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 that 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. The ``"\01"`` prefix
83 can be used on global variables to suppress mangling.
84 #. Unnamed values are represented as an unsigned numeric value with
85 their prefix. For example, ``%12``, ``@2``, ``%44``.
86 #. Constants, which are described in the section Constants_ below.
88 LLVM requires that values start with a prefix for two reasons: Compilers
89 don't need to worry about name clashes with reserved words, and the set
90 of reserved words may be expanded in the future without penalty.
91 Additionally, unnamed identifiers allow a compiler to quickly come up
92 with a temporary variable without having to avoid symbol table
95 Reserved words in LLVM are very similar to reserved words in other
96 languages. There are keywords for different opcodes ('``add``',
97 '``bitcast``', '``ret``', etc...), for primitive type names ('``void``',
98 '``i32``', etc...), and others. These reserved words cannot conflict
99 with variable names, because none of them start with a prefix character
100 (``'%'`` or ``'@'``).
102 Here is an example of LLVM code to multiply the integer variable
109 %result = mul i32 %X, 8
111 After strength reduction:
115 %result = shl i32 %X, 3
121 %0 = add i32 %X, %X ; yields i32:%0
122 %1 = add i32 %0, %0 ; yields i32:%1
123 %result = add i32 %1, %1
125 This last way of multiplying ``%X`` by 8 illustrates several important
126 lexical features of LLVM:
128 #. Comments are delimited with a '``;``' and go until the end of line.
129 #. Unnamed temporaries are created when the result of a computation is
130 not assigned to a named value.
131 #. Unnamed temporaries are numbered sequentially (using a per-function
132 incrementing counter, starting with 0). Note that basic blocks and unnamed
133 function parameters are included in this numbering. For example, if the
134 entry basic block is not given a label name and all function parameters are
135 named, then it will get number 0.
137 It also shows a convention that we follow in this document. When
138 demonstrating instructions, we will follow an instruction with a comment
139 that defines the type and name of value produced.
147 LLVM programs are composed of ``Module``'s, each of which is a
148 translation unit of the input programs. Each module consists of
149 functions, global variables, and symbol table entries. Modules may be
150 combined together with the LLVM linker, which merges function (and
151 global variable) definitions, resolves forward declarations, and merges
152 symbol table entries. Here is an example of the "hello world" module:
156 ; Declare the string constant as a global constant.
157 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
159 ; External declaration of the puts function
160 declare i32 @puts(i8* nocapture) nounwind
162 ; Definition of main function
163 define i32 @main() { ; i32()*
164 ; Convert [13 x i8]* to i8 *...
165 %cast210 = getelementptr [13 x i8]* @.str, i64 0, i64 0
167 ; Call puts function to write out the string to stdout.
168 call i32 @puts(i8* %cast210)
173 !0 = metadata !{i32 42, null, metadata !"string"}
176 This example is made up of a :ref:`global variable <globalvars>` named
177 "``.str``", an external declaration of the "``puts``" function, a
178 :ref:`function definition <functionstructure>` for "``main``" and
179 :ref:`named metadata <namedmetadatastructure>` "``foo``".
181 In general, a module is made up of a list of global values (where both
182 functions and global variables are global values). Global values are
183 represented by a pointer to a memory location (in this case, a pointer
184 to an array of char, and a pointer to a function), and have one of the
185 following :ref:`linkage types <linkage>`.
192 All Global Variables and Functions have one of the following types of
196 Global values with "``private``" linkage are only directly
197 accessible by objects in the current module. In particular, linking
198 code into a module with an private global value may cause the
199 private to be renamed as necessary to avoid collisions. Because the
200 symbol is private to the module, all references can be updated. This
201 doesn't show up in any symbol table in the object file.
203 Similar to private, but the value shows as a local symbol
204 (``STB_LOCAL`` in the case of ELF) in the object file. This
205 corresponds to the notion of the '``static``' keyword in C.
206 ``available_externally``
207 Globals with "``available_externally``" linkage are never emitted
208 into the object file corresponding to the LLVM module. They exist to
209 allow inlining and other optimizations to take place given knowledge
210 of the definition of the global, which is known to be somewhere
211 outside the module. Globals with ``available_externally`` linkage
212 are allowed to be discarded at will, and are otherwise the same as
213 ``linkonce_odr``. This linkage type is only allowed on definitions,
216 Globals with "``linkonce``" linkage are merged with other globals of
217 the same name when linkage occurs. This can be used to implement
218 some forms of inline functions, templates, or other code which must
219 be generated in each translation unit that uses it, but where the
220 body may be overridden with a more definitive definition later.
221 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
222 that ``linkonce`` linkage does not actually allow the optimizer to
223 inline the body of this function into callers because it doesn't
224 know if this definition of the function is the definitive definition
225 within the program or whether it will be overridden by a stronger
226 definition. To enable inlining and other optimizations, use
227 "``linkonce_odr``" linkage.
229 "``weak``" linkage has the same merging semantics as ``linkonce``
230 linkage, except that unreferenced globals with ``weak`` linkage may
231 not be discarded. This is used for globals that are declared "weak"
234 "``common``" linkage is most similar to "``weak``" linkage, but they
235 are used for tentative definitions in C, such as "``int X;``" at
236 global scope. Symbols with "``common``" linkage are merged in the
237 same way as ``weak symbols``, and they may not be deleted if
238 unreferenced. ``common`` symbols may not have an explicit section,
239 must have a zero initializer, and may not be marked
240 ':ref:`constant <globalvars>`'. Functions and aliases may not have
243 .. _linkage_appending:
246 "``appending``" linkage may only be applied to global variables of
247 pointer to array type. When two global variables with appending
248 linkage are linked together, the two global arrays are appended
249 together. This is the LLVM, typesafe, equivalent of having the
250 system linker append together "sections" with identical names when
253 The semantics of this linkage follow the ELF object file model: the
254 symbol is weak until linked, if not linked, the symbol becomes null
255 instead of being an undefined reference.
256 ``linkonce_odr``, ``weak_odr``
257 Some languages allow differing globals to be merged, such as two
258 functions with different semantics. Other languages, such as
259 ``C++``, ensure that only equivalent globals are ever merged (the
260 "one definition rule" --- "ODR"). Such languages can use the
261 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
262 global will only be merged with equivalent globals. These linkage
263 types are otherwise the same as their non-``odr`` versions.
265 If none of the above identifiers are used, the global is externally
266 visible, meaning that it participates in linkage and can be used to
267 resolve external symbol references.
269 It is illegal for a function *declaration* to have any linkage type
270 other than ``external`` or ``extern_weak``.
277 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
278 :ref:`invokes <i_invoke>` can all have an optional calling convention
279 specified for the call. The calling convention of any pair of dynamic
280 caller/callee must match, or the behavior of the program is undefined.
281 The following calling conventions are supported by LLVM, and more may be
284 "``ccc``" - The C calling convention
285 This calling convention (the default if no other calling convention
286 is specified) matches the target C calling conventions. This calling
287 convention supports varargs function calls and tolerates some
288 mismatch in the declared prototype and implemented declaration of
289 the function (as does normal C).
290 "``fastcc``" - The fast calling convention
291 This calling convention attempts to make calls as fast as possible
292 (e.g. by passing things in registers). This calling convention
293 allows the target to use whatever tricks it wants to produce fast
294 code for the target, without having to conform to an externally
295 specified ABI (Application Binary Interface). `Tail calls can only
296 be optimized when this, the GHC or the HiPE convention is
297 used. <CodeGenerator.html#id80>`_ This calling convention does not
298 support varargs and requires the prototype of all callees to exactly
299 match the prototype of the function definition.
300 "``coldcc``" - The cold calling convention
301 This calling convention attempts to make code in the caller as
302 efficient as possible under the assumption that the call is not
303 commonly executed. As such, these calls often preserve all registers
304 so that the call does not break any live ranges in the caller side.
305 This calling convention does not support varargs and requires the
306 prototype of all callees to exactly match the prototype of the
307 function definition. Furthermore the inliner doesn't consider such function
309 "``cc 10``" - GHC convention
310 This calling convention has been implemented specifically for use by
311 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
312 It passes everything in registers, going to extremes to achieve this
313 by disabling callee save registers. This calling convention should
314 not be used lightly but only for specific situations such as an
315 alternative to the *register pinning* performance technique often
316 used when implementing functional programming languages. At the
317 moment only X86 supports this convention and it has the following
320 - On *X86-32* only supports up to 4 bit type parameters. No
321 floating point types are supported.
322 - On *X86-64* only supports up to 10 bit type parameters and 6
323 floating point parameters.
325 This calling convention supports `tail call
326 optimization <CodeGenerator.html#id80>`_ but requires both the
327 caller and callee are using it.
328 "``cc 11``" - The HiPE calling convention
329 This calling convention has been implemented specifically for use by
330 the `High-Performance Erlang
331 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
332 native code compiler of the `Ericsson's Open Source Erlang/OTP
333 system <http://www.erlang.org/download.shtml>`_. It uses more
334 registers for argument passing than the ordinary C calling
335 convention and defines no callee-saved registers. The calling
336 convention properly supports `tail call
337 optimization <CodeGenerator.html#id80>`_ but requires that both the
338 caller and the callee use it. It uses a *register pinning*
339 mechanism, similar to GHC's convention, for keeping frequently
340 accessed runtime components pinned to specific hardware registers.
341 At the moment only X86 supports this convention (both 32 and 64
343 "``webkit_jscc``" - WebKit's JavaScript calling convention
344 This calling convention has been implemented for `WebKit FTL JIT
345 <https://trac.webkit.org/wiki/FTLJIT>`_. It passes arguments on the
346 stack right to left (as cdecl does), and returns a value in the
347 platform's customary return register.
348 "``anyregcc``" - Dynamic calling convention for code patching
349 This is a special convention that supports patching an arbitrary code
350 sequence in place of a call site. This convention forces the call
351 arguments into registers but allows them to be dynamcially
352 allocated. This can currently only be used with calls to
353 llvm.experimental.patchpoint because only this intrinsic records
354 the location of its arguments in a side table. See :doc:`StackMaps`.
355 "``preserve_mostcc``" - The `PreserveMost` calling convention
356 This calling convention attempts to make the code in the caller as little
357 intrusive as possible. This calling convention behaves identical to the `C`
358 calling convention on how arguments and return values are passed, but it
359 uses a different set of caller/callee-saved registers. This alleviates the
360 burden of saving and recovering a large register set before and after the
361 call in the caller. If the arguments are passed in callee-saved registers,
362 then they will be preserved by the callee across the call. This doesn't
363 apply for values returned in callee-saved registers.
365 - On X86-64 the callee preserves all general purpose registers, except for
366 R11. R11 can be used as a scratch register. Floating-point registers
367 (XMMs/YMMs) are not preserved and need to be saved by the caller.
369 The idea behind this convention is to support calls to runtime functions
370 that have a hot path and a cold path. The hot path is usually a small piece
371 of code that doesn't many registers. The cold path might need to call out to
372 another function and therefore only needs to preserve the caller-saved
373 registers, which haven't already been saved by the caller. The
374 `PreserveMost` calling convention is very similar to the `cold` calling
375 convention in terms of caller/callee-saved registers, but they are used for
376 different types of function calls. `coldcc` is for function calls that are
377 rarely executed, whereas `preserve_mostcc` function calls are intended to be
378 on the hot path and definitely executed a lot. Furthermore `preserve_mostcc`
379 doesn't prevent the inliner from inlining the function call.
381 This calling convention will be used by a future version of the ObjectiveC
382 runtime and should therefore still be considered experimental at this time.
383 Although this convention was created to optimize certain runtime calls to
384 the ObjectiveC runtime, it is not limited to this runtime and might be used
385 by other runtimes in the future too. The current implementation only
386 supports X86-64, but the intention is to support more architectures in the
388 "``preserve_allcc``" - The `PreserveAll` calling convention
389 This calling convention attempts to make the code in the caller even less
390 intrusive than the `PreserveMost` calling convention. This calling
391 convention also behaves identical to the `C` calling convention on how
392 arguments and return values are passed, but it uses a different set of
393 caller/callee-saved registers. This removes the burden of saving and
394 recovering a large register set before and after the call in the caller. If
395 the arguments are passed in callee-saved registers, then they will be
396 preserved by the callee across the call. This doesn't apply for values
397 returned in callee-saved registers.
399 - On X86-64 the callee preserves all general purpose registers, except for
400 R11. R11 can be used as a scratch register. Furthermore it also preserves
401 all floating-point registers (XMMs/YMMs).
403 The idea behind this convention is to support calls to runtime functions
404 that don't need to call out to any other functions.
406 This calling convention, like the `PreserveMost` calling convention, will be
407 used by a future version of the ObjectiveC runtime and should be considered
408 experimental at this time.
409 "``cc <n>``" - Numbered convention
410 Any calling convention may be specified by number, allowing
411 target-specific calling conventions to be used. Target specific
412 calling conventions start at 64.
414 More calling conventions can be added/defined on an as-needed basis, to
415 support Pascal conventions or any other well-known target-independent
418 .. _visibilitystyles:
423 All Global Variables and Functions have one of the following visibility
426 "``default``" - Default style
427 On targets that use the ELF object file format, default visibility
428 means that the declaration is visible to other modules and, in
429 shared libraries, means that the declared entity may be overridden.
430 On Darwin, default visibility means that the declaration is visible
431 to other modules. Default visibility corresponds to "external
432 linkage" in the language.
433 "``hidden``" - Hidden style
434 Two declarations of an object with hidden visibility refer to the
435 same object if they are in the same shared object. Usually, hidden
436 visibility indicates that the symbol will not be placed into the
437 dynamic symbol table, so no other module (executable or shared
438 library) can reference it directly.
439 "``protected``" - Protected style
440 On ELF, protected visibility indicates that the symbol will be
441 placed in the dynamic symbol table, but that references within the
442 defining module will bind to the local symbol. That is, the symbol
443 cannot be overridden by another module.
445 A symbol with ``internal`` or ``private`` linkage must have ``default``
453 All Global Variables, Functions and Aliases can have one of the following
457 "``dllimport``" causes the compiler to reference a function or variable via
458 a global pointer to a pointer that is set up by the DLL exporting the
459 symbol. On Microsoft Windows targets, the pointer name is formed by
460 combining ``__imp_`` and the function or variable name.
462 "``dllexport``" causes the compiler to provide a global pointer to a pointer
463 in a DLL, so that it can be referenced with the ``dllimport`` attribute. On
464 Microsoft Windows targets, the pointer name is formed by combining
465 ``__imp_`` and the function or variable name. Since this storage class
466 exists for defining a dll interface, the compiler, assembler and linker know
467 it is externally referenced and must refrain from deleting the symbol.
471 Thread Local Storage Models
472 ---------------------------
474 A variable may be defined as ``thread_local``, which means that it will
475 not be shared by threads (each thread will have a separated copy of the
476 variable). Not all targets support thread-local variables. Optionally, a
477 TLS model may be specified:
480 For variables that are only used within the current shared library.
482 For variables in modules that will not be loaded dynamically.
484 For variables defined in the executable and only used within it.
486 If no explicit model is given, the "general dynamic" model is used.
488 The models correspond to the ELF TLS models; see `ELF Handling For
489 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
490 more information on under which circumstances the different models may
491 be used. The target may choose a different TLS model if the specified
492 model is not supported, or if a better choice of model can be made.
494 A model can also be specified in a alias, but then it only governs how
495 the alias is accessed. It will not have any effect in the aliasee.
502 LLVM IR allows you to specify both "identified" and "literal" :ref:`structure
503 types <t_struct>`. Literal types are uniqued structurally, but identified types
504 are never uniqued. An :ref:`opaque structural type <t_opaque>` can also be used
505 to forward declare a type that is not yet available.
507 An example of a identified structure specification is:
511 %mytype = type { %mytype*, i32 }
513 Prior to the LLVM 3.0 release, identified types were structurally uniqued. Only
514 literal types are uniqued in recent versions of LLVM.
521 Global variables define regions of memory allocated at compilation time
524 Global variables definitions must be initialized.
526 Global variables in other translation units can also be declared, in which
527 case they don't have an initializer.
529 Either global variable definitions or declarations may have an explicit section
530 to be placed in and may have an optional explicit alignment specified.
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.
567 Additionally, the global can placed in a comdat if the target has the necessary
570 By default, global initializers are optimized by assuming that global
571 variables defined within the module are not modified from their
572 initial values before the start of the global initializer. This is
573 true even for variables potentially accessible from outside the
574 module, including those with external linkage or appearing in
575 ``@llvm.used`` or dllexported variables. This assumption may be suppressed
576 by marking the variable with ``externally_initialized``.
578 An explicit alignment may be specified for a global, which must be a
579 power of 2. If not present, or if the alignment is set to zero, the
580 alignment of the global is set by the target to whatever it feels
581 convenient. If an explicit alignment is specified, the global is forced
582 to have exactly that alignment. Targets and optimizers are not allowed
583 to over-align the global if the global has an assigned section. In this
584 case, the extra alignment could be observable: for example, code could
585 assume that the globals are densely packed in their section and try to
586 iterate over them as an array, alignment padding would break this
587 iteration. The maximum alignment is ``1 << 29``.
589 Globals can also have a :ref:`DLL storage class <dllstorageclass>`.
591 Variables and aliasaes can have a
592 :ref:`Thread Local Storage Model <tls_model>`.
596 [@<GlobalVarName> =] [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal]
597 [unnamed_addr] [AddrSpace] [ExternallyInitialized]
598 <global | constant> <Type> [<InitializerConstant>]
599 [, section "name"] [, align <Alignment>]
601 For example, the following defines a global in a numbered address space
602 with an initializer, section, and alignment:
606 @G = addrspace(5) constant float 1.0, section "foo", align 4
608 The following example just declares a global variable
612 @G = external global i32
614 The following example defines a thread-local global with the
615 ``initialexec`` TLS model:
619 @G = thread_local(initialexec) global i32 0, align 4
621 .. _functionstructure:
626 LLVM function definitions consist of the "``define``" keyword, an
627 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
628 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
629 an optional :ref:`calling convention <callingconv>`,
630 an optional ``unnamed_addr`` attribute, a return type, an optional
631 :ref:`parameter attribute <paramattrs>` for the return type, a function
632 name, a (possibly empty) argument list (each with optional :ref:`parameter
633 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
634 an optional section, an optional alignment,
635 an optional :ref:`comdat <langref_comdats>`,
636 an optional :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
637 an optional :ref:`prologue <prologuedata>`, an opening
638 curly brace, a list of basic blocks, and a closing curly brace.
640 LLVM function declarations consist of the "``declare``" keyword, an
641 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
642 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
643 an optional :ref:`calling convention <callingconv>`,
644 an optional ``unnamed_addr`` attribute, a return type, an optional
645 :ref:`parameter attribute <paramattrs>` for the return type, a function
646 name, a possibly empty list of arguments, an optional alignment, an optional
647 :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
648 and an optional :ref:`prologue <prologuedata>`.
650 A function definition contains a list of basic blocks, forming the CFG (Control
651 Flow Graph) for the function. Each basic block may optionally start with a label
652 (giving the basic block a symbol table entry), contains a list of instructions,
653 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
654 function return). If an explicit label is not provided, a block is assigned an
655 implicit numbered label, using the next value from the same counter as used for
656 unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
657 entry block does not have an explicit label, it will be assigned label "%0",
658 then the first unnamed temporary in that block will be "%1", etc.
660 The first basic block in a function is special in two ways: it is
661 immediately executed on entrance to the function, and it is not allowed
662 to have predecessor basic blocks (i.e. there can not be any branches to
663 the entry block of a function). Because the block can have no
664 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
666 LLVM allows an explicit section to be specified for functions. If the
667 target supports it, it will emit functions to the section specified.
668 Additionally, the function can placed in a COMDAT.
670 An explicit alignment may be specified for a function. If not present,
671 or if the alignment is set to zero, the alignment of the function is set
672 by the target to whatever it feels convenient. If an explicit alignment
673 is specified, the function is forced to have at least that much
674 alignment. All alignments must be a power of 2.
676 If the ``unnamed_addr`` attribute is given, the address is know to not
677 be significant and two identical functions can be merged.
681 define [linkage] [visibility] [DLLStorageClass]
683 <ResultType> @<FunctionName> ([argument list])
684 [unnamed_addr] [fn Attrs] [section "name"] [comdat $<ComdatName>]
685 [align N] [gc] [prefix Constant] [prologue Constant] { ... }
687 The argument list is a comma seperated sequence of arguments where each
688 argument is of the following form
692 <type> [parameter Attrs] [name]
700 Aliases, unlike function or variables, don't create any new data. They
701 are just a new symbol and metadata for an existing position.
