 intrinsics do not return a value, takes extra alignment/isvolatile
7046 arguments and the pointers can be in specified address spaces.
7051 The first argument is a pointer to the destination, the second is a
7052 pointer to the source. The third argument is an integer argument
7053 specifying the number of bytes to copy, the fourth argument is the
7054 alignment of the source and destination locations, and the fifth is a
7055 boolean indicating a volatile access.
7057 If the call to this intrinsic has an alignment value that is not 0 or 1,
7058 then the caller guarantees that the source and destination pointers are
7059 aligned to that boundary.
7061 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
7062 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
7063 not very cleanly specified and it is unwise to depend on it.
7068 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
7069 source location to the destination location, which may overlap. It
7070 copies "len" bytes of memory over. If the argument is known to be
7071 aligned to some boundary, this can be specified as the fourth argument,
7072 otherwise it should be set to 0 or 1 (both meaning no alignment).
7074 '``llvm.memset.*``' Intrinsics
7075 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7080 This is an overloaded intrinsic. You can use llvm.memset on any integer
7081 bit width and for different address spaces. However, not all targets
7082 support all bit widths.
7086 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
7087 i32 <len>, i32 <align>, i1 <isvolatile>)
7088 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
7089 i64 <len>, i32 <align>, i1 <isvolatile>)
7094 The '``llvm.memset.*``' intrinsics fill a block of memory with a
7095 particular byte value.
7097 Note that, unlike the standard libc function, the ``llvm.memset``
7098 intrinsic does not return a value and takes extra alignment/volatile
7099 arguments. Also, the destination can be in an arbitrary address space.
7104 The first argument is a pointer to the destination to fill, the second
7105 is the byte value with which to fill it, the third argument is an
7106 integer argument specifying the number of bytes to fill, and the fourth
7107 argument is the known alignment of the destination location.
7109 If the call to this intrinsic has an alignment value that is not 0 or 1,
7110 then the caller guarantees that the destination pointer is aligned to
7113 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
7114 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7115 very cleanly specified and it is unwise to depend on it.
7120 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
7121 at the destination location. If the argument is known to be aligned to
7122 some boundary, this can be specified as the fourth argument, otherwise
7123 it should be set to 0 or 1 (both meaning no alignment).
7125 '``llvm.sqrt.*``' Intrinsic
7126 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7131 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
7132 floating point or vector of floating point type. Not all targets support
7137 declare float @llvm.sqrt.f32(float %Val)
7138 declare double @llvm.sqrt.f64(double %Val)
7139 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
7140 declare fp128 @llvm.sqrt.f128(fp128 %Val)
7141 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
7146 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
7147 returning the same value as the libm '``sqrt``' functions would. Unlike
7148 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
7149 negative numbers other than -0.0 (which allows for better optimization,
7150 because there is no need to worry about errno being set).
7151 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
7156 The argument and return value are floating point numbers of the same
7162 This function returns the sqrt of the specified operand if it is a
7163 nonnegative floating point number.
7165 '``llvm.powi.*``' Intrinsic
7166 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7171 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
7172 floating point or vector of floating point type. Not all targets support
7177 declare float @llvm.powi.f32(float %Val, i32 %power)
7178 declare double @llvm.powi.f64(double %Val, i32 %power)
7179 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7180 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7181 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
7186 The '``llvm.powi.*``' intrinsics return the first operand raised to the
7187 specified (positive or negative) power. The order of evaluation of
7188 multiplications is not defined. When a vector of floating point type is
7189 used, the second argument remains a scalar integer value.
7194 The second argument is an integer power, and the first is a value to
7195 raise to that power.
7200 This function returns the first value raised to the second power with an
7201 unspecified sequence of rounding operations.
7203 '``llvm.sin.*``' Intrinsic
7204 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7209 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
7210 floating point or vector of floating point type. Not all targets support
7215 declare float @llvm.sin.f32(float %Val)
7216 declare double @llvm.sin.f64(double %Val)
7217 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7218 declare fp128 @llvm.sin.f128(fp128 %Val)
7219 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7224 The '``llvm.sin.*``' intrinsics return the sine of the operand.
7229 The argument and return value are floating point numbers of the same
7235 This function returns the sine of the specified operand, returning the
7236 same values as the libm ``sin`` functions would, and handles error
7237 conditions in the same way.