703 Aliases have a name and an aliasee that is either a global value or a
706 Aliases may have an optional :ref:`linkage type <linkage>`, an optional
707 :ref:`visibility style <visibility>`, an optional :ref:`DLL storage class
708 <dllstorageclass>` and an optional :ref:`tls model <tls_model>`.
712 @<Name> = [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal] [unnamed_addr] alias <AliaseeTy> @<Aliasee>
714 The linkage must be one of ``private``, ``internal``, ``linkonce``, ``weak``,
715 ``linkonce_odr``, ``weak_odr``, ``external``. Note that some system linkers
716 might not correctly handle dropping a weak symbol that is aliased.
718 Alias that are not ``unnamed_addr`` are guaranteed to have the same address as
719 the aliasee expression. ``unnamed_addr`` ones are only guaranteed to point
722 Since aliases are only a second name, some restrictions apply, of which
723 some can only be checked when producing an object file:
725 * The expression defining the aliasee must be computable at assembly
726 time. Since it is just a name, no relocations can be used.
728 * No alias in the expression can be weak as the possibility of the
729 intermediate alias being overridden cannot be represented in an
732 * No global value in the expression can be a declaration, since that
733 would require a relocation, which is not possible.
740 Comdat IR provides access to COFF and ELF object file COMDAT functionality.
742 Comdats have a name which represents the COMDAT key. All global objects that
743 specify this key will only end up in the final object file if the linker chooses
744 that key over some other key. Aliases are placed in the same COMDAT that their
745 aliasee computes to, if any.
747 Comdats have a selection kind to provide input on how the linker should
748 choose between keys in two different object files.
752 $<Name> = comdat SelectionKind
754 The selection kind must be one of the following:
757 The linker may choose any COMDAT key, the choice is arbitrary.
759 The linker may choose any COMDAT key but the sections must contain the
762 The linker will choose the section containing the largest COMDAT key.
764 The linker requires that only section with this COMDAT key exist.
766 The linker may choose any COMDAT key but the sections must contain the
769 Note that the Mach-O platform doesn't support COMDATs and ELF only supports
770 ``any`` as a selection kind.
772 Here is an example of a COMDAT group where a function will only be selected if
773 the COMDAT key's section is the largest:
777 $foo = comdat largest
778 @foo = global i32 2, comdat $foo
780 define void @bar() comdat $foo {
784 In a COFF object file, this will create a COMDAT section with selection kind
785 ``IMAGE_COMDAT_SELECT_LARGEST`` containing the contents of the ``@foo`` symbol
786 and another COMDAT section with selection kind
787 ``IMAGE_COMDAT_SELECT_ASSOCIATIVE`` which is associated with the first COMDAT
788 section and contains the contents of the ``@bar`` symbol.
790 There are some restrictions on the properties of the global object.
791 It, or an alias to it, must have the same name as the COMDAT group when
793 The contents and size of this object may be used during link-time to determine
794 which COMDAT groups get selected depending on the selection kind.
795 Because the name of the object must match the name of the COMDAT group, the
796 linkage of the global object must not be local; local symbols can get renamed
797 if a collision occurs in the symbol table.
799 The combined use of COMDATS and section attributes may yield surprising results.
806 @g1 = global i32 42, section "sec", comdat $foo
807 @g2 = global i32 42, section "sec", comdat $bar
809 From the object file perspective, this requires the creation of two sections
810 with the same name. This is necessary because both globals belong to different
811 COMDAT groups and COMDATs, at the object file level, are represented by
814 Note that certain IR constructs like global variables and functions may create
815 COMDATs in the object file in addition to any which are specified using COMDAT
816 IR. This arises, for example, when a global variable has linkonce_odr linkage.
818 .. _namedmetadatastructure:
823 Named metadata is a collection of metadata. :ref:`Metadata
824 nodes <metadata>` (but not metadata strings) are the only valid
825 operands for a named metadata.
829 ; Some unnamed metadata nodes, which are referenced by the named metadata.
830 !0 = metadata !{metadata !"zero"}
831 !1 = metadata !{metadata !"one"}
832 !2 = metadata !{metadata !"two"}
834 !name = !{!0, !1, !2}
841 The return type and each parameter of a function type may have a set of
842 *parameter attributes* associated with them. Parameter attributes are
843 used to communicate additional information about the result or
844 parameters of a function. Parameter attributes are considered to be part
845 of the function, not of the function type, so functions with different
846 parameter attributes can have the same function type.
848 Parameter attributes are simple keywords that follow the type specified.
849 If multiple parameter attributes are needed, they are space separated.
854 declare i32 @printf(i8* noalias nocapture, ...)
855 declare i32 @atoi(i8 zeroext)
856 declare signext i8 @returns_signed_char()
858 Note that any attributes for the function result (``nounwind``,
859 ``readonly``) come immediately after the argument list.
861 Currently, only the following parameter attributes are defined:
864 This indicates to the code generator that the parameter or return
865 value should be zero-extended to the extent required by the target's
866 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
867 the caller (for a parameter) or the callee (for a return value).
869 This indicates to the code generator that the parameter or return
870 value should be sign-extended to the extent required by the target's
871 ABI (which is usually 32-bits) by the caller (for a parameter) or
872 the callee (for a return value).
874 This indicates that this parameter or return value should be treated
875 in a special target-dependent fashion during while emitting code for
876 a function call or return (usually, by putting it in a register as
877 opposed to memory, though some targets use it to distinguish between
878 two different kinds of registers). Use of this attribute is
881 This indicates that the pointer parameter should really be passed by
882 value to the function. The attribute implies that a hidden copy of
883 the pointee is made between the caller and the callee, so the callee
884 is unable to modify the value in the caller. This attribute is only
885 valid on LLVM pointer arguments. It is generally used to pass
886 structs and arrays by value, but is also valid on pointers to
887 scalars. The copy is considered to belong to the caller not the
888 callee (for example, ``readonly`` functions should not write to
889 ``byval`` parameters). This is not a valid attribute for return
892 The byval attribute also supports specifying an alignment with the
893 align attribute. It indicates the alignment of the stack slot to
894 form and the known alignment of the pointer specified to the call
895 site. If the alignment is not specified, then the code generator
896 makes a target-specific assumption.
902 The ``inalloca`` argument attribute allows the caller to take the
903 address of outgoing stack arguments. An ``inalloca`` argument must
904 be a pointer to stack memory produced by an ``alloca`` instruction.
905 The alloca, or argument allocation, must also be tagged with the
906 inalloca keyword. Only the last argument may have the ``inalloca``
907 attribute, and that argument is guaranteed to be passed in memory.
909 An argument allocation may be used by a call at most once because
910 the call may deallocate it. The ``inalloca`` attribute cannot be
911 used in conjunction with other attributes that affect argument
912 storage, like ``inreg``, ``nest``, ``sret``, or ``byval``. The
913 ``inalloca`` attribute also disables LLVM's implicit lowering of
914 large aggregate return values, which means that frontend authors
915 must lower them with ``sret`` pointers.
917 When the call site is reached, the argument allocation must have
918 been the most recent stack allocation that is still live, or the
919 results are undefined. It is possible to allocate additional stack
920 space after an argument allocation and before its call site, but it
921 must be cleared off with :ref:`llvm.stackrestore
924 See :doc:`InAlloca` for more information on how to use this
928 This indicates that the pointer parameter specifies the address of a
929 structure that is the return value of the function in the source
930 program. This pointer must be guaranteed by the caller to be valid:
931 loads and stores to the structure may be assumed by the callee
932 not to trap and to be properly aligned. This may only be applied to
933 the first parameter. This is not a valid attribute for return
937 This indicates that the pointer value may be assumed by the optimizer to
938 have the specified alignment.
940 Note that this attribute has additional semantics when combined with the
946 This indicates that objects accessed via pointer values
947 :ref:`based <pointeraliasing>` on the argument or return value are not also
948 accessed, during the execution of the function, via pointer values not
949 *based* on the argument or return value. The attribute on a return value
950 also has additional semantics described below. The caller shares the
951 responsibility with the callee for ensuring that these requirements are met.
952 For further details, please see the discussion of the NoAlias response in
953 :ref:`alias analysis <Must, May, or No>`.
955 Note that this definition of ``noalias`` is intentionally similar
956 to the definition of ``restrict`` in C99 for function arguments.
958 For function return values, C99's ``restrict`` is not meaningful,
959 while LLVM's ``noalias`` is. Furthermore, the semantics of the ``noalias``
960 attribute on return values are stronger than the semantics of the attribute
961 when used on function arguments. On function return values, the ``noalias``
962 attribute indicates that the function acts like a system memory allocation
963 function, returning a pointer to allocated storage disjoint from the
964 storage for any other object accessible to the caller.
967 This indicates that the callee does not make any copies of the
968 pointer that outlive the callee itself. This is not a valid
969 attribute for return values.
974 This indicates that the pointer parameter can be excised using the
975 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
976 attribute for return values and can only be applied to one parameter.
979 This indicates that the function always returns the argument as its return
980 value. This is an optimization hint to the code generator when generating
981 the caller, allowing tail call optimization and omission of register saves
982 and restores in some cases; it is not checked or enforced when generating
983 the callee. The parameter and the function return type must be valid
984 operands for the :ref:`bitcast instruction <i_bitcast>`. This is not a
985 valid attribute for return values and can only be applied to one parameter.
988 This indicates that the parameter or return pointer is not null. This
989 attribute may only be applied to pointer typed parameters. This is not
990 checked or enforced by LLVM, the caller must ensure that the pointer
991 passed in is non-null, or the callee must ensure that the returned pointer
994 ``dereferenceable(<n>)``
995 This indicates that the parameter or return pointer is dereferenceable. This
996 attribute may only be applied to pointer typed parameters. A pointer that
997 is dereferenceable can be loaded from speculatively without a risk of
998 trapping. The number of bytes known to be dereferenceable must be provided
999 in parentheses. It is legal for the number of bytes to be less than the
1000 size of the pointee type. The ``nonnull`` attribute does not imply
1001 dereferenceability (consider a pointer to one element past the end of an
1002 array), however ``dereferenceable(<n>)`` does imply ``nonnull`` in
1003 ``addrspace(0)`` (which is the default address space).
1007 Garbage Collector Names
1008 -----------------------
1010 Each function may specify a garbage collector name, which is simply a
1013 .. code-block:: llvm
1015 define void @f() gc "name" { ... }
1017 The compiler declares the supported values of *name*. Specifying a
1018 collector will cause the compiler to alter its output in order to
1019 support the named garbage collection algorithm.
1026 Prefix data is data associated with a function which the code
1027 generator will emit immediately before the function's entrypoint.
1028 The purpose of this feature is to allow frontends to associate
1029 language-specific runtime metadata with specific functions and make it
1030 available through the function pointer while still allowing the
1031 function pointer to be called.
1033 To access the data for a given function, a program may bitcast the
1034 function pointer to a pointer to the constant's type and dereference
1035 index -1. This implies that the IR symbol points just past the end of
1036 the prefix data. For instance, take the example of a function annotated
1037 with a single ``i32``,
1039 .. code-block:: llvm
1041 define void @f() prefix i32 123 { ... }
1043 The prefix data can be referenced as,
1045 .. code-block:: llvm
1047 %0 = bitcast *void () @f to *i32
1048 %a = getelementptr inbounds *i32 %0, i32 -1
1051 Prefix data is laid out as if it were an initializer for a global variable
1052 of the prefix data's type. The function will be placed such that the
1053 beginning of the prefix data is aligned. This means that if the size
1054 of the prefix data is not a multiple of the alignment size, the
1055 function's entrypoint will not be aligned. If alignment of the
1056 function's entrypoint is desired, padding must be added to the prefix
1059 A function may have prefix data but no body. This has similar semantics
1060 to the ``available_externally`` linkage in that the data may be used by the
1061 optimizers but will not be emitted in the object file.
1068 The ``prologue`` attribute allows arbitrary code (encoded as bytes) to
1069 be inserted prior to the function body. This can be used for enabling
1070 function hot-patching and instrumentation.
1072 To maintain the semantics of ordinary function calls, the prologue data must
1073 have a particular format. Specifically, it must begin with a sequence of
1074 bytes which decode to a sequence of machine instructions, valid for the
1075 module's target, which transfer control to the point immediately succeeding
1076 the prologue data, without performing any other visible action. This allows
1077 the inliner and other passes to reason about the semantics of the function
1078 definition without needing to reason about the prologue data. Obviously this
1079 makes the format of the prologue data highly target dependent.
1081 A trivial example of valid prologue data for the x86 architecture is ``i8 144``,
1082 which encodes the ``nop`` instruction:
1084 .. code-block:: llvm
1086 define void @f() prologue i8 144 { ... }
1088 Generally prologue data can be formed by encoding a relative branch instruction
1089 which skips the metadata, as in this example of valid prologue data for the
1090 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
1092 .. code-block:: llvm
1094 %0 = type <{ i8, i8, i8* }>
1096 define void @f() prologue %0 <{ i8 235, i8 8, i8* @md}> { ... }
1098 A function may have prologue data but no body. This has similar semantics
1099 to the ``available_externally`` linkage in that the data may be used by the
1100 optimizers but will not be emitted in the object file.
1107 Attribute groups are groups of attributes that are referenced by objects within
1108 the IR. They are important for keeping ``.ll`` files readable, because a lot of
1109 functions will use the same set of attributes. In the degenerative case of a
1110 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
1111 group will capture the important command line flags used to build that file.
1113 An attribute group is a module-level object. To use an attribute group, an
1114 object references the attribute group's ID (e.g. ``#37``). An object may refer
1115 to more than one attribute group. In that situation, the attributes from the
1116 different groups are merged.
1118 Here is an example of attribute groups for a function that should always be
1119 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
1121 .. code-block:: llvm
1123 ; Target-independent attributes:
1124 attributes #0 = { alwaysinline alignstack=4 }
1126 ; Target-dependent attributes:
1127 attributes #1 = { "no-sse" }
1129 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
1130 define void @f() #0 #1 { ... }
1137 Function attributes are set to communicate additional information about
1138 a function. Function attributes are considered to be part of the
1139 function, not of the function type, so functions with different function
1140 attributes can have the same function type.
1142 Function attributes are simple keywords that follow the type specified.
1143 If multiple attributes are needed, they are space separated. For
1146 .. code-block:: llvm
1148 define void @f() noinline { ... }
1149 define void @f() alwaysinline { ... }
1150 define void @f() alwaysinline optsize { ... }
1151 define void @f() optsize { ... }
1154 This attribute indicates that, when emitting the prologue and
1155 epilogue, the backend should forcibly align the stack pointer.
1156 Specify the desired alignment, which must be a power of two, in
1159 This attribute indicates that the inliner should attempt to inline
1160 this function into callers whenever possible, ignoring any active
1161 inlining size threshold for this caller.
1163 This indicates that the callee function at a call site should be
1164 recognized as a built-in function, even though the function's declaration
1165 uses the ``nobuiltin`` attribute. This is only valid at call sites for
1166 direct calls to functions that are declared with the ``nobuiltin``
1169 This attribute indicates that this function is rarely called. When
1170 computing edge weights, basic blocks post-dominated by a cold
1171 function call are also considered to be cold; and, thus, given low
1174 This attribute indicates that the source code contained a hint that
1175 inlining this function is desirable (such as the "inline" keyword in
1176 C/C++). It is just a hint; it imposes no requirements on the
1179 This attribute indicates that the function should be added to a
1180 jump-instruction table at code-generation time, and that all address-taken
1181 references to this function should be replaced with a reference to the
1182 appropriate jump-instruction-table function pointer. Note that this creates
1183 a new pointer for the original function, which means that code that depends
1184 on function-pointer identity can break. So, any function annotated with
1185 ``jumptable`` must also be ``unnamed_addr``.
1187 This attribute suggests that optimization passes and code generator
1188 passes make choices that keep the code size of this function as small
1189 as possible and perform optimizations that may sacrifice runtime
1190 performance in order to minimize the size of the generated code.
1192 This attribute disables prologue / epilogue emission for the
1193 function. This can have very system-specific consequences.
1195 This indicates that the callee function at a call site is not recognized as
1196 a built-in function. LLVM will retain the original call and not replace it
1197 with equivalent code based on the semantics of the built-in function, unless
1198 the call site uses the ``builtin`` attribute. This is valid at call sites
1199 and on function declarations and definitions.
1201 This attribute indicates that calls to the function cannot be
1202 duplicated. A call to a ``noduplicate`` function may be moved
1203 within its parent function, but may not be duplicated within
1204 its parent function.
1206 A function containing a ``noduplicate`` call may still
1207 be an inlining candidate, provided that the call is not
1208 duplicated by inlining. That implies that the function has
1209 internal linkage and only has one call site, so the original
1210 call is dead after inlining.
1212 This attributes disables implicit floating point instructions.
1214 This attribute indicates that the inliner should never inline this
1215 function in any situation. This attribute may not be used together
1216 with the ``alwaysinline`` attribute.
1218 This attribute suppresses lazy symbol binding for the function. This
1219 may make calls to the function faster, at the cost of extra program
1220 startup time if the function is not called during program startup.
1222 This attribute indicates that the code generator should not use a
1223 red zone, even if the target-specific ABI normally permits it.
1225 This function attribute indicates that the function never returns
1226 normally. This produces undefined behavior at runtime if the
1227 function ever does dynamically return.
1229 This function attribute indicates that the function never returns
1230 with an unwind or exceptional control flow. If the function does
1231 unwind, its runtime behavior is undefined.
1233 This function attribute indicates that the function is not optimized
1234 by any optimization or code generator passes with the
1235 exception of interprocedural optimization passes.
1236 This attribute cannot be used together with the ``alwaysinline``
1237 attribute; this attribute is also incompatible
1238 with the ``minsize`` attribute and the ``optsize`` attribute.
1240 This attribute requires the ``noinline`` attribute to be specified on
1241 the function as well, so the function is never inlined into any caller.
1242 Only functions with the ``alwaysinline`` attribute are valid
1243 candidates for inlining into the body of this function.
1245 This attribute suggests that optimization passes and code generator
1246 passes make choices that keep the code size of this function low,
1247 and otherwise do optimizations specifically to reduce code size as
1248 long as they do not significantly impact runtime performance.
1250 On a function, this attribute indicates that the function computes its
1251 result (or decides to unwind an exception) based strictly on its arguments,
1252 without dereferencing any pointer arguments or otherwise accessing
1253 any mutable state (e.g. memory, control registers, etc) visible to
1254 caller functions. It does not write through any pointer arguments
1255 (including ``byval`` arguments) and never changes any state visible
1256 to callers. This means that it cannot unwind exceptions by calling
1257 the ``C++`` exception throwing methods.
1259 On an argument, this attribute indicates that the function does not
1260 dereference that pointer argument, even though it may read or write the
1261 memory that the pointer points to if accessed through other pointers.
1263 On a function, this attribute indicates that the function does not write
1264 through any pointer arguments (including ``byval`` arguments) or otherwise
1265 modify any state (e.g. memory, control registers, etc) visible to
1266 caller functions. It may dereference pointer arguments and read
1267 state that may be set in the caller. A readonly function always
1268 returns the same value (or unwinds an exception identically) when
1269 called with the same set of arguments and global state. It cannot
1270 unwind an exception by calling the ``C++`` exception throwing
1273 On an argument, this attribute indicates that the function does not write
1274 through this pointer argument, even though it may write to the memory that
1275 the pointer points to.
1277 This attribute indicates that this function can return twice. The C
1278 ``setjmp`` is an example of such a function. The compiler disables
1279 some optimizations (like tail calls) in the caller of these
1281 ``sanitize_address``
1282 This attribute indicates that AddressSanitizer checks
1283 (dynamic address safety analysis) are enabled for this function.
1285 This attribute indicates that MemorySanitizer checks (dynamic detection
1286 of accesses to uninitialized memory) are enabled for this function.
1288 This attribute indicates that ThreadSanitizer checks
1289 (dynamic thread safety analysis) are enabled for this function.
1291 This attribute indicates that the function should emit a stack
1292 smashing protector. It is in the form of a "canary" --- a random value
1293 placed on the stack before the local variables that's checked upon
1294 return from the function to see if it has been overwritten. A
1295 heuristic is used to determine if a function needs stack protectors
1296 or not. The heuristic used will enable protectors for functions with:
1298 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1299 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1300 - Calls to alloca() with variable sizes or constant sizes greater than
1301 ``ssp-buffer-size``.
1303 Variables that are identified as requiring a protector will be arranged
1304 on the stack such that they are adjacent to the stack protector guard.
1306 If a function that has an ``ssp`` attribute is inlined into a
1307 function that doesn't have an ``ssp`` attribute, then the resulting
1308 function will have an ``ssp`` attribute.
1310 This attribute indicates that the function should *always* emit a
1311 stack smashing protector. This overrides the ``ssp`` function
1314 Variables that are identified as requiring a protector will be arranged
1315 on the stack such that they are adjacent to the stack protector guard.
1316 The specific layout rules are:
1318 #. Large arrays and structures containing large arrays
1319 (``>= ssp-buffer-size``) are closest to the stack protector.
1320 #. Small arrays and structures containing small arrays
1321 (``< ssp-buffer-size``) are 2nd closest to the protector.
1322 #. Variables that have had their address taken are 3rd closest to the
1325 If a function that has an ``sspreq`` attribute is inlined into a
1326 function that doesn't have an ``sspreq`` attribute or which has an
1327 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1328 an ``sspreq`` attribute.
1330 This attribute indicates that the function should emit a stack smashing
1331 protector. This attribute causes a strong heuristic to be used when
1332 determining if a function needs stack protectors. The strong heuristic
1333 will enable protectors for functions with:
1335 - Arrays of any size and type
1336 - Aggregates containing an array of any size and type.
1337 - Calls to alloca().
1338 - Local variables that have had their address taken.
1340 Variables that are identified as requiring a protector will be arranged
1341 on the stack such that they are adjacent to the stack protector guard.
1342 The specific layout rules are:
1344 #. Large arrays and structures containing large arrays
1345 (``>= ssp-buffer-size``) are closest to the stack protector.
1346 #. Small arrays and structures containing small arrays
1347 (``< ssp-buffer-size``) are 2nd closest to the protector.
1348 #. Variables that have had their address taken are 3rd closest to the
1351 This overrides the ``ssp`` function attribute.
1353 If a function that has an ``sspstrong`` attribute is inlined into a
1354 function that doesn't have an ``sspstrong`` attribute, then the
1355 resulting function will have an ``sspstrong`` attribute.
1357 This attribute indicates that the ABI being targeted requires that
1358 an unwind table entry be produce for this function even if we can
1359 show that no exceptions passes by it. This is normally the case for
1360 the ELF x86-64 abi, but it can be disabled for some compilation
1365 Module-Level Inline Assembly
1366 ----------------------------
1368 Modules may contain "module-level inline asm" blocks, which corresponds
1369 to the GCC "file scope inline asm" blocks. These blocks are internally
1370 concatenated by LLVM and treated as a single unit, but may be separated
1371 in the ``.ll`` file if desired. The syntax is very simple:
1373 .. code-block:: llvm
1375 module asm "inline asm code goes here"
1376 module asm "more can go here"
1378 The strings can contain any character by escaping non-printable
1379 characters. The escape sequence used is simply "\\xx" where "xx" is the
1380 two digit hex code for the number.
1382 The inline asm code is simply printed to the machine code .s file when
1383 assembly code is generated.
1385 .. _langref_datalayout:
1390 A module may specify a target specific data layout string that specifies
1391 how data is to be laid out in memory. The syntax for the data layout is
1394 .. code-block:: llvm
1396 target datalayout = "layout specification"
1398 The *layout specification* consists of a list of specifications
1399 separated by the minus sign character ('-'). Each specification starts
1400 with a letter and may include other information after the letter to
1401 define some aspect of the data layout. The specifications accepted are
1405 Specifies that the target lays out data in big-endian form. That is,
1406 the bits with the most significance have the lowest address
1409 Specifies that the target lays out data in little-endian form. That
1410 is, the bits with the least significance have the lowest address
1413 Specifies the natural alignment of the stack in bits. Alignment
1414 promotion of stack variables is limited to the natural stack
1415 alignment to avoid dynamic stack realignment. The stack alignment
1416 must be a multiple of 8-bits. If omitted, the natural stack
1417 alignment defaults to "unspecified", which does not prevent any
1418 alignment promotions.
1419 ``p[n]:<size>:<abi>:<pref>``
1420 This specifies the *size* of a pointer and its ``<abi>`` and
1421 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1422 bits. The address space, ``n`` is optional, and if not specified,
1423 denotes the default address space 0. The value of ``n`` must be
1424 in the range [1,2^23).
1425 ``i<size>:<abi>:<pref>``
1426 This specifies the alignment for an integer type of a given bit
1427 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1428 ``v<size>:<abi>:<pref>``
1429 This specifies the alignment for a vector type of a given bit
1431 ``f<size>:<abi>:<pref>``
1432 This specifies the alignment for a floating point type of a given bit
1433 ``<size>``. Only values of ``<size>`` that are supported by the target
1434 will work. 32 (float) and 64 (double) are supported on all targets; 80
1435 or 128 (different flavors of long double) are also supported on some
1438 This specifies the alignment for an object of aggregate type.
1440 If present, specifies that llvm names are mangled in the output. The
1443 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1444 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1445 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1446 symbols get a ``_`` prefix.
1447 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1448 functions also get a suffix based on the frame size.
1449 ``n<size1>:<size2>:<size3>...``
1450 This specifies a set of native integer widths for the target CPU in
1451 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1452 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1453 this set are considered to support most general arithmetic operations
1456 On every specification that takes a ``<abi>:<pref>``, specifying the
1457 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1458 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1460 When constructing the data layout for a given target, LLVM starts with a
1461 default set of specifications which are then (possibly) overridden by
1462 the specifications in the ``datalayout`` keyword. The default
1463 specifications are given in this list:
1465 - ``E`` - big endian
1466 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1467 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1468 same as the default address space.
1469 - ``S0`` - natural stack alignment is unspecified
1470 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1471 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1472 - ``i16:16:16`` - i16 is 16-bit aligned
1473 - ``i32:32:32`` - i32 is 32-bit aligned
1474 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1475 alignment of 64-bits
1476 - ``f16:16:16`` - half is 16-bit aligned
1477 - ``f32:32:32`` - float is 32-bit aligned
1478 - ``f64:64:64`` - double is 64-bit aligned
1479 - ``f128:128:128`` - quad is 128-bit aligned
1480 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1481 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1482 - ``a:0:64`` - aggregates are 64-bit aligned
1484 When LLVM is determining the alignment for a given type, it uses the
1487 #. If the type sought is an exact match for one of the specifications,
1488 that specification is used.
1489 #. If no match is found, and the type sought is an integer type, then
1490 the smallest integer type that is larger than the bitwidth of the
1491 sought type is used. If none of the specifications are larger than
1492 the bitwidth then the largest integer type is used. For example,
1493 given the default specifications above, the i7 type will use the
1494 alignment of i8 (next largest) while both i65 and i256 will use the
1495 alignment of i64 (largest specified).
1496 #. If no match is found, and the type sought is a vector type, then the
1497 largest vector type that is smaller than the sought vector type will
1498 be used as a fall back. This happens because <128 x double> can be
1499 implemented in terms of 64 <2 x double>, for example.
1501 The function of the data layout string may not be what you expect.
1502 Notably, this is not a specification from the frontend of what alignment
1503 the code generator should use.
1505 Instead, if specified, the target data layout is required to match what
1506 the ultimate *code generator* expects. This string is used by the
1507 mid-level optimizers to improve code, and this only works if it matches
1508 what the ultimate code generator uses. If you would like to generate IR
1509 that does not embed this target-specific detail into the IR, then you
1510 don't have to specify the string. This will disable some optimizations
1511 that require precise layout information, but this also prevents those
1512 optimizations from introducing target specificity into the IR.
1519 A module may specify a target triple string that describes the target
1520 host. The syntax for the target triple is simply:
1522 .. code-block:: llvm
1524 target triple = "x86_64-apple-macosx10.7.0"
1526 The *target triple* string consists of a series of identifiers delimited
1527 by the minus sign character ('-'). The canonical forms are:
1531 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1532 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1534 This information is passed along to the backend so that it generates
1535 code for the proper architecture. It's possible to override this on the
1536 command line with the ``-mtriple`` command line option.
1538 .. _pointeraliasing:
1540 Pointer Aliasing Rules
1541 ----------------------
1543 Any memory access must be done through a pointer value associated with
1544 an address range of the memory access, otherwise the behavior is
1545 undefined. Pointer values are associated with address ranges according
1546 to the following rules:
1548 - A pointer value is associated with the addresses associated with any
1549 value it is *based* on.
1550 - An address of a global variable is associated with the address range
1551 of the variable's storage.
1552 - The result value of an allocation instruction is associated with the
1553 address range of the allocated storage.
1554 - A null pointer in the default address-space is associated with no
1556 - An integer constant other than zero or a pointer value returned from
1557 a function not defined within LLVM may be associated with address
1558 ranges allocated through mechanisms other than those provided by
1559 LLVM. Such ranges shall not overlap with any ranges of addresses
1560 allocated by mechanisms provided by LLVM.
1562 A pointer value is *based* on another pointer value according to the
1565 - A pointer value formed from a ``getelementptr`` operation is *based*
1566 on the first operand of the ``getelementptr``.
1567 - The result value of a ``bitcast`` is *based* on the operand of the
1569 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1570 values that contribute (directly or indirectly) to the computation of
1571 the pointer's value.
1572 - The "*based* on" relationship is transitive.
1574 Note that this definition of *"based"* is intentionally similar to the
1575 definition of *"based"* in C99, though it is slightly weaker.
1577 LLVM IR does not associate types with memory. The result type of a
1578 ``load`` merely indicates the size and alignment of the memory from
1579 which to load, as well as the interpretation of the value. The first
1580 operand type of a ``store`` similarly only indicates the size and
1581 alignment of the store.
1583 Consequently, type-based alias analysis, aka TBAA, aka
1584 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1585 :ref:`Metadata <metadata>` may be used to encode additional information
1586 which specialized optimization passes may use to implement type-based
1591 Volatile Memory Accesses
1592 ------------------------
1594 Certain memory accesses, such as :ref:`load <i_load>`'s,
1595 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1596 marked ``volatile``. The optimizers must not change the number of
1597 volatile operations or change their order of execution relative to other
1598 volatile operations. The optimizers *may* change the order of volatile
1599 operations relative to non-volatile operations. This is not Java's
1600 "volatile" and has no cross-thread synchronization behavior.
1602 IR-level volatile loads and stores cannot safely be optimized into
1603 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1604 flagged volatile. Likewise, the backend should never split or merge
1605 target-legal volatile load/store instructions.
1607 .. admonition:: Rationale
1609 Platforms may rely on volatile loads and stores of natively supported
1610 data width to be executed as single instruction. For example, in C
1611 this holds for an l-value of volatile primitive type with native
1612 hardware support, but not necessarily for aggregate types. The
1613 frontend upholds these expectations, which are intentionally
1614 unspecified in the IR. The rules above ensure that IR transformation
1615 do not violate the frontend's contract with the language.
1619 Memory Model for Concurrent Operations
1620 --------------------------------------
1622 The LLVM IR does not define any way to start parallel threads of
1623 execution or to register signal handlers. Nonetheless, there are
1624 platform-specific ways to create them, and we define LLVM IR's behavior
1625 in their presence. This model is inspired by the C++0x memory model.
1627 For a more informal introduction to this model, see the :doc:`Atomics`.
1629 We define a *happens-before* partial order as the least partial order
1632 - Is a superset of single-thread program order, and
1633 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1634 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1635 techniques, like pthread locks, thread creation, thread joining,
1636 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1637 Constraints <ordering>`).
1639 Note that program order does not introduce *happens-before* edges
1640 between a thread and signals executing inside that thread.
1642 Every (defined) read operation (load instructions, memcpy, atomic
1643 loads/read-modify-writes, etc.) R reads a series of bytes written by
1644 (defined) write operations (store instructions, atomic
1645 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1646 section, initialized globals are considered to have a write of the
1647 initializer which is atomic and happens before any other read or write
1648 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1649 may see any write to the same byte, except:
1651 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1652 write\ :sub:`2` happens before R\ :sub:`byte`, then
1653 R\ :sub:`byte` does not see write\ :sub:`1`.
1654 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1655 R\ :sub:`byte` does not see write\ :sub:`3`.
1657 Given that definition, R\ :sub:`byte` is defined as follows:
1659 - If R is volatile, the result is target-dependent. (Volatile is
1660 supposed to give guarantees which can support ``sig_atomic_t`` in
1661 C/C++, and may be used for accesses to addresses that do not behave
1662 like normal memory. It does not generally provide cross-thread
1664 - Otherwise, if there is no write to the same byte that happens before
1665 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1666 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1667 R\ :sub:`byte` returns the value written by that write.
1668 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1669 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1670 Memory Ordering Constraints <ordering>` section for additional
1671 constraints on how the choice is made.
1672 - Otherwise R\ :sub:`byte` returns ``undef``.
1674 R returns the value composed of the series of bytes it read. This
1675 implies that some bytes within the value may be ``undef`` **without**
1676 the entire value being ``undef``. Note that this only defines the
1677 semantics of the operation; it doesn't mean that targets will emit more
1678 than one instruction to read the series of bytes.
1680 Note that in cases where none of the atomic intrinsics are used, this
1681 model places only one restriction on IR transformations on top of what
1682 is required for single-threaded execution: introducing a store to a byte
1683 which might not otherwise be stored is not allowed in general.
1684 (Specifically, in the case where another thread might write to and read
1685 from an address, introducing a store can change a load that may see
1686 exactly one write into a load that may see multiple writes.)
1690 Atomic Memory Ordering Constraints
1691 ----------------------------------
1693 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1694 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1695 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1696 ordering parameters that determine which other atomic instructions on
1697 the same address they *synchronize with*. These semantics are borrowed
1698 from Java and C++0x, but are somewhat more colloquial. If these
1699 descriptions aren't precise enough, check those specs (see spec
1700 references in the :doc:`atomics guide <Atomics>`).
1701 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1702 differently since they don't take an address. See that instruction's
1703 documentation for details.
1705 For a simpler introduction to the ordering constraints, see the
1709 The set of values that can be read is governed by the happens-before
1710 partial order. A value cannot be read unless some operation wrote
1711 it. This is intended to provide a guarantee strong enough to model
1712 Java's non-volatile shared variables. This ordering cannot be
1713 specified for read-modify-write operations; it is not strong enough
1714 to make them atomic in any interesting way.
1716 In addition to the guarantees of ``unordered``, there is a single
1717 total order for modifications by ``monotonic`` operations on each
1718 address. All modification orders must be compatible with the
1719 happens-before order. There is no guarantee that the modification
1720 orders can be combined to a global total order for the whole program
1721 (and this often will not be possible). The read in an atomic
1722 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1723 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1724 order immediately before the value it writes. If one atomic read
1725 happens before another atomic read of the same address, the later
1726 read must see the same value or a later value in the address's
1727 modification order. This disallows reordering of ``monotonic`` (or
1728 stronger) operations on the same address. If an address is written
1729 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1730 read that address repeatedly, the other threads must eventually see
1731 the write. This corresponds to the C++0x/C1x
1732 ``memory_order_relaxed``.
1734 In addition to the guarantees of ``monotonic``, a
1735 *synchronizes-with* edge may be formed with a ``release`` operation.
1736 This is intended to model C++'s ``memory_order_acquire``.
1738 In addition to the guarantees of ``monotonic``, if this operation
1739 writes a value which is subsequently read by an ``acquire``
1740 operation, it *synchronizes-with* that operation. (This isn't a
1741 complete description; see the C++0x definition of a release
1742 sequence.) This corresponds to the C++0x/C1x
1743 ``memory_order_release``.
1744 ``acq_rel`` (acquire+release)
1745 Acts as both an ``acquire`` and ``release`` operation on its
1746 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1747 ``seq_cst`` (sequentially consistent)
1748 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1749 operation that only reads, ``release`` for an operation that only
1750 writes), there is a global total order on all
1751 sequentially-consistent operations on all addresses, which is
1752 consistent with the *happens-before* partial order and with the
1753 modification orders of all the affected addresses. Each
1754 sequentially-consistent read sees the last preceding write to the
1755 same address in this global order. This corresponds to the C++0x/C1x
1756 ``memory_order_seq_cst`` and Java volatile.
1760 If an atomic operation is marked ``singlethread``, it only *synchronizes
1761 with* or participates in modification and seq\_cst total orderings with
1762 other operations running in the same thread (for example, in signal
1770 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1771 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1772 :ref:`frem <i_frem>`) have the following flags that can set to enable
1773 otherwise unsafe floating point operations
1776 No NaNs - Allow optimizations to assume the arguments and result are not
1777 NaN. Such optimizations are required to retain defined behavior over
1778 NaNs, but the value of the result is undefined.
1781 No Infs - Allow optimizations to assume the arguments and result are not
1782 +/-Inf. Such optimizations are required to retain defined behavior over
1783 +/-Inf, but the value of the result is undefined.
1786 No Signed Zeros - Allow optimizations to treat the sign of a zero
1787 argument or result as insignificant.
1790 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1791 argument rather than perform division.
1794 Fast - Allow algebraically equivalent transformations that may
1795 dramatically change results in floating point (e.g. reassociate). This
1796 flag implies all the others.
1800 Use-list Order Directives
1801 -------------------------
1803 Use-list directives encode the in-memory order of each use-list, allowing the
1804 order to be recreated. ``<order-indexes>`` is a comma-separated list of
1805 indexes that are assigned to the referenced value's uses. The referenced
1806 value's use-list is immediately sorted by these indexes.
1808 Use-list directives may appear at function scope or global scope. They are not
1809 instructions, and have no effect on the semantics of the IR. When they're at
1810 function scope, they must appear after the terminator of the final basic block.
1812 If basic blocks have their address taken via ``blockaddress()`` expressions,
1813 ``uselistorder_bb`` can be used to reorder their use-lists from outside their
1820 uselistorder <ty> <value>, { <order-indexes> }
1821 uselistorder_bb @function, %block { <order-indexes> }
1827 define void @foo(i32 %arg1, i32 %arg2) {
1829 ; ... instructions ...
1831 ; ... instructions ...
1833 ; At function scope.
1834 uselistorder i32 %arg1, { 1, 0, 2 }
1835 uselistorder label %bb, { 1, 0 }
1839 uselistorder i32* @global, { 1, 2, 0 }
1840 uselistorder i32 7, { 1, 0 }
1841 uselistorder i32 (i32) @bar, { 1, 0 }
1842 uselistorder_bb @foo, %bb, { 5, 1, 3, 2, 0, 4 }
1849 The LLVM type system is one of the most important features of the
1850 intermediate representation. Being typed enables a number of
1851 optimizations to be performed on the intermediate representation
1852 directly, without having to do extra analyses on the side before the
1853 transformation. A strong type system makes it easier to read the
1854 generated code and enables novel analyses and transformations that are
1855 not feasible to perform on normal three address code representations.
1865 The void type does not represent any value and has no size.
1883 The function type can be thought of as a function signature. It consists of a
1884 return type and a list of formal parameter types. The return type of a function
1885 type is a void type or first class type --- except for :ref:`label <t_label>`
1886 and :ref:`metadata <t_metadata>` types.
1892 <returntype> (<parameter list>)
1894 ...where '``<parameter list>``' is a comma-separated list of type
1895 specifiers. Optionally, the parameter list may include a type ``...``, which
1896 indicates that the function takes a variable number of arguments. Variable
1897 argument functions can access their arguments with the :ref:`variable argument
1898 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
1899 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
1903 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1904 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1905 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1906 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1907 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1908 | ``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. |
1909 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1910 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1911 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1918 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1919 Values of these types are the only ones which can be produced by
1927 These are the types that are valid in registers from CodeGen's perspective.
1936 The integer type is a very simple type that simply specifies an
1937 arbitrary bit width for the integer type desired. Any bit width from 1
1938 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1946 The number of bits the integer will occupy is specified by the ``N``
1952 +----------------+------------------------------------------------+
1953 | ``i1`` | a single-bit integer. |
1954 +----------------+------------------------------------------------+
1955 | ``i32`` | a 32-bit integer. |
1956 +----------------+------------------------------------------------+
1957 | ``i1942652`` | a really big integer of over 1 million bits. |
1958 +----------------+------------------------------------------------+
1962 Floating Point Types
1963 """"""""""""""""""""
1972 - 16-bit floating point value
1975 - 32-bit floating point value
1978 - 64-bit floating point value
1981 - 128-bit floating point value (112-bit mantissa)
1984 - 80-bit floating point value (X87)
1987 - 128-bit floating point value (two 64-bits)
1994 The x86_mmx type represents a value held in an MMX register on an x86
1995 machine. The operations allowed on it are quite limited: parameters and
1996 return values, load and store, and bitcast. User-specified MMX
1997 instructions are represented as intrinsic or asm calls with arguments
1998 and/or results of this type. There are no arrays, vectors or constants
2015 The pointer type is used to specify memory locations. Pointers are
2016 commonly used to reference objects in memory.