7239 '``llvm.cos.*``' Intrinsic
7240 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7245 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
7246 floating point or vector of floating point type. Not all targets support
7251 declare float @llvm.cos.f32(float %Val)
7252 declare double @llvm.cos.f64(double %Val)
7253 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7254 declare fp128 @llvm.cos.f128(fp128 %Val)
7255 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7260 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
7265 The argument and return value are floating point numbers of the same
7271 This function returns the cosine of the specified operand, returning the
7272 same values as the libm ``cos`` functions would, and handles error
7273 conditions in the same way.
7275 '``llvm.pow.*``' Intrinsic
7276 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7281 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
7282 floating point or vector of floating point type. Not all targets support
7287 declare float @llvm.pow.f32(float %Val, float %Power)
7288 declare double @llvm.pow.f64(double %Val, double %Power)
7289 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7290 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7291 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7296 The '``llvm.pow.*``' intrinsics return the first operand raised to the
7297 specified (positive or negative) power.
7302 The second argument is a floating point power, and the first is a value
7303 to raise to that power.
7308 This function returns the first value raised to the second power,
7309 returning the same values as the libm ``pow`` functions would, and
7310 handles error conditions in the same way.
7312 '``llvm.exp.*``' Intrinsic
7313 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7318 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
7319 floating point or vector of floating point type. Not all targets support
7324 declare float @llvm.exp.f32(float %Val)
7325 declare double @llvm.exp.f64(double %Val)
7326 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7327 declare fp128 @llvm.exp.f128(fp128 %Val)
7328 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7333 The '``llvm.exp.*``' intrinsics perform the exp function.
7338 The argument and return value are floating point numbers of the same
7344 This function returns the same values as the libm ``exp`` functions
7345 would, and handles error conditions in the same way.
7347 '``llvm.exp2.*``' Intrinsic
7348 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7353 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
7354 floating point or vector of floating point type. Not all targets support
7359 declare float @llvm.exp2.f32(float %Val)
7360 declare double @llvm.exp2.f64(double %Val)
7361 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
7362 declare fp128 @llvm.exp2.f128(fp128 %Val)
7363 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
7368 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
7373 The argument and return value are floating point numbers of the same
7379 This function returns the same values as the libm ``exp2`` functions
7380 would, and handles error conditions in the same way.
7382 '``llvm.log.*``' Intrinsic
7383 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7388 This is an overloaded intrinsic. You can use ``llvm.log`` on any
7389 floating point or vector of floating point type. Not all targets support
7394 declare float @llvm.log.f32(float %Val)
7395 declare double @llvm.log.f64(double %Val)
7396 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7397 declare fp128 @llvm.log.f128(fp128 %Val)
7398 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7403 The '``llvm.log.*``' intrinsics perform the log function.
7408 The argument and return value are floating point numbers of the same
7414 This function returns the same values as the libm ``log`` functions
7415 would, and handles error conditions in the same way.
7417 '``llvm.log10.*``' Intrinsic
7418 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7423 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
7424 floating point or vector of floating point type. Not all targets support
7429 declare float @llvm.log10.f32(float %Val)
7430 declare double @llvm.log10.f64(double %Val)
7431 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
7432 declare fp128 @llvm.log10.f128(fp128 %Val)
7433 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
7438 The '``llvm.log10.*``' intrinsics perform the log10 function.
7443 The argument and return value are floating point numbers of the same
7449 This function returns the same values as the libm ``log10`` functions
7450 would, and handles error conditions in the same way.
7452 '``llvm.log2.*``' Intrinsic
7453 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7458 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
7459 floating point or vector of floating point type. Not all targets support
7464 declare float @llvm.log2.f32(float %Val)
7465 declare double @llvm.log2.f64(double %Val)
7466 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
7467 declare fp128 @llvm.log2.f128(fp128 %Val)
7468 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
7473 The '``llvm.log2.*``' intrinsics perform the log2 function.
7478 The argument and return value are floating point numbers of the same
7484 This function returns the same values as the libm ``log2`` functions
7485 would, and handles error conditions in the same way.
7487 '``llvm.fma.*``' Intrinsic
7488 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7493 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
7494 floating point or vector of floating point type. Not all targets support
7499 declare float @llvm.fma.f32(float %a, float %b, float %c)
7500 declare double @llvm.fma.f64(double %a, double %b, double %c)
7501 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7502 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7503 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7508 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
7514 The argument and return value are floating point numbers of the same
7520 This function returns the same values as the libm ``fma`` functions
7521 would, and does not set errno.