2018 Pointer types may have an optional address space attribute defining the
2019 numbered address space where the pointed-to object resides. The default
2020 address space is number zero. The semantics of non-zero address spaces
2021 are target-specific.
2023 Note that LLVM does not permit pointers to void (``void*``) nor does it
2024 permit pointers to labels (``label*``). Use ``i8*`` instead.
2034 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2035 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
2036 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2037 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
2038 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2039 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
2040 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2049 A vector type is a simple derived type that represents a vector of
2050 elements. Vector types are used when multiple primitive data are
2051 operated in parallel using a single instruction (SIMD). A vector type
2052 requires a size (number of elements) and an underlying primitive data
2053 type. Vector types are considered :ref:`first class <t_firstclass>`.
2059 < <# elements> x <elementtype> >
2061 The number of elements is a constant integer value larger than 0;
2062 elementtype may be any integer, floating point or pointer type. Vectors
2063 of size zero are not allowed.
2067 +-------------------+--------------------------------------------------+
2068 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
2069 +-------------------+--------------------------------------------------+
2070 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
2071 +-------------------+--------------------------------------------------+
2072 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
2073 +-------------------+--------------------------------------------------+
2074 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
2075 +-------------------+--------------------------------------------------+
2084 The label type represents code labels.
2099 The metadata type represents embedded metadata. No derived types may be
2100 created from metadata except for :ref:`function <t_function>` arguments.
2113 Aggregate Types are a subset of derived types that can contain multiple
2114 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
2115 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
2125 The array type is a very simple derived type that arranges elements
2126 sequentially in memory. The array type requires a size (number of
2127 elements) and an underlying data type.
2133 [<# elements> x <elementtype>]
2135 The number of elements is a constant integer value; ``elementtype`` may
2136 be any type with a size.
2140 +------------------+--------------------------------------+
2141 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
2142 +------------------+--------------------------------------+
2143 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
2144 +------------------+--------------------------------------+
2145 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
2146 +------------------+--------------------------------------+
2148 Here are some examples of multidimensional arrays:
2150 +-----------------------------+----------------------------------------------------------+
2151 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
2152 +-----------------------------+----------------------------------------------------------+
2153 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
2154 +-----------------------------+----------------------------------------------------------+
2155 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
2156 +-----------------------------+----------------------------------------------------------+
2158 There is no restriction on indexing beyond the end of the array implied
2159 by a static type (though there are restrictions on indexing beyond the
2160 bounds of an allocated object in some cases). This means that
2161 single-dimension 'variable sized array' addressing can be implemented in
2162 LLVM with a zero length array type. An implementation of 'pascal style
2163 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
2173 The structure type is used to represent a collection of data members
2174 together in memory. The elements of a structure may be any type that has
2177 Structures in memory are accessed using '``load``' and '``store``' by
2178 getting a pointer to a field with the '``getelementptr``' instruction.
2179 Structures in registers are accessed using the '``extractvalue``' and
2180 '``insertvalue``' instructions.
2182 Structures may optionally be "packed" structures, which indicate that
2183 the alignment of the struct is one byte, and that there is no padding
2184 between the elements. In non-packed structs, padding between field types
2185 is inserted as defined by the DataLayout string in the module, which is
2186 required to match what the underlying code generator expects.
2188 Structures can either be "literal" or "identified". A literal structure
2189 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
2190 identified types are always defined at the top level with a name.
2191 Literal types are uniqued by their contents and can never be recursive
2192 or opaque since there is no way to write one. Identified types can be
2193 recursive, can be opaqued, and are never uniqued.
2199 %T1 = type { <type list> } ; Identified normal struct type
2200 %T2 = type <{ <type list> }> ; Identified packed struct type
2204 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2205 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
2206 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2207 | ``{ 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``. |
2208 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2209 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
2210 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2214 Opaque Structure Types
2215 """"""""""""""""""""""
2219 Opaque structure types are used to represent named structure types that
2220 do not have a body specified. This corresponds (for example) to the C
2221 notion of a forward declared structure.
2232 +--------------+-------------------+
2233 | ``opaque`` | An opaque type. |
2234 +--------------+-------------------+
2241 LLVM has several different basic types of constants. This section
2242 describes them all and their syntax.
2247 **Boolean constants**
2248 The two strings '``true``' and '``false``' are both valid constants
2250 **Integer constants**
2251 Standard integers (such as '4') are constants of the
2252 :ref:`integer <t_integer>` type. Negative numbers may be used with
2254 **Floating point constants**
2255 Floating point constants use standard decimal notation (e.g.
2256 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2257 hexadecimal notation (see below). The assembler requires the exact
2258 decimal value of a floating-point constant. For example, the
2259 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2260 decimal in binary. Floating point constants must have a :ref:`floating
2261 point <t_floating>` type.
2262 **Null pointer constants**
2263 The identifier '``null``' is recognized as a null pointer constant
2264 and must be of :ref:`pointer type <t_pointer>`.
2266 The one non-intuitive notation for constants is the hexadecimal form of
2267 floating point constants. For example, the form
2268 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2269 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2270 constants are required (and the only time that they are generated by the
2271 disassembler) is when a floating point constant must be emitted but it
2272 cannot be represented as a decimal floating point number in a reasonable
2273 number of digits. For example, NaN's, infinities, and other special
2274 values are represented in their IEEE hexadecimal format so that assembly
2275 and disassembly do not cause any bits to change in the constants.
2277 When using the hexadecimal form, constants of types half, float, and
2278 double are represented using the 16-digit form shown above (which
2279 matches the IEEE754 representation for double); half and float values
2280 must, however, be exactly representable as IEEE 754 half and single
2281 precision, respectively. Hexadecimal format is always used for long
2282 double, and there are three forms of long double. The 80-bit format used
2283 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2284 128-bit format used by PowerPC (two adjacent doubles) is represented by
2285 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2286 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2287 will only work if they match the long double format on your target.
2288 The IEEE 16-bit format (half precision) is represented by ``0xH``
2289 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2290 (sign bit at the left).
2292 There are no constants of type x86_mmx.
2294 .. _complexconstants:
2299 Complex constants are a (potentially recursive) combination of simple
2300 constants and smaller complex constants.
2302 **Structure constants**
2303 Structure constants are represented with notation similar to
2304 structure type definitions (a comma separated list of elements,
2305 surrounded by braces (``{}``)). For example:
2306 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2307 "``@G = external global i32``". Structure constants must have
2308 :ref:`structure type <t_struct>`, and the number and types of elements
2309 must match those specified by the type.
2311 Array constants are represented with notation similar to array type
2312 definitions (a comma separated list of elements, surrounded by
2313 square brackets (``[]``)). For example:
2314 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2315 :ref:`array type <t_array>`, and the number and types of elements must
2316 match those specified by the type. As a special case, character array
2317 constants may also be represented as a double-quoted string using the ``c``
2318 prefix. For example: "``c"Hello World\0A\00"``".
2319 **Vector constants**
2320 Vector constants are represented with notation similar to vector
2321 type definitions (a comma separated list of elements, surrounded by
2322 less-than/greater-than's (``<>``)). For example:
2323 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2324 must have :ref:`vector type <t_vector>`, and the number and types of
2325 elements must match those specified by the type.
2326 **Zero initialization**
2327 The string '``zeroinitializer``' can be used to zero initialize a
2328 value to zero of *any* type, including scalar and
2329 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2330 having to print large zero initializers (e.g. for large arrays) and
2331 is always exactly equivalent to using explicit zero initializers.
2333 A metadata node is a structure-like constant with :ref:`metadata
2334 type <t_metadata>`. For example:
2335 "``metadata !{ i32 0, metadata !"test" }``". Unlike other
2336 constants that are meant to be interpreted as part of the
2337 instruction stream, metadata is a place to attach additional
2338 information such as debug info.
2340 Global Variable and Function Addresses
2341 --------------------------------------
2343 The addresses of :ref:`global variables <globalvars>` and
2344 :ref:`functions <functionstructure>` are always implicitly valid
2345 (link-time) constants. These constants are explicitly referenced when
2346 the :ref:`identifier for the global <identifiers>` is used and always have
2347 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2350 .. code-block:: llvm
2354 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2361 The string '``undef``' can be used anywhere a constant is expected, and
2362 indicates that the user of the value may receive an unspecified
2363 bit-pattern. Undefined values may be of any type (other than '``label``'
2364 or '``void``') and be used anywhere a constant is permitted.
2366 Undefined values are useful because they indicate to the compiler that
2367 the program is well defined no matter what value is used. This gives the
2368 compiler more freedom to optimize. Here are some examples of
2369 (potentially surprising) transformations that are valid (in pseudo IR):
2371 .. code-block:: llvm
2381 This is safe because all of the output bits are affected by the undef
2382 bits. Any output bit can have a zero or one depending on the input bits.
2384 .. code-block:: llvm
2395 These logical operations have bits that are not always affected by the
2396 input. For example, if ``%X`` has a zero bit, then the output of the
2397 '``and``' operation will always be a zero for that bit, no matter what
2398 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2399 optimize or assume that the result of the '``and``' is '``undef``'.
2400 However, it is safe to assume that all bits of the '``undef``' could be
2401 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2402 all the bits of the '``undef``' operand to the '``or``' could be set,
2403 allowing the '``or``' to be folded to -1.
2405 .. code-block:: llvm
2407 %A = select undef, %X, %Y
2408 %B = select undef, 42, %Y
2409 %C = select %X, %Y, undef
2419 This set of examples shows that undefined '``select``' (and conditional
2420 branch) conditions can go *either way*, but they have to come from one
2421 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2422 both known to have a clear low bit, then ``%A`` would have to have a
2423 cleared low bit. However, in the ``%C`` example, the optimizer is
2424 allowed to assume that the '``undef``' operand could be the same as
2425 ``%Y``, allowing the whole '``select``' to be eliminated.
2427 .. code-block:: llvm
2429 %A = xor undef, undef
2446 This example points out that two '``undef``' operands are not
2447 necessarily the same. This can be surprising to people (and also matches
2448 C semantics) where they assume that "``X^X``" is always zero, even if
2449 ``X`` is undefined. This isn't true for a number of reasons, but the
2450 short answer is that an '``undef``' "variable" can arbitrarily change
2451 its value over its "live range". This is true because the variable
2452 doesn't actually *have a live range*. Instead, the value is logically
2453 read from arbitrary registers that happen to be around when needed, so
2454 the value is not necessarily consistent over time. In fact, ``%A`` and
2455 ``%C`` need to have the same semantics or the core LLVM "replace all
2456 uses with" concept would not hold.
2458 .. code-block:: llvm
2466 These examples show the crucial difference between an *undefined value*
2467 and *undefined behavior*. An undefined value (like '``undef``') is
2468 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2469 operation can be constant folded to '``undef``', because the '``undef``'
2470 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2471 However, in the second example, we can make a more aggressive
2472 assumption: because the ``undef`` is allowed to be an arbitrary value,
2473 we are allowed to assume that it could be zero. Since a divide by zero
2474 has *undefined behavior*, we are allowed to assume that the operation
2475 does not execute at all. This allows us to delete the divide and all
2476 code after it. Because the undefined operation "can't happen", the
2477 optimizer can assume that it occurs in dead code.
2479 .. code-block:: llvm
2481 a: store undef -> %X
2482 b: store %X -> undef
2487 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2488 value can be assumed to not have any effect; we can assume that the
2489 value is overwritten with bits that happen to match what was already
2490 there. However, a store *to* an undefined location could clobber
2491 arbitrary memory, therefore, it has undefined behavior.
2498 Poison values are similar to :ref:`undef values <undefvalues>`, however
2499 they also represent the fact that an instruction or constant expression
2500 that cannot evoke side effects has nevertheless detected a condition
2501 that results in undefined behavior.
2503 There is currently no way of representing a poison value in the IR; they
2504 only exist when produced by operations such as :ref:`add <i_add>` with
2507 Poison value behavior is defined in terms of value *dependence*:
2509 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2510 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2511 their dynamic predecessor basic block.
2512 - Function arguments depend on the corresponding actual argument values
2513 in the dynamic callers of their functions.
2514 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2515 instructions that dynamically transfer control back to them.
2516 - :ref:`Invoke <i_invoke>` instructions depend on the
2517 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2518 call instructions that dynamically transfer control back to them.
2519 - Non-volatile loads and stores depend on the most recent stores to all
2520 of the referenced memory addresses, following the order in the IR
2521 (including loads and stores implied by intrinsics such as
2522 :ref:`@llvm.memcpy <int_memcpy>`.)
2523 - An instruction with externally visible side effects depends on the
2524 most recent preceding instruction with externally visible side
2525 effects, following the order in the IR. (This includes :ref:`volatile
2526 operations <volatile>`.)
2527 - An instruction *control-depends* on a :ref:`terminator
2528 instruction <terminators>` if the terminator instruction has
2529 multiple successors and the instruction is always executed when
2530 control transfers to one of the successors, and may not be executed
2531 when control is transferred to another.
2532 - Additionally, an instruction also *control-depends* on a terminator
2533 instruction if the set of instructions it otherwise depends on would
2534 be different if the terminator had transferred control to a different
2536 - Dependence is transitive.
2538 Poison values have the same behavior as :ref:`undef values <undefvalues>`,
2539 with the additional effect that any instruction that has a *dependence*
2540 on a poison value has undefined behavior.
2542 Here are some examples:
2544 .. code-block:: llvm
2547 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2548 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2549 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
2550 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2552 store i32 %poison, i32* @g ; Poison value stored to memory.
2553 %poison2 = load i32* @g ; Poison value loaded back from memory.
2555 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2557 %narrowaddr = bitcast i32* @g to i16*
2558 %wideaddr = bitcast i32* @g to i64*
2559 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2560 %poison4 = load i64* %wideaddr ; Returns a poison value.
2562 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2563 br i1 %cmp, label %true, label %end ; Branch to either destination.
2566 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2567 ; it has undefined behavior.
2571 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2572 ; Both edges into this PHI are
2573 ; control-dependent on %cmp, so this
2574 ; always results in a poison value.
2576 store volatile i32 0, i32* @g ; This would depend on the store in %true
2577 ; if %cmp is true, or the store in %entry
2578 ; otherwise, so this is undefined behavior.
2580 br i1 %cmp, label %second_true, label %second_end
2581 ; The same branch again, but this time the
2582 ; true block doesn't have side effects.
2589 store volatile i32 0, i32* @g ; This time, the instruction always depends
2590 ; on the store in %end. Also, it is
2591 ; control-equivalent to %end, so this is
2592 ; well-defined (ignoring earlier undefined
2593 ; behavior in this example).
2597 Addresses of Basic Blocks
2598 -------------------------
2600 ``blockaddress(@function, %block)``
2602 The '``blockaddress``' constant computes the address of the specified
2603 basic block in the specified function, and always has an ``i8*`` type.
2604 Taking the address of the entry block is illegal.
2606 This value only has defined behavior when used as an operand to the
2607 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2608 against null. Pointer equality tests between labels addresses results in
2609 undefined behavior --- though, again, comparison against null is ok, and
2610 no label is equal to the null pointer. This may be passed around as an
2611 opaque pointer sized value as long as the bits are not inspected. This
2612 allows ``ptrtoint`` and arithmetic to be performed on these values so
2613 long as the original value is reconstituted before the ``indirectbr``
2616 Finally, some targets may provide defined semantics when using the value
2617 as the operand to an inline assembly, but that is target specific.
2621 Constant Expressions
2622 --------------------
2624 Constant expressions are used to allow expressions involving other
2625 constants to be used as constants. Constant expressions may be of any
2626 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2627 that does not have side effects (e.g. load and call are not supported).
2628 The following is the syntax for constant expressions:
2630 ``trunc (CST to TYPE)``
2631 Truncate a constant to another type. The bit size of CST must be
2632 larger than the bit size of TYPE. Both types must be integers.
2633 ``zext (CST to TYPE)``
2634 Zero extend a constant to another type. The bit size of CST must be
2635 smaller than the bit size of TYPE. Both types must be integers.
2636 ``sext (CST to TYPE)``
2637 Sign extend a constant to another type. The bit size of CST must be
2638 smaller than the bit size of TYPE. Both types must be integers.
2639 ``fptrunc (CST to TYPE)``
2640 Truncate a floating point constant to another floating point type.
2641 The size of CST must be larger than the size of TYPE. Both types
2642 must be floating point.
2643 ``fpext (CST to TYPE)``
2644 Floating point extend a constant to another type. The size of CST
2645 must be smaller or equal to the size of TYPE. Both types must be
2647 ``fptoui (CST to TYPE)``
2648 Convert a floating point constant to the corresponding unsigned
2649 integer constant. TYPE must be a scalar or vector integer type. CST
2650 must be of scalar or vector floating point type. Both CST and TYPE
2651 must be scalars, or vectors of the same number of elements. If the
2652 value won't fit in the integer type, the results are undefined.
2653 ``fptosi (CST to TYPE)``
2654 Convert a floating point constant to the corresponding signed
2655 integer constant. TYPE must be a scalar or vector integer type. CST
2656 must be of scalar or vector floating point type. Both CST and TYPE
2657 must be scalars, or vectors of the same number of elements. If the
2658 value won't fit in the integer type, the results are undefined.
2659 ``uitofp (CST to TYPE)``
2660 Convert an unsigned integer constant to the corresponding floating
2661 point constant. TYPE must be a scalar or vector floating point type.
2662 CST must be of scalar or vector integer type. Both CST and TYPE must
2663 be scalars, or vectors of the same number of elements. If the value
2664 won't fit in the floating point type, the results are undefined.
2665 ``sitofp (CST to TYPE)``
2666 Convert a signed integer constant to the corresponding floating
2667 point constant. TYPE must be a scalar or vector floating point type.
2668 CST must be of scalar or vector integer type. Both CST and TYPE must
2669 be scalars, or vectors of the same number of elements. If the value
2670 won't fit in the floating point type, the results are undefined.
2671 ``ptrtoint (CST to TYPE)``
2672 Convert a pointer typed constant to the corresponding integer
2673 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2674 pointer type. The ``CST`` value is zero extended, truncated, or
2675 unchanged to make it fit in ``TYPE``.
2676 ``inttoptr (CST to TYPE)``
2677 Convert an integer constant to a pointer constant. TYPE must be a
2678 pointer type. CST must be of integer type. The CST value is zero
2679 extended, truncated, or unchanged to make it fit in a pointer size.
2680 This one is *really* dangerous!
2681 ``bitcast (CST to TYPE)``
2682 Convert a constant, CST, to another TYPE. The constraints of the
2683 operands are the same as those for the :ref:`bitcast
2684 instruction <i_bitcast>`.
2685 ``addrspacecast (CST to TYPE)``
2686 Convert a constant pointer or constant vector of pointer, CST, to another
2687 TYPE in a different address space. The constraints of the operands are the
2688 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2689 ``getelementptr (CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)``
2690 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2691 constants. As with the :ref:`getelementptr <i_getelementptr>`
2692 instruction, the index list may have zero or more indexes, which are
2693 required to make sense for the type of "CSTPTR".
2694 ``select (COND, VAL1, VAL2)``
2695 Perform the :ref:`select operation <i_select>` on constants.
2696 ``icmp COND (VAL1, VAL2)``
2697 Performs the :ref:`icmp operation <i_icmp>` on constants.
2698 ``fcmp COND (VAL1, VAL2)``
2699 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2700 ``extractelement (VAL, IDX)``
2701 Perform the :ref:`extractelement operation <i_extractelement>` on
2703 ``insertelement (VAL, ELT, IDX)``
2704 Perform the :ref:`insertelement operation <i_insertelement>` on
2706 ``shufflevector (VEC1, VEC2, IDXMASK)``
2707 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2709 ``extractvalue (VAL, IDX0, IDX1, ...)``
2710 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2711 constants. The index list is interpreted in a similar manner as
2712 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2713 least one index value must be specified.
2714 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2715 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2716 The index list is interpreted in a similar manner as indices in a
2717 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2718 value must be specified.
2719 ``OPCODE (LHS, RHS)``
2720 Perform the specified operation of the LHS and RHS constants. OPCODE
2721 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2722 binary <bitwiseops>` operations. The constraints on operands are
2723 the same as those for the corresponding instruction (e.g. no bitwise
2724 operations on floating point values are allowed).
2731 Inline Assembler Expressions
2732 ----------------------------
2734 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2735 Inline Assembly <moduleasm>`) through the use of a special value. This
2736 value represents the inline assembler as a string (containing the
2737 instructions to emit), a list of operand constraints (stored as a
2738 string), a flag that indicates whether or not the inline asm expression
2739 has side effects, and a flag indicating whether the function containing
2740 the asm needs to align its stack conservatively. An example inline
2741 assembler expression is:
2743 .. code-block:: llvm
2745 i32 (i32) asm "bswap $0", "=r,r"
2747 Inline assembler expressions may **only** be used as the callee operand
2748 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2749 Thus, typically we have:
2751 .. code-block:: llvm
2753 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2755 Inline asms with side effects not visible in the constraint list must be
2756 marked as having side effects. This is done through the use of the
2757 '``sideeffect``' keyword, like so:
2759 .. code-block:: llvm
2761 call void asm sideeffect "eieio", ""()
2763 In some cases inline asms will contain code that will not work unless
2764 the stack is aligned in some way, such as calls or SSE instructions on
2765 x86, yet will not contain code that does that alignment within the asm.
2766 The compiler should make conservative assumptions about what the asm
2767 might contain and should generate its usual stack alignment code in the
2768 prologue if the '``alignstack``' keyword is present:
2770 .. code-block:: llvm
2772 call void asm alignstack "eieio", ""()
2774 Inline asms also support using non-standard assembly dialects. The
2775 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2776 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2777 the only supported dialects. An example is:
2779 .. code-block:: llvm
2781 call void asm inteldialect "eieio", ""()
2783 If multiple keywords appear the '``sideeffect``' keyword must come
2784 first, the '``alignstack``' keyword second and the '``inteldialect``'
2790 The call instructions that wrap inline asm nodes may have a
2791 "``!srcloc``" MDNode attached to it that contains a list of constant
2792 integers. If present, the code generator will use the integer as the
2793 location cookie value when report errors through the ``LLVMContext``
2794 error reporting mechanisms. This allows a front-end to correlate backend
2795 errors that occur with inline asm back to the source code that produced
2798 .. code-block:: llvm
2800 call void asm sideeffect "something bad", ""(), !srcloc !42
2802 !42 = !{ i32 1234567 }
2804 It is up to the front-end to make sense of the magic numbers it places
2805 in the IR. If the MDNode contains multiple constants, the code generator
2806 will use the one that corresponds to the line of the asm that the error
2811 Metadata Nodes and Metadata Strings
2812 -----------------------------------
2814 LLVM IR allows metadata to be attached to instructions in the program
2815 that can convey extra information about the code to the optimizers and
2816 code generator. One example application of metadata is source-level
2817 debug information. There are two metadata primitives: strings and nodes.
2818 All metadata has the ``metadata`` type and is identified in syntax by a
2819 preceding exclamation point ('``!``').
2821 A metadata string is a string surrounded by double quotes. It can
2822 contain any character by escaping non-printable characters with
2823 "``\xx``" where "``xx``" is the two digit hex code. For example:
2826 Metadata nodes are represented with notation similar to structure
2827 constants (a comma separated list of elements, surrounded by braces and
2828 preceded by an exclamation point). Metadata nodes can have any values as
2829 their operand. For example:
2831 .. code-block:: llvm
2833 !{ metadata !"test\00", i32 10}
2835 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2836 metadata nodes, which can be looked up in the module symbol table. For
2839 .. code-block:: llvm
2841 !foo = metadata !{!4, !3}
2843 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2844 function is using two metadata arguments:
2846 .. code-block:: llvm
2848 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2850 Metadata can be attached with an instruction. Here metadata ``!21`` is
2851 attached to the ``add`` instruction using the ``!dbg`` identifier:
2853 .. code-block:: llvm
2855 %indvar.next = add i64 %indvar, 1, !dbg !21
2857 More information about specific metadata nodes recognized by the
2858 optimizers and code generator is found below.
2863 In LLVM IR, memory does not have types, so LLVM's own type system is not
2864 suitable for doing TBAA. Instead, metadata is added to the IR to
2865 describe a type system of a higher level language. This can be used to
2866 implement typical C/C++ TBAA, but it can also be used to implement
2867 custom alias analysis behavior for other languages.
2869 The current metadata format is very simple. TBAA metadata nodes have up
2870 to three fields, e.g.:
2872 .. code-block:: llvm
2874 !0 = metadata !{ metadata !"an example type tree" }
2875 !1 = metadata !{ metadata !"int", metadata !0 }
2876 !2 = metadata !{ metadata !"float", metadata !0 }
2877 !3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2879 The first field is an identity field. It can be any value, usually a
2880 metadata string, which uniquely identifies the type. The most important
2881 name in the tree is the name of the root node. Two trees with different
2882 root node names are entirely disjoint, even if they have leaves with
2885 The second field identifies the type's parent node in the tree, or is
2886 null or omitted for a root node. A type is considered to alias all of
2887 its descendants and all of its ancestors in the tree. Also, a type is
2888 considered to alias all types in other trees, so that bitcode produced
2889 from multiple front-ends is handled conservatively.
2891 If the third field is present, it's an integer which if equal to 1
2892 indicates that the type is "constant" (meaning
2893 ``pointsToConstantMemory`` should return true; see `other useful
2894 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
2896 '``tbaa.struct``' Metadata
2897 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2899 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
2900 aggregate assignment operations in C and similar languages, however it
2901 is defined to copy a contiguous region of memory, which is more than
2902 strictly necessary for aggregate types which contain holes due to
2903 padding. Also, it doesn't contain any TBAA information about the fields
2906 ``!tbaa.struct`` metadata can describe which memory subregions in a
2907 memcpy are padding and what the TBAA tags of the struct are.
2909 The current metadata format is very simple. ``!tbaa.struct`` metadata
2910 nodes are a list of operands which are in conceptual groups of three.
2911 For each group of three, the first operand gives the byte offset of a
2912 field in bytes, the second gives its size in bytes, and the third gives
2915 .. code-block:: llvm
2917 !4 = metadata !{ i64 0, i64 4, metadata !1, i64 8, i64 4, metadata !2 }
2919 This describes a struct with two fields. The first is at offset 0 bytes
2920 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
2921 and has size 4 bytes and has tbaa tag !2.
2923 Note that the fields need not be contiguous. In this example, there is a
2924 4 byte gap between the two fields. This gap represents padding which
2925 does not carry useful data and need not be preserved.
2927 '``noalias``' and '``alias.scope``' Metadata
2928 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2930 ``noalias`` and ``alias.scope`` metadata provide the ability to specify generic
2931 noalias memory-access sets. This means that some collection of memory access
2932 instructions (loads, stores, memory-accessing calls, etc.) that carry
2933 ``noalias`` metadata can specifically be specified not to alias with some other
2934 collection of memory access instructions that carry ``alias.scope`` metadata.
2935 Each type of metadata specifies a list of scopes where each scope has an id and
2936 a domain. When evaluating an aliasing query, if for some some domain, the set
2937 of scopes with that domain in one instruction's ``alias.scope`` list is a
2938 subset of (or qual to) the set of scopes for that domain in another
2939 instruction's ``noalias`` list, then the two memory accesses are assumed not to
2942 The metadata identifying each domain is itself a list containing one or two
2943 entries. The first entry is the name of the domain. Note that if the name is a
2944 string then it can be combined accross functions and translation units. A
2945 self-reference can be used to create globally unique domain names. A
2946 descriptive string may optionally be provided as a second list entry.
2948 The metadata identifying each scope is also itself a list containing two or
2949 three entries. The first entry is the name of the scope. Note that if the name
2950 is a string then it can be combined accross functions and translation units. A
2951 self-reference can be used to create globally unique scope names. A metadata
2952 reference to the scope's domain is the second entry. A descriptive string may
2953 optionally be provided as a third list entry.
2957 .. code-block:: llvm
2959 ; Two scope domains:
2960 !0 = metadata !{metadata !0}
2961 !1 = metadata !{metadata !1}
2963 ; Some scopes in these domains:
2964 !2 = metadata !{metadata !2, metadata !0}
2965 !3 = metadata !{metadata !3, metadata !0}
2966 !4 = metadata !{metadata !4, metadata !1}
2969 !5 = metadata !{metadata !4} ; A list containing only scope !4
2970 !6 = metadata !{metadata !4, metadata !3, metadata !2}
2971 !7 = metadata !{metadata !3}
2973 ; These two instructions don't alias:
2974 %0 = load float* %c, align 4, !alias.scope !5
2975 store float %0, float* %arrayidx.i, align 4, !noalias !5
2977 ; These two instructions also don't alias (for domain !1, the set of scopes
2978 ; in the !alias.scope equals that in the !noalias list):
2979 %2 = load float* %c, align 4, !alias.scope !5
2980 store float %2, float* %arrayidx.i2, align 4, !noalias !6
2982 ; These two instructions don't alias (for domain !0, the set of scopes in
2983 ; the !noalias list is not a superset of, or equal to, the scopes in the
2984 ; !alias.scope list):
2985 %2 = load float* %c, align 4, !alias.scope !6
2986 store float %0, float* %arrayidx.i, align 4, !noalias !7
2988 '``fpmath``' Metadata
2989 ^^^^^^^^^^^^^^^^^^^^^
2991 ``fpmath`` metadata may be attached to any instruction of floating point
2992 type. It can be used to express the maximum acceptable error in the
2993 result of that instruction, in ULPs, thus potentially allowing the
2994 compiler to use a more efficient but less accurate method of computing
2995 it. ULP is defined as follows:
2997 If ``x`` is a real number that lies between two finite consecutive
2998 floating-point numbers ``a`` and ``b``, without being equal to one
2999 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
3000 distance between the two non-equal finite floating-point numbers
3001 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
3003 The metadata node shall consist of a single positive floating point
3004 number representing the maximum relative error, for example:
3006 .. code-block:: llvm
3008 !0 = metadata !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
3010 '``range``' Metadata
3011 ^^^^^^^^^^^^^^^^^^^^
3013 ``range`` metadata may be attached only to ``load``, ``call`` and ``invoke`` of
3014 integer types. It expresses the possible ranges the loaded value or the value
3015 returned by the called function at this call site is in. The ranges are
3016 represented with a flattened list of integers. The loaded value or the value
3017 returned is known to be in the union of the ranges defined by each consecutive
3018 pair. Each pair has the following properties:
3020 - The type must match the type loaded by the instruction.
3021 - The pair ``a,b`` represents the range ``[a,b)``.
3022 - Both ``a`` and ``b`` are constants.
3023 - The range is allowed to wrap.
3024 - The range should not represent the full or empty set. That is,
3027 In addition, the pairs must be in signed order of the lower bound and
3028 they must be non-contiguous.
3032 .. code-block:: llvm
3034 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
3035 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
3036 %c = call i8 @foo(), !range !2 ; Can only be 0, 1, 3, 4 or 5
3037 %d = invoke i8 @bar() to label %cont
3038 unwind label %lpad, !range !3 ; Can only be -2, -1, 3, 4 or 5
3040 !0 = metadata !{ i8 0, i8 2 }
3041 !1 = metadata !{ i8 255, i8 2 }
3042 !2 = metadata !{ i8 0, i8 2, i8 3, i8 6 }
3043 !3 = metadata !{ i8 -2, i8 0, i8 3, i8 6 }
3048 It is sometimes useful to attach information to loop constructs. Currently,
3049 loop metadata is implemented as metadata attached to the branch instruction
3050 in the loop latch block. This type of metadata refer to a metadata node that is
3051 guaranteed to be separate for each loop. The loop identifier metadata is
3052 specified with the name ``llvm.loop``.
3054 The loop identifier metadata is implemented using a metadata that refers to
3055 itself to avoid merging it with any other identifier metadata, e.g.,
3056 during module linkage or function inlining. That is, each loop should refer
3057 to their own identification metadata even if they reside in separate functions.
3058 The following example contains loop identifier metadata for two separate loop
3061 .. code-block:: llvm
3063 !0 = metadata !{ metadata !0 }
3064 !1 = metadata !{ metadata !1 }
3066 The loop identifier metadata can be used to specify additional
3067 per-loop metadata. Any operands after the first operand can be treated
3068 as user-defined metadata. For example the ``llvm.loop.unroll.count``
3069 suggests an unroll factor to the loop unroller:
3071 .. code-block:: llvm
3073 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
3075 !0 = metadata !{ metadata !0, metadata !1 }
3076 !1 = metadata !{ metadata !"llvm.loop.unroll.count", i32 4 }
3078 '``llvm.loop.vectorize``' and '``llvm.loop.interleave``'
3079 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3081 Metadata prefixed with ``llvm.loop.vectorize`` or ``llvm.loop.interleave`` are
3082 used to control per-loop vectorization and interleaving parameters such as
3083 vectorization width and interleave count. These metadata should be used in
3084 conjunction with ``llvm.loop`` loop identification metadata. The
3085 ``llvm.loop.vectorize`` and ``llvm.loop.interleave`` metadata are only
3086 optimization hints and the optimizer will only interleave and vectorize loops if
3087 it believes it is safe to do so. The ``llvm.mem.parallel_loop_access`` metadata
3088 which contains information about loop-carried memory dependencies can be helpful
3089 in determining the safety of these transformations.
3091 '``llvm.loop.interleave.count``' Metadata
3092 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3094 This metadata suggests an interleave count to the loop interleaver.
3095 The first operand is the string ``llvm.loop.interleave.count`` and the
3096 second operand is an integer specifying the interleave count. For
3099 .. code-block:: llvm
3101 !0 = metadata !{ metadata !"llvm.loop.interleave.count", i32 4 }
3103 Note that setting ``llvm.loop.interleave.count`` to 1 disables interleaving
3104 multiple iterations of the loop. If ``llvm.loop.interleave.count`` is set to 0
3105 then the interleave count will be determined automatically.
3107 '``llvm.loop.vectorize.enable``' Metadata
3108 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3110 This metadata selectively enables or disables vectorization for the loop. The
3111 first operand is the string ``llvm.loop.vectorize.enable`` and the second operand
3112 is a bit. If the bit operand value is 1 vectorization is enabled. A value of
3113 0 disables vectorization:
3115 .. code-block:: llvm
3117 !0 = metadata !{ metadata !"llvm.loop.vectorize.enable", i1 0 }
3118 !1 = metadata !{ metadata !"llvm.loop.vectorize.enable", i1 1 }
3120 '``llvm.loop.vectorize.width``' Metadata
3121 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3123 This metadata sets the target width of the vectorizer. The first
3124 operand is the string ``llvm.loop.vectorize.width`` and the second
3125 operand is an integer specifying the width. For example:
3127 .. code-block:: llvm
3129 !0 = metadata !{ metadata !"llvm.loop.vectorize.width", i32 4 }
3131 Note that setting ``llvm.loop.vectorize.width`` to 1 disables
3132 vectorization of the loop. If ``llvm.loop.vectorize.width`` is set to
3133 0 or if the loop does not have this metadata the width will be
3134 determined automatically.
3136 '``llvm.loop.unroll``'
3137 ^^^^^^^^^^^^^^^^^^^^^^
3139 Metadata prefixed with ``llvm.loop.unroll`` are loop unrolling
3140 optimization hints such as the unroll factor. ``llvm.loop.unroll``
3141 metadata should be used in conjunction with ``llvm.loop`` loop
3142 identification metadata. The ``llvm.loop.unroll`` metadata are only
3143 optimization hints and the unrolling will only be performed if the
3144 optimizer believes it is safe to do so.
3146 '``llvm.loop.unroll.count``' Metadata
3147 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3149 This metadata suggests an unroll factor to the loop unroller. The
3150 first operand is the string ``llvm.loop.unroll.count`` and the second
3151 operand is a positive integer specifying the unroll factor. For
3154 .. code-block:: llvm
3156 !0 = metadata !{ metadata !"llvm.loop.unroll.count", i32 4 }
3158 If the trip count of the loop is less than the unroll count the loop
3159 will be partially unrolled.
3161 '``llvm.loop.unroll.disable``' Metadata
3162 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3164 This metadata either disables loop unrolling. The metadata has a single operand
3165 which is the string ``llvm.loop.unroll.disable``. For example:
3167 .. code-block:: llvm
3169 !0 = metadata !{ metadata !"llvm.loop.unroll.disable" }
3171 '``llvm.loop.unroll.full``' Metadata
3172 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3174 This metadata either suggests that the loop should be unrolled fully. The
3175 metadata has a single operand which is the string ``llvm.loop.unroll.disable``.
3178 .. code-block:: llvm
3180 !0 = metadata !{ metadata !"llvm.loop.unroll.full" }
3185 Metadata types used to annotate memory accesses with information helpful
3186 for optimizations are prefixed with ``llvm.mem``.
3188 '``llvm.mem.parallel_loop_access``' Metadata
3189 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3191 The ``llvm.mem.parallel_loop_access`` metadata refers to a loop identifier,
3192 or metadata containing a list of loop identifiers for nested loops.
3193 The metadata is attached to memory accessing instructions and denotes that
3194 no loop carried memory dependence exist between it and other instructions denoted
3195 with the same loop identifier.
3197 Precisely, given two instructions ``m1`` and ``m2`` that both have the
3198 ``llvm.mem.parallel_loop_access`` metadata, with ``L1`` and ``L2`` being the
3199 set of loops associated with that metadata, respectively, then there is no loop
3200 carried dependence between ``m1`` and ``m2`` for loops in both ``L1`` and
3203 As a special case, if all memory accessing instructions in a loop have
3204 ``llvm.mem.parallel_loop_access`` metadata that refers to that loop, then the
3205 loop has no loop carried memory dependences and is considered to be a parallel
3208 Note that if not all memory access instructions have such metadata referring to
3209 the loop, then the loop is considered not being trivially parallel. Additional
3210 memory dependence analysis is required to make that determination. As a fail
3211 safe mechanism, this causes loops that were originally parallel to be considered
3212 sequential (if optimization passes that are unaware of the parallel semantics
3213 insert new memory instructions into the loop body).
3215 Example of a loop that is considered parallel due to its correct use of
3216 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
3217 metadata types that refer to the same loop identifier metadata.
3219 .. code-block:: llvm
3223 %val0 = load i32* %arrayidx, !llvm.mem.parallel_loop_access !0
3225 store i32 %val0, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3227 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
3231 !0 = metadata !{ metadata !0 }
3233 It is also possible to have nested parallel loops. In that case the
3234 memory accesses refer to a list of loop identifier metadata nodes instead of
3235 the loop identifier metadata node directly:
3237 .. code-block:: llvm
3241 %val1 = load i32* %arrayidx3, !llvm.mem.parallel_loop_access !2
3243 br label %inner.for.body
3247 %val0 = load i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3249 store i32 %val0, i32* %arrayidx2, !llvm.mem.parallel_loop_access !0
3251 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
3255 store i32 %val1, i32* %arrayidx4, !llvm.mem.parallel_loop_access !2
3257 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
3259 outer.for.end: ; preds = %for.body
3261 !0 = metadata !{ metadata !1, metadata !2 } ; a list of loop identifiers
3262 !1 = metadata !{ metadata !1 } ; an identifier for the inner loop
3263 !2 = metadata !{ metadata !2 } ; an identifier for the outer loop
3265 Module Flags Metadata
3266 =====================
3268 Information about the module as a whole is difficult to convey to LLVM's
3269 subsystems. The LLVM IR isn't sufficient to transmit this information.
3270 The ``llvm.module.flags`` named metadata exists in order to facilitate
3271 this. These flags are in the form of key / value pairs --- much like a
3272 dictionary --- making it easy for any subsystem who cares about a flag to
3275 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
3276 Each triplet has the following form:
3278 - The first element is a *behavior* flag, which specifies the behavior
3279 when two (or more) modules are merged together, and it encounters two
3280 (or more) metadata with the same ID. The supported behaviors are
3282 - The second element is a metadata string that is a unique ID for the
3283 metadata. Each module may only have one flag entry for each unique ID (not
3284 including entries with the **Require** behavior).
3285 - The third element is the value of the flag.