7523 '``llvm.fabs.*``' Intrinsic
7524 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7529 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
7530 floating point or vector of floating point type. Not all targets support
7535 declare float @llvm.fabs.f32(float %Val)
7536 declare double @llvm.fabs.f64(double %Val)
7537 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
7538 declare fp128 @llvm.fabs.f128(fp128 %Val)
7539 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
7544 The '``llvm.fabs.*``' intrinsics return the absolute value of the
7550 The argument and return value are floating point numbers of the same
7556 This function returns the same values as the libm ``fabs`` functions
7557 would, and handles error conditions in the same way.
7559 '``llvm.copysign.*``' Intrinsic
7560 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7565 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
7566 floating point or vector of floating point type. Not all targets support
7571 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
7572 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
7573 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
7574 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
7575 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
7580 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
7581 first operand and the sign of the second operand.
7586 The arguments and return value are floating point numbers of the same
7592 This function returns the same values as the libm ``copysign``
7593 functions would, and handles error conditions in the same way.
7595 '``llvm.floor.*``' Intrinsic
7596 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7601 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
7602 floating point or vector of floating point type. Not all targets support
7607 declare float @llvm.floor.f32(float %Val)
7608 declare double @llvm.floor.f64(double %Val)
7609 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
7610 declare fp128 @llvm.floor.f128(fp128 %Val)
7611 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
7616 The '``llvm.floor.*``' intrinsics return the floor of the operand.
7621 The argument and return value are floating point numbers of the same
7627 This function returns the same values as the libm ``floor`` functions
7628 would, and handles error conditions in the same way.
7630 '``llvm.ceil.*``' Intrinsic
7631 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7636 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
7637 floating point or vector of floating point type. Not all targets support
7642 declare float @llvm.ceil.f32(float %Val)
7643 declare double @llvm.ceil.f64(double %Val)
7644 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
7645 declare fp128 @llvm.ceil.f128(fp128 %Val)
7646 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
7651 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
7656 The argument and return value are floating point numbers of the same
7662 This function returns the same values as the libm ``ceil`` functions
7663 would, and handles error conditions in the same way.
7665 '``llvm.trunc.*``' Intrinsic
7666 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7671 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
7672 floating point or vector of floating point type. Not all targets support
7677 declare float @llvm.trunc.f32(float %Val)
7678 declare double @llvm.trunc.f64(double %Val)
7679 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
7680 declare fp128 @llvm.trunc.f128(fp128 %Val)
7681 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
7686 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
7687 nearest integer not larger in magnitude than the operand.
7692 The argument and return value are floating point numbers of the same
7698 This function returns the same values as the libm ``trunc`` functions
7699 would, and handles error conditions in the same way.
7701 '``llvm.rint.*``' Intrinsic
7702 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7707 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
7708 floating point or vector of floating point type. Not all targets support
7713 declare float @llvm.rint.f32(float %Val)
7714 declare double @llvm.rint.f64(double %Val)
7715 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
7716 declare fp128 @llvm.rint.f128(fp128 %Val)
7717 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
7722 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
7723 nearest integer. It may raise an inexact floating-point exception if the
7724 operand isn't an integer.
7729 The argument and return value are floating point numbers of the same
7735 This function returns the same values as the libm ``rint`` functions
7736 would, and handles error conditions in the same way.
7738 '``llvm.nearbyint.*``' Intrinsic
7739 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7744 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
7745 floating point or vector of floating point type. Not all targets support
7750 declare float @llvm.nearbyint.f32(float %Val)
7751 declare double @llvm.nearbyint.f64(double %Val)
7752 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
7753 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
7754 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
7759 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
7765 The argument and return value are floating point numbers of the same
7771 This function returns the same values as the libm ``nearbyint``
7772 functions would, and handles error conditions in the same way.
7774 '``llvm.round.*``' Intrinsic
7775 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7780 This is an overloaded intrinsic. You can use ``llvm.round`` on any
7781 floating point or vector of floating point type. Not all targets support
7786 declare float @llvm.round.f32(float %Val)
7787 declare double @llvm.round.f64(double %Val)
7788 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
7789 declare fp128 @llvm.round.f128(fp128 %Val)
7790 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
7795 The '``llvm.round.*``' intrinsics returns the operand rounded to the
7801 The argument and return value are floating point numbers of the same
7807 This function returns the same values as the libm ``round``
7808 functions would, and handles error conditions in the same way.
7810 Bit Manipulation Intrinsics
7811 ---------------------------
7813 LLVM provides intrinsics for a few important bit manipulation
7814 operations. These allow efficient code generation for some algorithms.
7816 '``llvm.bswap.*``' Intrinsics
7817 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7822 This is an overloaded intrinsic function. You can use bswap on any
7823 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
7827 declare i16 @llvm.bswap.i16(i16 <id>)
7828 declare i32 @llvm.bswap.i32(i32 <id>)
7829 declare i64 @llvm.bswap.i64(i64 <id>)
7834 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
7835 values with an even number of bytes (positive multiple of 16 bits).
7836 These are useful for performing operations on data that is not in the
7837 target's native byte order.
7842 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
7843 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
7844 intrinsic returns an i32 value that has the four bytes of the input i32
7845 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
7846 returned i32 will have its bytes in 3, 2, 1, 0 order. The
7847 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
7848 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
7851 '``llvm.ctpop.*``' Intrinsic
7852 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7857 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
7858 bit width, or on any vector with integer elements. Not all targets
7859 support all bit widths or vector types, however.
7863 declare i8 @llvm.ctpop.i8(i8 <src>)
7864 declare i16 @llvm.ctpop.i16(i16 <src>)
7865 declare i32 @llvm.ctpop.i32(i32 <src>)
7866 declare i64 @llvm.ctpop.i64(i64 <src>)
7867 declare i256 @llvm.ctpop.i256(i256 <src>)
7868 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
7873 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
7879 The only argument is the value to be counted. The argument may be of any
7880 integer type, or a vector with integer elements. The return type must
7881 match the argument type.
7886 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
7887 each element of a vector.
7889 '``llvm.ctlz.*``' Intrinsic
7890 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7895 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
7896 integer bit width, or any vector whose elements are integers. Not all
7897 targets support all bit widths or vector types, however.
7901 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
7902 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
7903 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
7904 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
7905 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
7906 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
7911 The '``llvm.ctlz``' family of intrinsic functions counts the number of
7912 leading zeros in a variable.
7917 The first argument is the value to be counted. This argument may be of
7918 any integer type, or a vectory with integer element type. The return
7919 type must match the first argument type.
7921 The second argument must be a constant and is a flag to indicate whether
7922 the intrinsic should ensure that a zero as the first argument produces a
7923 defined result. Historically some architectures did not provide a
7924 defined result for zero values as efficiently, and many algorithms are
7925 now predicated on avoiding zero-value inputs.
7930 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
7931 zeros in a variable, or within each element of the vector. If
7932 ``src == 0`` then the result is the size in bits of the type of ``src``
7933 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
7934 ``llvm.ctlz(i32 2) = 30``.
7936 '``llvm.cttz.*``' Intrinsic
7937 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7942 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
7943 integer bit width, or any vector of integer elements. Not all targets
7944 support all bit widths or vector types, however.
7948 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
7949 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
7950 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
7951 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
7952 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
7953 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
7958 The '``llvm.cttz``' family of intrinsic functions counts the number of
7964 The first argument is the value to be counted. This argument may be of
7965 any integer type, or a vectory with integer element type. The return
7966 type must match the first argument type.
7968 The second argument must be a constant and is a flag to indicate whether
7969 the intrinsic should ensure that a zero as the first argument produces a
7970 defined result. Historically some architectures did not provide a
7971 defined result for zero values as efficiently, and many algorithms are
7972 now predicated on avoiding zero-value inputs.
7977 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
7978 zeros in a variable, or within each element of a vector. If ``src == 0``
7979 then the result is the size in bits of the type of ``src`` if
7980 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
7981 ``llvm.cttz(2) = 1``.
7983 Arithmetic with Overflow Intrinsics
7984 -----------------------------------
7986 LLVM provides intrinsics for some arithmetic with overflow operations.
7988 '``llvm.sadd.with.overflow.*``' Intrinsics
7989 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7994 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
7995 on any integer bit width.
7999 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
8000 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8001 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
8006 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8007 a signed addition of the two arguments, and indicate whether an overflow
8008 occurred during the signed summation.
8013 The arguments (%a and %b) and the first element of the result structure
8014 may be of integer types of any bit width, but they must have the same
8015 bit width. The second element of the result structure must be of type
8016 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8022 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8023 a signed addition of the two variables. They return a structure --- the
8024 first element of which is the signed summation, and the second element
8025 of which is a bit specifying if the signed summation resulted in an
8031 .. code-block:: llvm
8033 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8034 %sum = extractvalue {i32, i1} %res, 0
8035 %obit = extractvalue {i32, i1} %res, 1
8036 br i1 %obit, label %overflow, label %normal
8038 '``llvm.uadd.with.overflow.*``' Intrinsics
8039 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8044 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
8045 on any integer bit width.