3287 When two (or more) modules are merged together, the resulting
3288 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
3289 each unique metadata ID string, there will be exactly one entry in the merged
3290 modules ``llvm.module.flags`` metadata table, and the value for that entry will
3291 be determined by the merge behavior flag, as described below. The only exception
3292 is that entries with the *Require* behavior are always preserved.
3294 The following behaviors are supported:
3305 Emits an error if two values disagree, otherwise the resulting value
3306 is that of the operands.
3310 Emits a warning if two values disagree. The result value will be the
3311 operand for the flag from the first module being linked.
3315 Adds a requirement that another module flag be present and have a
3316 specified value after linking is performed. The value must be a
3317 metadata pair, where the first element of the pair is the ID of the
3318 module flag to be restricted, and the second element of the pair is
3319 the value the module flag should be restricted to. This behavior can
3320 be used to restrict the allowable results (via triggering of an
3321 error) of linking IDs with the **Override** behavior.
3325 Uses the specified value, regardless of the behavior or value of the
3326 other module. If both modules specify **Override**, but the values
3327 differ, an error will be emitted.
3331 Appends the two values, which are required to be metadata nodes.
3335 Appends the two values, which are required to be metadata
3336 nodes. However, duplicate entries in the second list are dropped
3337 during the append operation.
3339 It is an error for a particular unique flag ID to have multiple behaviors,
3340 except in the case of **Require** (which adds restrictions on another metadata
3341 value) or **Override**.
3343 An example of module flags:
3345 .. code-block:: llvm
3347 !0 = metadata !{ i32 1, metadata !"foo", i32 1 }
3348 !1 = metadata !{ i32 4, metadata !"bar", i32 37 }
3349 !2 = metadata !{ i32 2, metadata !"qux", i32 42 }
3350 !3 = metadata !{ i32 3, metadata !"qux",
3352 metadata !"foo", i32 1
3355 !llvm.module.flags = !{ !0, !1, !2, !3 }
3357 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
3358 if two or more ``!"foo"`` flags are seen is to emit an error if their
3359 values are not equal.
3361 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
3362 behavior if two or more ``!"bar"`` flags are seen is to use the value
3365 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
3366 behavior if two or more ``!"qux"`` flags are seen is to emit a
3367 warning if their values are not equal.
3369 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
3373 metadata !{ metadata !"foo", i32 1 }
3375 The behavior is to emit an error if the ``llvm.module.flags`` does not
3376 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
3379 Objective-C Garbage Collection Module Flags Metadata
3380 ----------------------------------------------------
3382 On the Mach-O platform, Objective-C stores metadata about garbage
3383 collection in a special section called "image info". The metadata
3384 consists of a version number and a bitmask specifying what types of
3385 garbage collection are supported (if any) by the file. If two or more
3386 modules are linked together their garbage collection metadata needs to
3387 be merged rather than appended together.
3389 The Objective-C garbage collection module flags metadata consists of the
3390 following key-value pairs:
3399 * - ``Objective-C Version``
3400 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
3402 * - ``Objective-C Image Info Version``
3403 - **[Required]** --- The version of the image info section. Currently
3406 * - ``Objective-C Image Info Section``
3407 - **[Required]** --- The section to place the metadata. Valid values are
3408 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
3409 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
3410 Objective-C ABI version 2.
3412 * - ``Objective-C Garbage Collection``
3413 - **[Required]** --- Specifies whether garbage collection is supported or
3414 not. Valid values are 0, for no garbage collection, and 2, for garbage
3415 collection supported.
3417 * - ``Objective-C GC Only``
3418 - **[Optional]** --- Specifies that only garbage collection is supported.
3419 If present, its value must be 6. This flag requires that the
3420 ``Objective-C Garbage Collection`` flag have the value 2.
3422 Some important flag interactions:
3424 - If a module with ``Objective-C Garbage Collection`` set to 0 is
3425 merged with a module with ``Objective-C Garbage Collection`` set to
3426 2, then the resulting module has the
3427 ``Objective-C Garbage Collection`` flag set to 0.
3428 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
3429 merged with a module with ``Objective-C GC Only`` set to 6.
3431 Automatic Linker Flags Module Flags Metadata
3432 --------------------------------------------
3434 Some targets support embedding flags to the linker inside individual object
3435 files. Typically this is used in conjunction with language extensions which
3436 allow source files to explicitly declare the libraries they depend on, and have
3437 these automatically be transmitted to the linker via object files.
3439 These flags are encoded in the IR using metadata in the module flags section,
3440 using the ``Linker Options`` key. The merge behavior for this flag is required
3441 to be ``AppendUnique``, and the value for the key is expected to be a metadata
3442 node which should be a list of other metadata nodes, each of which should be a
3443 list of metadata strings defining linker options.
3445 For example, the following metadata section specifies two separate sets of
3446 linker options, presumably to link against ``libz`` and the ``Cocoa``
3449 !0 = metadata !{ i32 6, metadata !"Linker Options",
3451 metadata !{ metadata !"-lz" },
3452 metadata !{ metadata !"-framework", metadata !"Cocoa" } } }
3453 !llvm.module.flags = !{ !0 }
3455 The metadata encoding as lists of lists of options, as opposed to a collapsed
3456 list of options, is chosen so that the IR encoding can use multiple option
3457 strings to specify e.g., a single library, while still having that specifier be
3458 preserved as an atomic element that can be recognized by a target specific
3459 assembly writer or object file emitter.
3461 Each individual option is required to be either a valid option for the target's
3462 linker, or an option that is reserved by the target specific assembly writer or
3463 object file emitter. No other aspect of these options is defined by the IR.
3465 C type width Module Flags Metadata
3466 ----------------------------------
3468 The ARM backend emits a section into each generated object file describing the
3469 options that it was compiled with (in a compiler-independent way) to prevent
3470 linking incompatible objects, and to allow automatic library selection. Some
3471 of these options are not visible at the IR level, namely wchar_t width and enum
3474 To pass this information to the backend, these options are encoded in module
3475 flags metadata, using the following key-value pairs:
3485 - * 0 --- sizeof(wchar_t) == 4
3486 * 1 --- sizeof(wchar_t) == 2
3489 - * 0 --- Enums are at least as large as an ``int``.
3490 * 1 --- Enums are stored in the smallest integer type which can
3491 represent all of its values.
3493 For example, the following metadata section specifies that the module was
3494 compiled with a ``wchar_t`` width of 4 bytes, and the underlying type of an
3495 enum is the smallest type which can represent all of its values::
3497 !llvm.module.flags = !{!0, !1}
3498 !0 = metadata !{i32 1, metadata !"short_wchar", i32 1}
3499 !1 = metadata !{i32 1, metadata !"short_enum", i32 0}
3501 .. _intrinsicglobalvariables:
3503 Intrinsic Global Variables
3504 ==========================
3506 LLVM has a number of "magic" global variables that contain data that
3507 affect code generation or other IR semantics. These are documented here.
3508 All globals of this sort should have a section specified as
3509 "``llvm.metadata``". This section and all globals that start with
3510 "``llvm.``" are reserved for use by LLVM.
3514 The '``llvm.used``' Global Variable
3515 -----------------------------------
3517 The ``@llvm.used`` global is an array which has
3518 :ref:`appending linkage <linkage_appending>`. This array contains a list of
3519 pointers to named global variables, functions and aliases which may optionally
3520 have a pointer cast formed of bitcast or getelementptr. For example, a legal
3523 .. code-block:: llvm
3528 @llvm.used = appending global [2 x i8*] [
3530 i8* bitcast (i32* @Y to i8*)
3531 ], section "llvm.metadata"
3533 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
3534 and linker are required to treat the symbol as if there is a reference to the
3535 symbol that it cannot see (which is why they have to be named). For example, if
3536 a variable has internal linkage and no references other than that from the
3537 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
3538 references from inline asms and other things the compiler cannot "see", and
3539 corresponds to "``attribute((used))``" in GNU C.
3541 On some targets, the code generator must emit a directive to the
3542 assembler or object file to prevent the assembler and linker from
3543 molesting the symbol.
3545 .. _gv_llvmcompilerused:
3547 The '``llvm.compiler.used``' Global Variable
3548 --------------------------------------------
3550 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
3551 directive, except that it only prevents the compiler from touching the
3552 symbol. On targets that support it, this allows an intelligent linker to
3553 optimize references to the symbol without being impeded as it would be
3556 This is a rare construct that should only be used in rare circumstances,
3557 and should not be exposed to source languages.
3559 .. _gv_llvmglobalctors:
3561 The '``llvm.global_ctors``' Global Variable
3562 -------------------------------------------
3564 .. code-block:: llvm
3566 %0 = type { i32, void ()*, i8* }
3567 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor, i8* @data }]
3569 The ``@llvm.global_ctors`` array contains a list of constructor
3570 functions, priorities, and an optional associated global or function.
3571 The functions referenced by this array will be called in ascending order
3572 of priority (i.e. lowest first) when the module is loaded. The order of
3573 functions with the same priority is not defined.
3575 If the third field is present, non-null, and points to a global variable
3576 or function, the initializer function will only run if the associated
3577 data from the current module is not discarded.
3579 .. _llvmglobaldtors:
3581 The '``llvm.global_dtors``' Global Variable
3582 -------------------------------------------
3584 .. code-block:: llvm
3586 %0 = type { i32, void ()*, i8* }
3587 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor, i8* @data }]
3589 The ``@llvm.global_dtors`` array contains a list of destructor
3590 functions, priorities, and an optional associated global or function.
3591 The functions referenced by this array will be called in descending
3592 order of priority (i.e. highest first) when the module is unloaded. The
3593 order of functions with the same priority is not defined.
3595 If the third field is present, non-null, and points to a global variable
3596 or function, the destructor function will only run if the associated
3597 data from the current module is not discarded.
3599 Instruction Reference
3600 =====================
3602 The LLVM instruction set consists of several different classifications
3603 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
3604 instructions <binaryops>`, :ref:`bitwise binary
3605 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
3606 :ref:`other instructions <otherops>`.
3610 Terminator Instructions
3611 -----------------------
3613 As mentioned :ref:`previously <functionstructure>`, every basic block in a
3614 program ends with a "Terminator" instruction, which indicates which
3615 block should be executed after the current block is finished. These
3616 terminator instructions typically yield a '``void``' value: they produce
3617 control flow, not values (the one exception being the
3618 ':ref:`invoke <i_invoke>`' instruction).
3620 The terminator instructions are: ':ref:`ret <i_ret>`',
3621 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
3622 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
3623 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
3627 '``ret``' Instruction
3628 ^^^^^^^^^^^^^^^^^^^^^
3635 ret <type> <value> ; Return a value from a non-void function
3636 ret void ; Return from void function
3641 The '``ret``' instruction is used to return control flow (and optionally
3642 a value) from a function back to the caller.
3644 There are two forms of the '``ret``' instruction: one that returns a
3645 value and then causes control flow, and one that just causes control
3651 The '``ret``' instruction optionally accepts a single argument, the
3652 return value. The type of the return value must be a ':ref:`first
3653 class <t_firstclass>`' type.
3655 A function is not :ref:`well formed <wellformed>` if it it has a non-void
3656 return type and contains a '``ret``' instruction with no return value or
3657 a return value with a type that does not match its type, or if it has a
3658 void return type and contains a '``ret``' instruction with a return
3664 When the '``ret``' instruction is executed, control flow returns back to
3665 the calling function's context. If the caller is a
3666 ":ref:`call <i_call>`" instruction, execution continues at the
3667 instruction after the call. If the caller was an
3668 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
3669 beginning of the "normal" destination block. If the instruction returns
3670 a value, that value shall set the call or invoke instruction's return
3676 .. code-block:: llvm
3678 ret i32 5 ; Return an integer value of 5
3679 ret void ; Return from a void function
3680 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
3684 '``br``' Instruction
3685 ^^^^^^^^^^^^^^^^^^^^
3692 br i1 <cond>, label <iftrue>, label <iffalse>
3693 br label <dest> ; Unconditional branch
3698 The '``br``' instruction is used to cause control flow to transfer to a
3699 different basic block in the current function. There are two forms of
3700 this instruction, corresponding to a conditional branch and an
3701 unconditional branch.
3706 The conditional branch form of the '``br``' instruction takes a single
3707 '``i1``' value and two '``label``' values. The unconditional form of the
3708 '``br``' instruction takes a single '``label``' value as a target.
3713 Upon execution of a conditional '``br``' instruction, the '``i1``'
3714 argument is evaluated. If the value is ``true``, control flows to the
3715 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
3716 to the '``iffalse``' ``label`` argument.
3721 .. code-block:: llvm
3724 %cond = icmp eq i32 %a, %b
3725 br i1 %cond, label %IfEqual, label %IfUnequal
3733 '``switch``' Instruction
3734 ^^^^^^^^^^^^^^^^^^^^^^^^
3741 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
3746 The '``switch``' instruction is used to transfer control flow to one of
3747 several different places. It is a generalization of the '``br``'
3748 instruction, allowing a branch to occur to one of many possible
3754 The '``switch``' instruction uses three parameters: an integer
3755 comparison value '``value``', a default '``label``' destination, and an
3756 array of pairs of comparison value constants and '``label``'s. The table
3757 is not allowed to contain duplicate constant entries.
3762 The ``switch`` instruction specifies a table of values and destinations.
3763 When the '``switch``' instruction is executed, this table is searched
3764 for the given value. If the value is found, control flow is transferred
3765 to the corresponding destination; otherwise, control flow is transferred
3766 to the default destination.
3771 Depending on properties of the target machine and the particular
3772 ``switch`` instruction, this instruction may be code generated in
3773 different ways. For example, it could be generated as a series of
3774 chained conditional branches or with a lookup table.
3779 .. code-block:: llvm
3781 ; Emulate a conditional br instruction
3782 %Val = zext i1 %value to i32
3783 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
3785 ; Emulate an unconditional br instruction
3786 switch i32 0, label %dest [ ]
3788 ; Implement a jump table:
3789 switch i32 %val, label %otherwise [ i32 0, label %onzero
3791 i32 2, label %ontwo ]
3795 '``indirectbr``' Instruction
3796 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3803 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
3808 The '``indirectbr``' instruction implements an indirect branch to a
3809 label within the current function, whose address is specified by
3810 "``address``". Address must be derived from a
3811 :ref:`blockaddress <blockaddress>` constant.
3816 The '``address``' argument is the address of the label to jump to. The
3817 rest of the arguments indicate the full set of possible destinations
3818 that the address may point to. Blocks are allowed to occur multiple
3819 times in the destination list, though this isn't particularly useful.
3821 This destination list is required so that dataflow analysis has an
3822 accurate understanding of the CFG.
3827 Control transfers to the block specified in the address argument. All
3828 possible destination blocks must be listed in the label list, otherwise
3829 this instruction has undefined behavior. This implies that jumps to
3830 labels defined in other functions have undefined behavior as well.
3835 This is typically implemented with a jump through a register.
3840 .. code-block:: llvm
3842 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
3846 '``invoke``' Instruction
3847 ^^^^^^^^^^^^^^^^^^^^^^^^
3854 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
3855 to label <normal label> unwind label <exception label>
3860 The '``invoke``' instruction causes control to transfer to a specified
3861 function, with the possibility of control flow transfer to either the
3862 '``normal``' label or the '``exception``' label. If the callee function
3863 returns with the "``ret``" instruction, control flow will return to the
3864 "normal" label. If the callee (or any indirect callees) returns via the
3865 ":ref:`resume <i_resume>`" instruction or other exception handling
3866 mechanism, control is interrupted and continued at the dynamically
3867 nearest "exception" label.
3869 The '``exception``' label is a `landing
3870 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
3871 '``exception``' label is required to have the
3872 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
3873 information about the behavior of the program after unwinding happens,
3874 as its first non-PHI instruction. The restrictions on the
3875 "``landingpad``" instruction's tightly couples it to the "``invoke``"
3876 instruction, so that the important information contained within the
3877 "``landingpad``" instruction can't be lost through normal code motion.
3882 This instruction requires several arguments:
3884 #. The optional "cconv" marker indicates which :ref:`calling
3885 convention <callingconv>` the call should use. If none is
3886 specified, the call defaults to using C calling conventions.
3887 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
3888 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
3890 #. '``ptr to function ty``': shall be the signature of the pointer to
3891 function value being invoked. In most cases, this is a direct
3892 function invocation, but indirect ``invoke``'s are just as possible,
3893 branching off an arbitrary pointer to function value.
3894 #. '``function ptr val``': An LLVM value containing a pointer to a
3895 function to be invoked.
3896 #. '``function args``': argument list whose types match the function
3897 signature argument types and parameter attributes. All arguments must
3898 be of :ref:`first class <t_firstclass>` type. If the function signature
3899 indicates the function accepts a variable number of arguments, the
3900 extra arguments can be specified.
3901 #. '``normal label``': the label reached when the called function
3902 executes a '``ret``' instruction.
3903 #. '``exception label``': the label reached when a callee returns via
3904 the :ref:`resume <i_resume>` instruction or other exception handling
3906 #. The optional :ref:`function attributes <fnattrs>` list. Only
3907 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
3908 attributes are valid here.
3913 This instruction is designed to operate as a standard '``call``'
3914 instruction in most regards. The primary difference is that it
3915 establishes an association with a label, which is used by the runtime
3916 library to unwind the stack.
3918 This instruction is used in languages with destructors to ensure that
3919 proper cleanup is performed in the case of either a ``longjmp`` or a
3920 thrown exception. Additionally, this is important for implementation of
3921 '``catch``' clauses in high-level languages that support them.
3923 For the purposes of the SSA form, the definition of the value returned
3924 by the '``invoke``' instruction is deemed to occur on the edge from the
3925 current block to the "normal" label. If the callee unwinds then no
3926 return value is available.
3931 .. code-block:: llvm
3933 %retval = invoke i32 @Test(i32 15) to label %Continue
3934 unwind label %TestCleanup ; i32:retval set
3935 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
3936 unwind label %TestCleanup ; i32:retval set
3940 '``resume``' Instruction
3941 ^^^^^^^^^^^^^^^^^^^^^^^^
3948 resume <type> <value>
3953 The '``resume``' instruction is a terminator instruction that has no
3959 The '``resume``' instruction requires one argument, which must have the
3960 same type as the result of any '``landingpad``' instruction in the same
3966 The '``resume``' instruction resumes propagation of an existing
3967 (in-flight) exception whose unwinding was interrupted with a
3968 :ref:`landingpad <i_landingpad>` instruction.
3973 .. code-block:: llvm
3975 resume { i8*, i32 } %exn
3979 '``unreachable``' Instruction
3980 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3992 The '``unreachable``' instruction has no defined semantics. This
3993 instruction is used to inform the optimizer that a particular portion of
3994 the code is not reachable. This can be used to indicate that the code
3995 after a no-return function cannot be reached, and other facts.
4000 The '``unreachable``' instruction has no defined semantics.
4007 Binary operators are used to do most of the computation in a program.
4008 They require two operands of the same type, execute an operation on
4009 them, and produce a single value. The operands might represent multiple
4010 data, as is the case with the :ref:`vector <t_vector>` data type. The
4011 result value has the same type as its operands.
4013 There are several different binary operators:
4017 '``add``' Instruction
4018 ^^^^^^^^^^^^^^^^^^^^^
4025 <result> = add <ty> <op1>, <op2> ; yields ty:result
4026 <result> = add nuw <ty> <op1>, <op2> ; yields ty:result
4027 <result> = add nsw <ty> <op1>, <op2> ; yields ty:result
4028 <result> = add nuw nsw <ty> <op1>, <op2> ; yields ty:result
4033 The '``add``' instruction returns the sum of its two operands.
4038 The two arguments to the '``add``' instruction must be
4039 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4040 arguments must have identical types.
4045 The value produced is the integer sum of the two operands.
4047 If the sum has unsigned overflow, the result returned is the
4048 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4051 Because LLVM integers use a two's complement representation, this
4052 instruction is appropriate for both signed and unsigned integers.
4054 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4055 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4056 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
4057 unsigned and/or signed overflow, respectively, occurs.
4062 .. code-block:: llvm
4064 <result> = add i32 4, %var ; yields i32:result = 4 + %var
4068 '``fadd``' Instruction
4069 ^^^^^^^^^^^^^^^^^^^^^^
4076 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4081 The '``fadd``' instruction returns the sum of its two operands.
4086 The two arguments to the '``fadd``' instruction must be :ref:`floating
4087 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4088 Both arguments must have identical types.
4093 The value produced is the floating point sum of the two operands. This
4094 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
4095 which are optimization hints to enable otherwise unsafe floating point
4101 .. code-block:: llvm
4103 <result> = fadd float 4.0, %var ; yields float:result = 4.0 + %var
4105 '``sub``' Instruction
4106 ^^^^^^^^^^^^^^^^^^^^^
4113 <result> = sub <ty> <op1>, <op2> ; yields ty:result
4114 <result> = sub nuw <ty> <op1>, <op2> ; yields ty:result
4115 <result> = sub nsw <ty> <op1>, <op2> ; yields ty:result
4116 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields ty:result
4121 The '``sub``' instruction returns the difference of its two operands.
4123 Note that the '``sub``' instruction is used to represent the '``neg``'
4124 instruction present in most other intermediate representations.
4129 The two arguments to the '``sub``' instruction must be
4130 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4131 arguments must have identical types.
4136 The value produced is the integer difference of the two operands.
4138 If the difference has unsigned overflow, the result returned is the
4139 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4142 Because LLVM integers use a two's complement representation, this
4143 instruction is appropriate for both signed and unsigned integers.
4145 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4146 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4147 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
4148 unsigned and/or signed overflow, respectively, occurs.
4153 .. code-block:: llvm
4155 <result> = sub i32 4, %var ; yields i32:result = 4 - %var
4156 <result> = sub i32 0, %val ; yields i32:result = -%var
4160 '``fsub``' Instruction
4161 ^^^^^^^^^^^^^^^^^^^^^^
4168 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4173 The '``fsub``' instruction returns the difference of its two operands.
4175 Note that the '``fsub``' instruction is used to represent the '``fneg``'
4176 instruction present in most other intermediate representations.
4181 The two arguments to the '``fsub``' instruction must be :ref:`floating
4182 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4183 Both arguments must have identical types.
4188 The value produced is the floating point difference of the two operands.
4189 This instruction can also take any number of :ref:`fast-math
4190 flags <fastmath>`, which are optimization hints to enable otherwise
4191 unsafe floating point optimizations:
4196 .. code-block:: llvm
4198 <result> = fsub float 4.0, %var ; yields float:result = 4.0 - %var
4199 <result> = fsub float -0.0, %val ; yields float:result = -%var
4201 '``mul``' Instruction
4202 ^^^^^^^^^^^^^^^^^^^^^
4209 <result> = mul <ty> <op1>, <op2> ; yields ty:result
4210 <result> = mul nuw <ty> <op1>, <op2> ; yields ty:result
4211 <result> = mul nsw <ty> <op1>, <op2> ; yields ty:result
4212 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields ty:result
4217 The '``mul``' instruction returns the product of its two operands.
4222 The two arguments to the '``mul``' instruction must be
4223 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4224 arguments must have identical types.
4229 The value produced is the integer product of the two operands.
4231 If the result of the multiplication has unsigned overflow, the result
4232 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
4233 bit width of the result.
4235 Because LLVM integers use a two's complement representation, and the
4236 result is the same width as the operands, this instruction returns the
4237 correct result for both signed and unsigned integers. If a full product
4238 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
4239 sign-extended or zero-extended as appropriate to the width of the full
4242 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4243 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4244 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
4245 unsigned and/or signed overflow, respectively, occurs.
4250 .. code-block:: llvm
4252 <result> = mul i32 4, %var ; yields i32:result = 4 * %var
4256 '``fmul``' Instruction
4257 ^^^^^^^^^^^^^^^^^^^^^^
4264 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4269 The '``fmul``' instruction returns the product of its two operands.
4274 The two arguments to the '``fmul``' instruction must be :ref:`floating
4275 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4276 Both arguments must have identical types.
4281 The value produced is the floating point product of the two operands.
4282 This instruction can also take any number of :ref:`fast-math
4283 flags <fastmath>`, which are optimization hints to enable otherwise
4284 unsafe floating point optimizations:
4289 .. code-block:: llvm
4291 <result> = fmul float 4.0, %var ; yields float:result = 4.0 * %var
4293 '``udiv``' Instruction
4294 ^^^^^^^^^^^^^^^^^^^^^^
4301 <result> = udiv <ty> <op1>, <op2> ; yields ty:result
4302 <result> = udiv exact <ty> <op1>, <op2> ; yields ty:result
4307 The '``udiv``' instruction returns the quotient of its two operands.
4312 The two arguments to the '``udiv``' instruction must be
4313 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4314 arguments must have identical types.
4319 The value produced is the unsigned integer quotient of the two operands.
4321 Note that unsigned integer division and signed integer division are
4322 distinct operations; for signed integer division, use '``sdiv``'.
4324 Division by zero leads to undefined behavior.
4326 If the ``exact`` keyword is present, the result value of the ``udiv`` is
4327 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
4328 such, "((a udiv exact b) mul b) == a").
4333 .. code-block:: llvm
4335 <result> = udiv i32 4, %var ; yields i32:result = 4 / %var
4337 '``sdiv``' Instruction
4338 ^^^^^^^^^^^^^^^^^^^^^^
4345 <result> = sdiv <ty> <op1>, <op2> ; yields ty:result
4346 <result> = sdiv exact <ty> <op1>, <op2> ; yields ty:result
4351 The '``sdiv``' instruction returns the quotient of its two operands.
4356 The two arguments to the '``sdiv``' instruction must be
4357 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4358 arguments must have identical types.
4363 The value produced is the signed integer quotient of the two operands
4364 rounded towards zero.
4366 Note that signed integer division and unsigned integer division are
4367 distinct operations; for unsigned integer division, use '``udiv``'.
4369 Division by zero leads to undefined behavior. Overflow also leads to
4370 undefined behavior; this is a rare case, but can occur, for example, by
4371 doing a 32-bit division of -2147483648 by -1.
4373 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
4374 a :ref:`poison value <poisonvalues>` if the result would be rounded.
4379 .. code-block:: llvm
4381 <result> = sdiv i32 4, %var ; yields i32:result = 4 / %var
4385 '``fdiv``' Instruction
4386 ^^^^^^^^^^^^^^^^^^^^^^
4393 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4398 The '``fdiv``' instruction returns the quotient of its two operands.
4403 The two arguments to the '``fdiv``' instruction must be :ref:`floating
4404 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4405 Both arguments must have identical types.
4410 The value produced is the floating point quotient of the two operands.
4411 This instruction can also take any number of :ref:`fast-math
4412 flags <fastmath>`, which are optimization hints to enable otherwise
4413 unsafe floating point optimizations:
4418 .. code-block:: llvm
4420 <result> = fdiv float 4.0, %var ; yields float:result = 4.0 / %var
4422 '``urem``' Instruction
4423 ^^^^^^^^^^^^^^^^^^^^^^
4430 <result> = urem <ty> <op1>, <op2> ; yields ty:result
4435 The '``urem``' instruction returns the remainder from the unsigned
4436 division of its two arguments.
4441 The two arguments to the '``urem``' instruction must be
4442 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4443 arguments must have identical types.
4448 This instruction returns the unsigned integer *remainder* of a division.
4449 This instruction always performs an unsigned division to get the
4452 Note that unsigned integer remainder and signed integer remainder are
4453 distinct operations; for signed integer remainder, use '``srem``'.
4455 Taking the remainder of a division by zero leads to undefined behavior.
4460 .. code-block:: llvm
4462 <result> = urem i32 4, %var ; yields i32:result = 4 % %var
4464 '``srem``' Instruction
4465 ^^^^^^^^^^^^^^^^^^^^^^
4472 <result> = srem <ty> <op1>, <op2> ; yields ty:result
4477 The '``srem``' instruction returns the remainder from the signed
4478 division of its two operands. This instruction can also take
4479 :ref:`vector <t_vector>` versions of the values in which case the elements
4485 The two arguments to the '``srem``' instruction must be
4486 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4487 arguments must have identical types.
4492 This instruction returns the *remainder* of a division (where the result
4493 is either zero or has the same sign as the dividend, ``op1``), not the
4494 *modulo* operator (where the result is either zero or has the same sign
4495 as the divisor, ``op2``) of a value. For more information about the
4496 difference, see `The Math
4497 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
4498 table of how this is implemented in various languages, please see
4500 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
4502 Note that signed integer remainder and unsigned integer remainder are
4503 distinct operations; for unsigned integer remainder, use '``urem``'.
4505 Taking the remainder of a division by zero leads to undefined behavior.
4506 Overflow also leads to undefined behavior; this is a rare case, but can
4507 occur, for example, by taking the remainder of a 32-bit division of
4508 -2147483648 by -1. (The remainder doesn't actually overflow, but this
4509 rule lets srem be implemented using instructions that return both the
4510 result of the division and the remainder.)
4515 .. code-block:: llvm
4517 <result> = srem i32 4, %var ; yields i32:result = 4 % %var
4521 '``frem``' Instruction
4522 ^^^^^^^^^^^^^^^^^^^^^^
4529 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4534 The '``frem``' instruction returns the remainder from the division of
4540 The two arguments to the '``frem``' instruction must be :ref:`floating
4541 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4542 Both arguments must have identical types.
4547 This instruction returns the *remainder* of a division. The remainder
4548 has the same sign as the dividend. This instruction can also take any
4549 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
4550 to enable otherwise unsafe floating point optimizations:
4555 .. code-block:: llvm
4557 <result> = frem float 4.0, %var ; yields float:result = 4.0 % %var
4561 Bitwise Binary Operations
4562 -------------------------
4564 Bitwise binary operators are used to do various forms of bit-twiddling
4565 in a program. They are generally very efficient instructions and can
4566 commonly be strength reduced from other instructions. They require two
4567 operands of the same type, execute an operation on them, and produce a
4568 single value. The resulting value is the same type as its operands.
4570 '``shl``' Instruction
4571 ^^^^^^^^^^^^^^^^^^^^^
4578 <result> = shl <ty> <op1>, <op2> ; yields ty:result
4579 <result> = shl nuw <ty> <op1>, <op2> ; yields ty:result
4580 <result> = shl nsw <ty> <op1>, <op2> ; yields ty:result
4581 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields ty:result
4586 The '``shl``' instruction returns the first operand shifted to the left
4587 a specified number of bits.
4592 Both arguments to the '``shl``' instruction must be the same
4593 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4594 '``op2``' is treated as an unsigned value.
4599 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
4600 where ``n`` is the width of the result. If ``op2`` is (statically or
4601 dynamically) negative or equal to or larger than the number of bits in
4602 ``op1``, the result is undefined. If the arguments are vectors, each
4603 vector element of ``op1`` is shifted by the corresponding shift amount
4606 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
4607 value <poisonvalues>` if it shifts out any non-zero bits. If the
4608 ``nsw`` keyword is present, then the shift produces a :ref:`poison
4609 value <poisonvalues>` if it shifts out any bits that disagree with the
4610 resultant sign bit. As such, NUW/NSW have the same semantics as they
4611 would if the shift were expressed as a mul instruction with the same
4612 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
4617 .. code-block:: llvm
4619 <result> = shl i32 4, %var ; yields i32: 4 << %var
4620 <result> = shl i32 4, 2 ; yields i32: 16
4621 <result> = shl i32 1, 10 ; yields i32: 1024
4622 <result> = shl i32 1, 32 ; undefined
4623 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
4625 '``lshr``' Instruction
4626 ^^^^^^^^^^^^^^^^^^^^^^
4633 <result> = lshr <ty> <op1>, <op2> ; yields ty:result
4634 <result> = lshr exact <ty> <op1>, <op2> ; yields ty:result
4639 The '``lshr``' instruction (logical shift right) returns the first
4640 operand shifted to the right a specified number of bits with zero fill.
4645 Both arguments to the '``lshr``' instruction must be the same
4646 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4647 '``op2``' is treated as an unsigned value.
4652 This instruction always performs a logical shift right operation. The
4653 most significant bits of the result will be filled with zero bits after
4654 the shift. If ``op2`` is (statically or dynamically) equal to or larger
4655 than the number of bits in ``op1``, the result is undefined. If the
4656 arguments are vectors, each vector element of ``op1`` is shifted by the
4657 corresponding shift amount in ``op2``.
4659 If the ``exact`` keyword is present, the result value of the ``lshr`` is
4660 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4666 .. code-block:: llvm
4668 <result> = lshr i32 4, 1 ; yields i32:result = 2
4669 <result> = lshr i32 4, 2 ; yields i32:result = 1
4670 <result> = lshr i8 4, 3 ; yields i8:result = 0
4671 <result> = lshr i8 -2, 1 ; yields i8:result = 0x7F
4672 <result> = lshr i32 1, 32 ; undefined
4673 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
4675 '``ashr``' Instruction
4676 ^^^^^^^^^^^^^^^^^^^^^^
4683 <result> = ashr <ty> <op1>, <op2> ; yields ty:result
4684 <result> = ashr exact <ty> <op1>, <op2> ; yields ty:result
4689 The '``ashr``' instruction (arithmetic shift right) returns the first
4690 operand shifted to the right a specified number of bits with sign
4696 Both arguments to the '``ashr``' instruction must be the same
4697 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4698 '``op2``' is treated as an unsigned value.
4703 This instruction always performs an arithmetic shift right operation,
4704 The most significant bits of the result will be filled with the sign bit
4705 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
4706 than the number of bits in ``op1``, the result is undefined. If the
4707 arguments are vectors, each vector element of ``op1`` is shifted by the
4708 corresponding shift amount in ``op2``.
4710 If the ``exact`` keyword is present, the result value of the ``ashr`` is
4711 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4717 .. code-block:: llvm
4719 <result> = ashr i32 4, 1 ; yields i32:result = 2
4720 <result> = ashr i32 4, 2 ; yields i32:result = 1
4721 <result> = ashr i8 4, 3 ; yields i8:result = 0
4722 <result> = ashr i8 -2, 1 ; yields i8:result = -1
4723 <result> = ashr i32 1, 32 ; undefined
4724 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
4726 '``and``' Instruction
4727 ^^^^^^^^^^^^^^^^^^^^^
4734 <result> = and <ty> <op1>, <op2> ; yields ty:result
4739 The '``and``' instruction returns the bitwise logical and of its two
4745 The two arguments to the '``and``' instruction must be
4746 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4747 arguments must have identical types.
4752 The truth table used for the '``and``' instruction is:
4769 .. code-block:: llvm
4771 <result> = and i32 4, %var ; yields i32:result = 4 & %var
4772 <result> = and i32 15, 40 ; yields i32:result = 8
4773 <result> = and i32 4, 8 ; yields i32:result = 0
4775 '``or``' Instruction
4776 ^^^^^^^^^^^^^^^^^^^^
4783 <result> = or <ty> <op1>, <op2> ; yields ty:result
4788 The '``or``' instruction returns the bitwise logical inclusive or of its
4794 The two arguments to the '``or``' instruction must be
4795 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4796 arguments must have identical types.
4801 The truth table used for the '``or``' instruction is:
4820 <result> = or i32 4, %var ; yields i32:result = 4 | %var
4821 <result> = or i32 15, 40 ; yields i32:result = 47
4822 <result> = or i32 4, 8 ; yields i32:result = 12
4824 '``xor``' Instruction
4825 ^^^^^^^^^^^^^^^^^^^^^
4832 <result> = xor <ty> <op1>, <op2> ; yields ty:result
4837 The '``xor``' instruction returns the bitwise logical exclusive or of
4838 its two operands. The ``xor`` is used to implement the "one's
4839 complement" operation, which is the "~" operator in C.
4844 The two arguments to the '``xor``' instruction must be
4845 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4846 arguments must have identical types.
4851 The truth table used for the '``xor``' instruction is:
4868 .. code-block:: llvm
4870 <result> = xor i32 4, %var ; yields i32:result = 4 ^ %var
4871 <result> = xor i32 15, 40 ; yields i32:result = 39
4872 <result> = xor i32 4, 8 ; yields i32:result = 12
4873 <result> = xor i32 %V, -1 ; yields i32:result = ~%V
4878 LLVM supports several instructions to represent vector operations in a
4879 target-independent manner. These instructions cover the element-access
4880 and vector-specific operations needed to process vectors effectively.
4881 While LLVM does directly support these vector operations, many
4882 sophisticated algorithms will want to use target-specific intrinsics to
4883 take full advantage of a specific target.
4885 .. _i_extractelement:
4887 '``extractelement``' Instruction
4888 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4895 <result> = extractelement <n x <ty>> <val>, <ty2> <idx> ; yields <ty>
4900 The '``extractelement``' instruction extracts a single scalar element
4901 from a vector at a specified index.
4906 The first operand of an '``extractelement``' instruction is a value of
4907 :ref:`vector <t_vector>` type. The second operand is an index indicating
4908 the position from which to extract the element. The index may be a
4909 variable of any integer type.
4914 The result is a scalar of the same type as the element type of ``val``.
4915 Its value is the value at position ``idx`` of ``val``. If ``idx``
4916 exceeds the length of ``val``, the results are undefined.
4921 .. code-block:: llvm
4923 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
4925 .. _i_insertelement:
4927 '``insertelement``' Instruction
4928 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4935 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, <ty2> <idx> ; yields <n x <ty>>
4940 The '``insertelement``' instruction inserts a scalar element into a
4941 vector at a specified index.
4946 The first operand of an '``insertelement``' instruction is a value of
4947 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
4948 type must equal the element type of the first operand. The third operand
4949 is an index indicating the position at which to insert the value. The
4950 index may be a variable of any integer type.
4955 The result is a vector of the same type as ``val``. Its element values
4956 are those of ``val`` except at position ``idx``, where it gets the value
4957 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
4963 .. code-block:: llvm
4965 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
4967 .. _i_shufflevector:
4969 '``shufflevector``' Instruction
4970 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4977 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
4982 The '``shufflevector``' instruction constructs a permutation of elements
4983 from two input vectors, returning a vector with the same element type as
4984 the input and length that is the same as the shuffle mask.
4989 The first two operands of a '``shufflevector``' instruction are vectors
4990 with the same type. The third argument is a shuffle mask whose element
4991 type is always 'i32'. The result of the instruction is a vector whose
4992 length is the same as the shuffle mask and whose element type is the
4993 same as the element type of the first two operands.
4995 The shuffle mask operand is required to be a constant vector with either
4996 constant integer or undef values.
5001 The elements of the two input vectors are numbered from left to right
5002 across both of the vectors. The shuffle mask operand specifies, for each
5003 element of the result vector, which element of the two input vectors the
5004 result element gets. The element selector may be undef (meaning "don't
5005 care") and the second operand may be undef if performing a shuffle from
5011 .. code-block:: llvm
5013 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
5014 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
5015 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
5016 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
5017 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
5018 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
5019 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
5020 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
5022 Aggregate Operations
5023 --------------------
5025 LLVM supports several instructions for working with
5026 :ref:`aggregate <t_aggregate>` values.
5030 '``extractvalue``' Instruction
5031 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5038 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
5043 The '``extractvalue``' instruction extracts the value of a member field
5044 from an :ref:`aggregate <t_aggregate>` value.
5049 The first operand of an '``extractvalue``' instruction is a value of
5050 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
5051 constant indices to specify which value to extract in a similar manner
5052 as indices in a '``getelementptr``' instruction.
5054 The major differences to ``getelementptr`` indexing are:
5056 - Since the value being indexed is not a pointer, the first index is
5057 omitted and assumed to be zero.
5058 - At least one index must be specified.
5059 - Not only struct indices but also array indices must be in bounds.
5064 The result is the value at the position in the aggregate specified by
5070 .. code-block:: llvm
5072 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
5076 '``insertvalue``' Instruction
5077 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5084 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
5089 The '``insertvalue``' instruction inserts a value into a member field in
5090 an :ref:`aggregate <t_aggregate>` value.
5095 The first operand of an '``insertvalue``' instruction is a value of
5096 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
5097 a first-class value to insert. The following operands are constant
5098 indices indicating the position at which to insert the value in a
5099 similar manner as indices in a '``extractvalue``' instruction. The value
5100 to insert must have the same type as the value identified by the
5106 The result is an aggregate of the same type as ``val``. Its value is
5107 that of ``val`` except that the value at the position specified by the
5108 indices is that of ``elt``.