8049 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
8050 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8051 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
8056 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8057 an unsigned addition of the two arguments, and indicate whether a carry
8058 occurred during the unsigned summation.
8063 The arguments (%a and %b) and the first element of the result structure
8064 may be of integer types of any bit width, but they must have the same
8065 bit width. The second element of the result structure must be of type
8066 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8072 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8073 an unsigned addition of the two arguments. They return a structure --- the
8074 first element of which is the sum, and the second element of which is a
8075 bit specifying if the unsigned summation resulted in a carry.
8080 .. code-block:: llvm
8082 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8083 %sum = extractvalue {i32, i1} %res, 0
8084 %obit = extractvalue {i32, i1} %res, 1
8085 br i1 %obit, label %carry, label %normal
8087 '``llvm.ssub.with.overflow.*``' Intrinsics
8088 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8093 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
8094 on any integer bit width.
8098 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
8099 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8100 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
8105 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8106 a signed subtraction of the two arguments, and indicate whether an
8107 overflow occurred during the signed subtraction.
8112 The arguments (%a and %b) and the first element of the result structure
8113 may be of integer types of any bit width, but they must have the same
8114 bit width. The second element of the result structure must be of type
8115 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8121 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8122 a signed subtraction of the two arguments. They return a structure --- the
8123 first element of which is the subtraction, and the second element of
8124 which is a bit specifying if the signed subtraction resulted in an
8130 .. code-block:: llvm
8132 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8133 %sum = extractvalue {i32, i1} %res, 0
8134 %obit = extractvalue {i32, i1} %res, 1
8135 br i1 %obit, label %overflow, label %normal
8137 '``llvm.usub.with.overflow.*``' Intrinsics
8138 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8143 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
8144 on any integer bit width.
8148 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
8149 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8150 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
8155 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8156 an unsigned subtraction of the two arguments, and indicate whether an
8157 overflow occurred during the unsigned subtraction.
8162 The arguments (%a and %b) and the first element of the result structure
8163 may be of integer types of any bit width, but they must have the same
8164 bit width. The second element of the result structure must be of type
8165 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8171 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8172 an unsigned subtraction of the two arguments. They return a structure ---
8173 the first element of which is the subtraction, and the second element of
8174 which is a bit specifying if the unsigned subtraction resulted in an
8180 .. code-block:: llvm
8182 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8183 %sum = extractvalue {i32, i1} %res, 0
8184 %obit = extractvalue {i32, i1} %res, 1
8185 br i1 %obit, label %overflow, label %normal
8187 '``llvm.smul.with.overflow.*``' Intrinsics
8188 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8193 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
8194 on any integer bit width.
8198 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
8199 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8200 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
8205 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8206 a signed multiplication of the two arguments, and indicate whether an
8207 overflow occurred during the signed multiplication.
8212 The arguments (%a and %b) and the first element of the result structure
8213 may be of integer types of any bit width, but they must have the same
8214 bit width. The second element of the result structure must be of type
8215 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8221 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8222 a signed multiplication of the two arguments. They return a structure ---
8223 the first element of which is the multiplication, and the second element
8224 of which is a bit specifying if the signed multiplication resulted in an
8230 .. code-block:: llvm
8232 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8233 %sum = extractvalue {i32, i1} %res, 0
8234 %obit = extractvalue {i32, i1} %res, 1
8235 br i1 %obit, label %overflow, label %normal
8237 '``llvm.umul.with.overflow.*``' Intrinsics
8238 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8243 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
8244 on any integer bit width.
8248 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
8249 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8250 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
8255 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
8256 a unsigned multiplication of the two arguments, and indicate whether an
8257 overflow occurred during the unsigned multiplication.
8262 The arguments (%a and %b) and the first element of the result structure
8263 may be of integer types of any bit width, but they must have the same
8264 bit width. The second element of the result structure must be of type
8265 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8271 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
8272 an unsigned multiplication of the two arguments. They return a structure ---
8273 the first element of which is the multiplication, and the second
8274 element of which is a bit specifying if the unsigned multiplication
8275 resulted in an overflow.
8280 .. code-block:: llvm
8282 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8283 %sum = extractvalue {i32, i1} %res, 0
8284 %obit = extractvalue {i32, i1} %res, 1
8285 br i1 %obit, label %overflow, label %normal
8287 Specialised Arithmetic Intrinsics
8288 ---------------------------------
8290 '``llvm.fmuladd.*``' Intrinsic
8291 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8298 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
8299 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
8304 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
8305 expressions that can be fused if the code generator determines that (a) the
8306 target instruction set has support for a fused operation, and (b) that the
8307 fused operation is more efficient than the equivalent, separate pair of mul
8308 and add instructions.
8313 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
8314 multiplicands, a and b, and an addend c.
8323 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
8325 is equivalent to the expression a \* b + c, except that rounding will
8326 not be performed between the multiplication and addition steps if the
8327 code generator fuses the operations. Fusion is not guaranteed, even if
8328 the target platform supports it. If a fused multiply-add is required the
8329 corresponding llvm.fma.\* intrinsic function should be used
8330 instead. This never sets errno, just as '``llvm.fma.*``'.
8335 .. code-block:: llvm
8337 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields {float}:r2 = (a * b) + c
8339 Half Precision Floating Point Intrinsics
8340 ----------------------------------------
8342 For most target platforms, half precision floating point is a
8343 storage-only format. This means that it is a dense encoding (in memory)
8344 but does not support computation in the format.
8346 This means that code must first load the half-precision floating point
8347 value as an i16, then convert it to float with
8348 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
8349 then be performed on the float value (including extending to double
8350 etc). To store the value back to memory, it is first converted to float
8351 if needed, then converted to i16 with
8352 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
8355 .. _int_convert_to_fp16:
8357 '``llvm.convert.to.fp16``' Intrinsic
8358 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8365 declare i16 @llvm.convert.to.fp16(f32 %a)
8370 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion
8371 from single precision floating point format to half precision floating
8377 The intrinsic function contains single argument - the value to be
8383 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion
8384 from single precision floating point format to half precision floating
8385 point format. The return value is an ``i16`` which contains the
8391 .. code-block:: llvm
8393 %res = call i16 @llvm.convert.to.fp16(f32 %a)
8394 store i16 %res, i16* @x, align 2
8396 .. _int_convert_from_fp16:
8398 '``llvm.convert.from.fp16``' Intrinsic
8399 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8406 declare f32 @llvm.convert.from.fp16(i16 %a)
8411 The '``llvm.convert.from.fp16``' intrinsic function performs a
8412 conversion from half precision floating point format to single precision
8413 floating point format.
8418 The intrinsic function contains single argument - the value to be
8424 The '``llvm.convert.from.fp16``' intrinsic function performs a
8425 conversion from half single precision floating point format to single
8426 precision floating point format. The input half-float value is
8427 represented by an ``i16`` value.
8432 .. code-block:: llvm
8434 %a = load i16* @x, align 2
8435 %res = call f32 @llvm.convert.from.fp16(i16 %a)
8440 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
8441 prefix), are described in the `LLVM Source Level
8442 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
8445 Exception Handling Intrinsics
8446 -----------------------------
8448 The LLVM exception handling intrinsics (which all start with
8449 ``llvm.eh.`` prefix), are described in the `LLVM Exception
8450 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
8454 Trampoline Intrinsics
8455 ---------------------
8457 These intrinsics make it possible to excise one parameter, marked with
8458 the :ref:`nest <nest>` attribute, from a function. The result is a
8459 callable function pointer lacking the nest parameter - the caller does
8460 not need to provide a value for it. Instead, the value to use is stored
8461 in advance in a "trampoline", a block of memory usually allocated on the
8462 stack, which also contains code to splice the nest value into the
8463 argument list. This is used to implement the GCC nested function address
8466 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
8467 then the resulting function pointer has signature ``i32 (i32, i32)*``.
8468 It can be created as follows:
8470 .. code-block:: llvm
8472 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
8473 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
8474 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
8475 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
8476 %fp = bitcast i8* %p to i32 (i32, i32)*
8478 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
8479 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
8483 '``llvm.init.trampoline``' Intrinsic
8484 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8491 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
8496 This fills the memory pointed to by ``tramp`` with executable code,
8497 turning it into a trampoline.
8502 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
8503 pointers. The ``tramp`` argument must point to a sufficiently large and
8504 sufficiently aligned block of memory; this memory is written to by the
8505 intrinsic. Note that the size and the alignment are target-specific -
8506 LLVM currently provides no portable way of determining them, so a
8507 front-end that generates this intrinsic needs to have some
8508 target-specific knowledge. The ``func`` argument must hold a function
8509 bitcast to an ``i8*``.
8514 The block of memory pointed to by ``tramp`` is filled with target
8515 dependent code, turning it into a function. Then ``tramp`` needs to be
8516 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
8517 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
8518 function's signature is the same as that of ``func`` with any arguments
8519 marked with the ``nest`` attribute removed. At most one such ``nest``
8520 argument is allowed, and it must be of pointer type. Calling the new
8521 function is equivalent to calling ``func`` with the same argument list,
8522 but with ``nval`` used for the missing ``nest`` argument. If, after
8523 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
8524 modified, then the effect of any later call to the returned function
8525 pointer is undefined.
8529 '``llvm.adjust.trampoline``' Intrinsic
8530 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8537 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
8542 This performs any required machine-specific adjustment to the address of
8543 a trampoline (passed as ``tramp``).
8548 ``tramp`` must point to a block of memory which already has trampoline
8549 code filled in by a previous call to
8550 :ref:`llvm.init.trampoline <int_it>`.
8555 On some architectures the address of the code to be executed needs to be
8556 different to the address where the trampoline is actually stored. This
8557 intrinsic returns the executable address corresponding to ``tramp``
8558 after performing the required machine specific adjustments. The pointer
8559 returned can then be :ref:`bitcast and executed <int_trampoline>`.
8564 This class of intrinsics exists to information about the lifetime of
8565 memory objects and ranges where variables are immutable.
8569 '``llvm.lifetime.start``' Intrinsic
8570 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8577 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
8582 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
8588 The first argument is a constant integer representing the size of the
8589 object, or -1 if it is variable sized. The second argument is a pointer
8595 This intrinsic indicates that before this point in the code, the value
8596 of the memory pointed to by ``ptr`` is dead. This means that it is known
8597 to never be used and has an undefined value. A load from the pointer
8598 that precedes this intrinsic can be replaced with ``'undef'``.
8602 '``llvm.lifetime.end``' Intrinsic
8603 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8610 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
8615 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
8621 The first argument is a constant integer representing the size of the
8622 object, or -1 if it is variable sized. The second argument is a pointer
8628 This intrinsic indicates that after this point in the code, the value of
8629 the memory pointed to by ``ptr`` is dead. This means that it is known to
8630 never be used and has an undefined value. Any stores into the memory
8631 object following this intrinsic may be removed as dead.
8633 '``llvm.invariant.start``' Intrinsic
8634 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8641 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
8646 The '``llvm.invariant.start``' intrinsic specifies that the contents of
8647 a memory object will not change.
8652 The first argument is a constant integer representing the size of the
8653 object, or -1 if it is variable sized. The second argument is a pointer
8659 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
8660 the return value, the referenced memory location is constant and
8663 '``llvm.invariant.end``' Intrinsic
8664 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8671 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
8676 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
8677 memory object are mutable.
8682 The first argument is the matching ``llvm.invariant.start`` intrinsic.
8683 The second argument is a constant integer representing the size of the
8684 object, or -1 if it is variable sized and the third argument is a
8685 pointer to the object.
8690 This intrinsic indicates that the memory is mutable again.
8695 This class of intrinsics is designed to be generic and has no specific
8698 '``llvm.var.annotation``' Intrinsic
8699 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8706 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
8711 The '``llvm.var.annotation``' intrinsic.
8716 The first argument is a pointer to a value, the second is a pointer to a
8717 global string, the third is a pointer to a global string which is the
8718 source file name, and the last argument is the line number.
8723 This intrinsic allows annotation of local variables with arbitrary
8724 strings. This can be useful for special purpose optimizations that want
8725 to look for these annotations. These have no other defined use; they are
8726 ignored by code generation and optimization.
8728 '``llvm.ptr.annotation.*``' Intrinsic
8729 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8734 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
8735 pointer to an integer of any width. *NOTE* you must specify an address space for
8736 the pointer. The identifier for the default address space is the integer
8741 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
8742 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
8743 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
8744 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
8745 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
8750 The '``llvm.ptr.annotation``' intrinsic.
8755 The first argument is a pointer to an integer value of arbitrary bitwidth
8756 (result of some expression), the second is a pointer to a global string, the
8757 third is a pointer to a global string which is the source file name, and the
8758 last argument is the line number. It returns the value of the first argument.
8763 This intrinsic allows annotation of a pointer to an integer with arbitrary
8764 strings. This can be useful for special purpose optimizations that want to look
8765 for these annotations. These have no other defined use; they are ignored by code
8766 generation and optimization.
8768 '``llvm.annotation.*``' Intrinsic
8769 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8774 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
8775 any integer bit width.
8779 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
8780 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
8781 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
8782 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
8783 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
8788 The '``llvm.annotation``' intrinsic.
8793 The first argument is an integer value (result of some expression), the
8794 second is a pointer to a global string, the third is a pointer to a
8795 global string which is the source file name, and the last argument is
8796 the line number. It returns the value of the first argument.
8801 This intrinsic allows annotations to be put on arbitrary expressions
8802 with arbitrary strings. This can be useful for special purpose
8803 optimizations that want to look for these annotations. These have no
8804 other defined use; they are ignored by code generation and optimization.
8806 '``llvm.trap``' Intrinsic
8807 ^^^^^^^^^^^^^^^^^^^^^^^^^
8814 declare void @llvm.trap() noreturn nounwind
8819 The '``llvm.trap``' intrinsic.
8829 This intrinsic is lowered to the target dependent trap instruction. If
8830 the target does not have a trap instruction, this intrinsic will be
8831 lowered to a call of the ``abort()`` function.
8833 '``llvm.debugtrap``' Intrinsic
8834 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8841 declare void @llvm.debugtrap() nounwind
8846 The '``llvm.debugtrap``' intrinsic.
8856 This intrinsic is lowered to code which is intended to cause an
8857 execution trap with the intention of requesting the attention of a
8860 '``llvm.stackprotector``' Intrinsic
8861 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8868 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
8873 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
8874 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
8875 is placed on the stack before local variables.
8880 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
8881 The first argument is the value loaded from the stack guard
8882 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
8883 enough space to hold the value of the guard.
8888 This intrinsic causes the prologue/epilogue inserter to force the position of
8889 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
8890 to ensure that if a local variable on the stack is overwritten, it will destroy
8891 the value of the guard. When the function exits, the guard on the stack is
8892 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
8893 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
8894 calling the ``__stack_chk_fail()`` function.
8896 '``llvm.stackprotectorcheck``' Intrinsic
8897 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8904 declare void @llvm.stackprotectorcheck(i8** <guard>)
8909 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
8910 created stack protector and if they are not equal calls the
8911 ``__stack_chk_fail()`` function.
8916 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
8917 the variable ``@__stack_chk_guard``.
8922 This intrinsic is provided to perform the stack protector check by comparing
8923 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
8924 values do not match call the ``__stack_chk_fail()`` function.
8926 The reason to provide this as an IR level intrinsic instead of implementing it
8927 via other IR operations is that in order to perform this operation at the IR
8928 level without an intrinsic, one would need to create additional basic blocks to
8929 handle the success/failure cases. This makes it difficult to stop the stack
8930 protector check from disrupting sibling tail calls in Codegen. With this
8931 intrinsic, we are able to generate the stack protector basic blocks late in
8932 codegen after the tail call decision has occurred.
8934 '``llvm.objectsize``' Intrinsic
8935 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8942 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
8943 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
8948 The ``llvm.objectsize`` intrinsic is designed to provide information to
8949 the optimizers to determine at compile time whether a) an operation
8950 (like memcpy) will overflow a buffer that corresponds to an object, or
8951 b) that a runtime check for overflow isn't necessary. An object in this
8952 context means an allocation of a specific class, structure, array, or
8958 The ``llvm.objectsize`` intrinsic takes two arguments. The first
8959 argument is a pointer to or into the ``object``. The second argument is
8960 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
8961 or -1 (if false) when the object size is unknown. The second argument
8962 only accepts constants.
8967 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
8968 the size of the object concerned. If the size cannot be determined at
8969 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
8970 on the ``min`` argument).
8972 '``llvm.expect``' Intrinsic
8973 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8978 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
8983 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
8984 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
8985 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
8990 The ``llvm.expect`` intrinsic provides information about expected (the
8991 most probable) value of ``val``, which can be used by optimizers.
8996 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
8997 a value. The second argument is an expected value, this needs to be a
8998 constant value, variables are not allowed.
9003 This intrinsic is lowered to the ``val``.
9005 '``llvm.donothing``' Intrinsic
9006 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9013 declare void @llvm.donothing() nounwind readnone
9018 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's the
9019 only intrinsic that can be called with an invoke instruction.
9029 This intrinsic does nothing, and it's removed by optimizers and ignored
9032 Stack Map Intrinsics
9033 --------------------
9035 LLVM provides experimental intrinsics to support runtime patching
9036 mechanisms commonly desired in dynamic language JITs. These intrinsics
9037 are described in :doc:`StackMaps`.