5113 .. code-block:: llvm
5115 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
5116 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
5117 %agg3 = insertvalue {i32, {float}} undef, float %val, 1, 0 ; yields {i32 undef, {float %val}}
5121 Memory Access and Addressing Operations
5122 ---------------------------------------
5124 A key design point of an SSA-based representation is how it represents
5125 memory. In LLVM, no memory locations are in SSA form, which makes things
5126 very simple. This section describes how to read, write, and allocate
5131 '``alloca``' Instruction
5132 ^^^^^^^^^^^^^^^^^^^^^^^^
5139 <result> = alloca [inalloca] <type> [, <ty> <NumElements>] [, align <alignment>] ; yields type*:result
5144 The '``alloca``' instruction allocates memory on the stack frame of the
5145 currently executing function, to be automatically released when this
5146 function returns to its caller. The object is always allocated in the
5147 generic address space (address space zero).
5152 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
5153 bytes of memory on the runtime stack, returning a pointer of the
5154 appropriate type to the program. If "NumElements" is specified, it is
5155 the number of elements allocated, otherwise "NumElements" is defaulted
5156 to be one. If a constant alignment is specified, the value result of the
5157 allocation is guaranteed to be aligned to at least that boundary. The
5158 alignment may not be greater than ``1 << 29``. If not specified, or if
5159 zero, the target can choose to align the allocation on any convenient
5160 boundary compatible with the type.
5162 '``type``' may be any sized type.
5167 Memory is allocated; a pointer is returned. The operation is undefined
5168 if there is insufficient stack space for the allocation. '``alloca``'d
5169 memory is automatically released when the function returns. The
5170 '``alloca``' instruction is commonly used to represent automatic
5171 variables that must have an address available. When the function returns
5172 (either with the ``ret`` or ``resume`` instructions), the memory is
5173 reclaimed. Allocating zero bytes is legal, but the result is undefined.
5174 The order in which memory is allocated (ie., which way the stack grows)
5180 .. code-block:: llvm
5182 %ptr = alloca i32 ; yields i32*:ptr
5183 %ptr = alloca i32, i32 4 ; yields i32*:ptr
5184 %ptr = alloca i32, i32 4, align 1024 ; yields i32*:ptr
5185 %ptr = alloca i32, align 1024 ; yields i32*:ptr
5189 '``load``' Instruction
5190 ^^^^^^^^^^^^^^^^^^^^^^
5197 <result> = load [volatile] <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>][, !nonnull !<index>]
5198 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
5199 !<index> = !{ i32 1 }
5204 The '``load``' instruction is used to read from memory.
5209 The argument to the ``load`` instruction specifies the memory address
5210 from which to load. The pointer must point to a :ref:`first
5211 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
5212 then the optimizer is not allowed to modify the number or order of
5213 execution of this ``load`` with other :ref:`volatile
5214 operations <volatile>`.
5216 If the ``load`` is marked as ``atomic``, it takes an extra
5217 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5218 ``release`` and ``acq_rel`` orderings are not valid on ``load``
5219 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5220 when they may see multiple atomic stores. The type of the pointee must
5221 be an integer type whose bit width is a power of two greater than or
5222 equal to eight and less than or equal to a target-specific size limit.
5223 ``align`` must be explicitly specified on atomic loads, and the load has
5224 undefined behavior if the alignment is not set to a value which is at
5225 least the size in bytes of the pointee. ``!nontemporal`` does not have
5226 any defined semantics for atomic loads.
5228 The optional constant ``align`` argument specifies the alignment of the
5229 operation (that is, the alignment of the memory address). A value of 0
5230 or an omitted ``align`` argument means that the operation has the ABI
5231 alignment for the target. It is the responsibility of the code emitter
5232 to ensure that the alignment information is correct. Overestimating the
5233 alignment results in undefined behavior. Underestimating the alignment
5234 may produce less efficient code. An alignment of 1 is always safe. The
5235 maximum possible alignment is ``1 << 29``.
5237 The optional ``!nontemporal`` metadata must reference a single
5238 metadata name ``<index>`` corresponding to a metadata node with one
5239 ``i32`` entry of value 1. The existence of the ``!nontemporal``
5240 metadata on the instruction tells the optimizer and code generator
5241 that this load is not expected to be reused in the cache. The code
5242 generator may select special instructions to save cache bandwidth, such
5243 as the ``MOVNT`` instruction on x86.
5245 The optional ``!invariant.load`` metadata must reference a single
5246 metadata name ``<index>`` corresponding to a metadata node with no
5247 entries. The existence of the ``!invariant.load`` metadata on the
5248 instruction tells the optimizer and code generator that the address
5249 operand to this load points to memory which can be assumed unchanged.
5250 Being invariant does not imply that a location is dereferenceable,
5251 but it does imply that once the location is known dereferenceable
5252 its value is henceforth unchanging.
5254 The optional ``!nonnull`` metadata must reference a single
5255 metadata name ``<index>`` corresponding to a metadata node with no
5256 entries. The existence of the ``!nonnull`` metadata on the
5257 instruction tells the optimizer that the value loaded is known to
5258 never be null. This is analogous to the ''nonnull'' attribute
5259 on parameters and return values. This metadata can only be applied
5260 to loads of a pointer type.
5265 The location of memory pointed to is loaded. If the value being loaded
5266 is of scalar type then the number of bytes read does not exceed the
5267 minimum number of bytes needed to hold all bits of the type. For
5268 example, loading an ``i24`` reads at most three bytes. When loading a
5269 value of a type like ``i20`` with a size that is not an integral number
5270 of bytes, the result is undefined if the value was not originally
5271 written using a store of the same type.
5276 .. code-block:: llvm
5278 %ptr = alloca i32 ; yields i32*:ptr
5279 store i32 3, i32* %ptr ; yields void
5280 %val = load i32* %ptr ; yields i32:val = i32 3
5284 '``store``' Instruction
5285 ^^^^^^^^^^^^^^^^^^^^^^^
5292 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields void
5293 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields void
5298 The '``store``' instruction is used to write to memory.
5303 There are two arguments to the ``store`` instruction: a value to store
5304 and an address at which to store it. The type of the ``<pointer>``
5305 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
5306 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
5307 then the optimizer is not allowed to modify the number or order of
5308 execution of this ``store`` with other :ref:`volatile
5309 operations <volatile>`.
5311 If the ``store`` is marked as ``atomic``, it takes an extra
5312 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5313 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
5314 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5315 when they may see multiple atomic stores. The type of the pointee must
5316 be an integer type whose bit width is a power of two greater than or
5317 equal to eight and less than or equal to a target-specific size limit.
5318 ``align`` must be explicitly specified on atomic stores, and the store
5319 has undefined behavior if the alignment is not set to a value which is
5320 at least the size in bytes of the pointee. ``!nontemporal`` does not
5321 have any defined semantics for atomic stores.
5323 The optional constant ``align`` argument specifies the alignment of the
5324 operation (that is, the alignment of the memory address). A value of 0
5325 or an omitted ``align`` argument means that the operation has the ABI
5326 alignment for the target. It is the responsibility of the code emitter
5327 to ensure that the alignment information is correct. Overestimating the
5328 alignment results in undefined behavior. Underestimating the
5329 alignment may produce less efficient code. An alignment of 1 is always
5330 safe. The maximum possible alignment is ``1 << 29``.
5332 The optional ``!nontemporal`` metadata must reference a single metadata
5333 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
5334 value 1. The existence of the ``!nontemporal`` metadata on the instruction
5335 tells the optimizer and code generator that this load is not expected to
5336 be reused in the cache. The code generator may select special
5337 instructions to save cache bandwidth, such as the MOVNT instruction on
5343 The contents of memory are updated to contain ``<value>`` at the
5344 location specified by the ``<pointer>`` operand. If ``<value>`` is
5345 of scalar type then the number of bytes written does not exceed the
5346 minimum number of bytes needed to hold all bits of the type. For
5347 example, storing an ``i24`` writes at most three bytes. When writing a
5348 value of a type like ``i20`` with a size that is not an integral number
5349 of bytes, it is unspecified what happens to the extra bits that do not
5350 belong to the type, but they will typically be overwritten.
5355 .. code-block:: llvm
5357 %ptr = alloca i32 ; yields i32*:ptr
5358 store i32 3, i32* %ptr ; yields void
5359 %val = load i32* %ptr ; yields i32:val = i32 3
5363 '``fence``' Instruction
5364 ^^^^^^^^^^^^^^^^^^^^^^^
5371 fence [singlethread] <ordering> ; yields void
5376 The '``fence``' instruction is used to introduce happens-before edges
5382 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
5383 defines what *synchronizes-with* edges they add. They can only be given
5384 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
5389 A fence A which has (at least) ``release`` ordering semantics
5390 *synchronizes with* a fence B with (at least) ``acquire`` ordering
5391 semantics if and only if there exist atomic operations X and Y, both
5392 operating on some atomic object M, such that A is sequenced before X, X
5393 modifies M (either directly or through some side effect of a sequence
5394 headed by X), Y is sequenced before B, and Y observes M. This provides a
5395 *happens-before* dependency between A and B. Rather than an explicit
5396 ``fence``, one (but not both) of the atomic operations X or Y might
5397 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
5398 still *synchronize-with* the explicit ``fence`` and establish the
5399 *happens-before* edge.
5401 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
5402 ``acquire`` and ``release`` semantics specified above, participates in
5403 the global program order of other ``seq_cst`` operations and/or fences.
5405 The optional ":ref:`singlethread <singlethread>`" argument specifies
5406 that the fence only synchronizes with other fences in the same thread.
5407 (This is useful for interacting with signal handlers.)
5412 .. code-block:: llvm
5414 fence acquire ; yields void
5415 fence singlethread seq_cst ; yields void
5419 '``cmpxchg``' Instruction
5420 ^^^^^^^^^^^^^^^^^^^^^^^^^
5427 cmpxchg [weak] [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <success ordering> <failure ordering> ; yields { ty, i1 }
5432 The '``cmpxchg``' instruction is used to atomically modify memory. It
5433 loads a value in memory and compares it to a given value. If they are
5434 equal, it tries to store a new value into the memory.
5439 There are three arguments to the '``cmpxchg``' instruction: an address
5440 to operate on, a value to compare to the value currently be at that
5441 address, and a new value to place at that address if the compared values
5442 are equal. The type of '<cmp>' must be an integer type whose bit width
5443 is a power of two greater than or equal to eight and less than or equal
5444 to a target-specific size limit. '<cmp>' and '<new>' must have the same
5445 type, and the type of '<pointer>' must be a pointer to that type. If the
5446 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
5447 to modify the number or order of execution of this ``cmpxchg`` with
5448 other :ref:`volatile operations <volatile>`.
5450 The success and failure :ref:`ordering <ordering>` arguments specify how this
5451 ``cmpxchg`` synchronizes with other atomic operations. Both ordering parameters
5452 must be at least ``monotonic``, the ordering constraint on failure must be no
5453 stronger than that on success, and the failure ordering cannot be either
5454 ``release`` or ``acq_rel``.
5456 The optional "``singlethread``" argument declares that the ``cmpxchg``
5457 is only atomic with respect to code (usually signal handlers) running in
5458 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
5459 respect to all other code in the system.
5461 The pointer passed into cmpxchg must have alignment greater than or
5462 equal to the size in memory of the operand.
5467 The contents of memory at the location specified by the '``<pointer>``' operand
5468 is read and compared to '``<cmp>``'; if the read value is the equal, the
5469 '``<new>``' is written. The original value at the location is returned, together
5470 with a flag indicating success (true) or failure (false).
5472 If the cmpxchg operation is marked as ``weak`` then a spurious failure is
5473 permitted: the operation may not write ``<new>`` even if the comparison
5476 If the cmpxchg operation is strong (the default), the i1 value is 1 if and only
5477 if the value loaded equals ``cmp``.
5479 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose of
5480 identifying release sequences. A failed ``cmpxchg`` is equivalent to an atomic
5481 load with an ordering parameter determined the second ordering parameter.
5486 .. code-block:: llvm
5489 %orig = atomic load i32* %ptr unordered ; yields i32
5493 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
5494 %squared = mul i32 %cmp, %cmp
5495 %val_success = cmpxchg i32* %ptr, i32 %cmp, i32 %squared acq_rel monotonic ; yields { i32, i1 }
5496 %value_loaded = extractvalue { i32, i1 } %val_success, 0
5497 %success = extractvalue { i32, i1 } %val_success, 1
5498 br i1 %success, label %done, label %loop
5505 '``atomicrmw``' Instruction
5506 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
5513 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields ty
5518 The '``atomicrmw``' instruction is used to atomically modify memory.
5523 There are three arguments to the '``atomicrmw``' instruction: an
5524 operation to apply, an address whose value to modify, an argument to the
5525 operation. The operation must be one of the following keywords:
5539 The type of '<value>' must be an integer type whose bit width is a power
5540 of two greater than or equal to eight and less than or equal to a
5541 target-specific size limit. The type of the '``<pointer>``' operand must
5542 be a pointer to that type. If the ``atomicrmw`` is marked as
5543 ``volatile``, then the optimizer is not allowed to modify the number or
5544 order of execution of this ``atomicrmw`` with other :ref:`volatile
5545 operations <volatile>`.
5550 The contents of memory at the location specified by the '``<pointer>``'
5551 operand are atomically read, modified, and written back. The original
5552 value at the location is returned. The modification is specified by the
5555 - xchg: ``*ptr = val``
5556 - add: ``*ptr = *ptr + val``
5557 - sub: ``*ptr = *ptr - val``
5558 - and: ``*ptr = *ptr & val``
5559 - nand: ``*ptr = ~(*ptr & val)``
5560 - or: ``*ptr = *ptr | val``
5561 - xor: ``*ptr = *ptr ^ val``
5562 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
5563 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
5564 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
5566 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
5572 .. code-block:: llvm
5574 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields i32
5576 .. _i_getelementptr:
5578 '``getelementptr``' Instruction
5579 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5586 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
5587 <result> = getelementptr inbounds <pty>* <ptrval>{, <ty> <idx>}*
5588 <result> = getelementptr <ptr vector> ptrval, <vector index type> idx
5593 The '``getelementptr``' instruction is used to get the address of a
5594 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
5595 address calculation only and does not access memory.
5600 The first argument is always a pointer or a vector of pointers, and
5601 forms the basis of the calculation. The remaining arguments are indices
5602 that indicate which of the elements of the aggregate object are indexed.
5603 The interpretation of each index is dependent on the type being indexed
5604 into. The first index always indexes the pointer value given as the
5605 first argument, the second index indexes a value of the type pointed to
5606 (not necessarily the value directly pointed to, since the first index
5607 can be non-zero), etc. The first type indexed into must be a pointer
5608 value, subsequent types can be arrays, vectors, and structs. Note that
5609 subsequent types being indexed into can never be pointers, since that
5610 would require loading the pointer before continuing calculation.
5612 The type of each index argument depends on the type it is indexing into.
5613 When indexing into a (optionally packed) structure, only ``i32`` integer
5614 **constants** are allowed (when using a vector of indices they must all
5615 be the **same** ``i32`` integer constant). When indexing into an array,
5616 pointer or vector, integers of any width are allowed, and they are not
5617 required to be constant. These integers are treated as signed values
5620 For example, let's consider a C code fragment and how it gets compiled
5636 int *foo(struct ST *s) {
5637 return &s[1].Z.B[5][13];
5640 The LLVM code generated by Clang is:
5642 .. code-block:: llvm
5644 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
5645 %struct.ST = type { i32, double, %struct.RT }
5647 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
5649 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5656 In the example above, the first index is indexing into the
5657 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
5658 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
5659 indexes into the third element of the structure, yielding a
5660 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
5661 structure. The third index indexes into the second element of the
5662 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
5663 dimensions of the array are subscripted into, yielding an '``i32``'
5664 type. The '``getelementptr``' instruction returns a pointer to this
5665 element, thus computing a value of '``i32*``' type.
5667 Note that it is perfectly legal to index partially through a structure,
5668 returning a pointer to an inner element. Because of this, the LLVM code
5669 for the given testcase is equivalent to:
5671 .. code-block:: llvm
5673 define i32* @foo(%struct.ST* %s) {
5674 %t1 = getelementptr %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
5675 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
5676 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
5677 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
5678 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
5682 If the ``inbounds`` keyword is present, the result value of the
5683 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
5684 pointer is not an *in bounds* address of an allocated object, or if any
5685 of the addresses that would be formed by successive addition of the
5686 offsets implied by the indices to the base address with infinitely
5687 precise signed arithmetic are not an *in bounds* address of that
5688 allocated object. The *in bounds* addresses for an allocated object are
5689 all the addresses that point into the object, plus the address one byte
5690 past the end. In cases where the base is a vector of pointers the
5691 ``inbounds`` keyword applies to each of the computations element-wise.
5693 If the ``inbounds`` keyword is not present, the offsets are added to the
5694 base address with silently-wrapping two's complement arithmetic. If the
5695 offsets have a different width from the pointer, they are sign-extended
5696 or truncated to the width of the pointer. The result value of the
5697 ``getelementptr`` may be outside the object pointed to by the base
5698 pointer. The result value may not necessarily be used to access memory
5699 though, even if it happens to point into allocated storage. See the
5700 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
5703 The getelementptr instruction is often confusing. For some more insight
5704 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
5709 .. code-block:: llvm
5711 ; yields [12 x i8]*:aptr
5712 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5714 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
5716 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
5718 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
5720 In cases where the pointer argument is a vector of pointers, each index
5721 must be a vector with the same number of elements. For example:
5723 .. code-block:: llvm
5725 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5727 Conversion Operations
5728 ---------------------
5730 The instructions in this category are the conversion instructions
5731 (casting) which all take a single operand and a type. They perform
5732 various bit conversions on the operand.
5734 '``trunc .. to``' Instruction
5735 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5742 <result> = trunc <ty> <value> to <ty2> ; yields ty2
5747 The '``trunc``' instruction truncates its operand to the type ``ty2``.
5752 The '``trunc``' instruction takes a value to trunc, and a type to trunc
5753 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
5754 of the same number of integers. The bit size of the ``value`` must be
5755 larger than the bit size of the destination type, ``ty2``. Equal sized
5756 types are not allowed.
5761 The '``trunc``' instruction truncates the high order bits in ``value``
5762 and converts the remaining bits to ``ty2``. Since the source size must
5763 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
5764 It will always truncate bits.
5769 .. code-block:: llvm
5771 %X = trunc i32 257 to i8 ; yields i8:1
5772 %Y = trunc i32 123 to i1 ; yields i1:true
5773 %Z = trunc i32 122 to i1 ; yields i1:false
5774 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
5776 '``zext .. to``' Instruction
5777 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5784 <result> = zext <ty> <value> to <ty2> ; yields ty2
5789 The '``zext``' instruction zero extends its operand to type ``ty2``.
5794 The '``zext``' instruction takes a value to cast, and a type to cast it
5795 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5796 the same number of integers. The bit size of the ``value`` must be
5797 smaller than the bit size of the destination type, ``ty2``.
5802 The ``zext`` fills the high order bits of the ``value`` with zero bits
5803 until it reaches the size of the destination type, ``ty2``.
5805 When zero extending from i1, the result will always be either 0 or 1.
5810 .. code-block:: llvm
5812 %X = zext i32 257 to i64 ; yields i64:257
5813 %Y = zext i1 true to i32 ; yields i32:1
5814 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5816 '``sext .. to``' Instruction
5817 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5824 <result> = sext <ty> <value> to <ty2> ; yields ty2
5829 The '``sext``' sign extends ``value`` to the type ``ty2``.
5834 The '``sext``' instruction takes a value to cast, and a type to cast it
5835 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5836 the same number of integers. The bit size of the ``value`` must be
5837 smaller than the bit size of the destination type, ``ty2``.
5842 The '``sext``' instruction performs a sign extension by copying the sign
5843 bit (highest order bit) of the ``value`` until it reaches the bit size
5844 of the type ``ty2``.
5846 When sign extending from i1, the extension always results in -1 or 0.
5851 .. code-block:: llvm
5853 %X = sext i8 -1 to i16 ; yields i16 :65535
5854 %Y = sext i1 true to i32 ; yields i32:-1
5855 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5857 '``fptrunc .. to``' Instruction
5858 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5865 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
5870 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
5875 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
5876 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
5877 The size of ``value`` must be larger than the size of ``ty2``. This
5878 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
5883 The '``fptrunc``' instruction truncates a ``value`` from a larger
5884 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
5885 point <t_floating>` type. If the value cannot fit within the
5886 destination type, ``ty2``, then the results are undefined.
5891 .. code-block:: llvm
5893 %X = fptrunc double 123.0 to float ; yields float:123.0
5894 %Y = fptrunc double 1.0E+300 to float ; yields undefined
5896 '``fpext .. to``' Instruction
5897 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5904 <result> = fpext <ty> <value> to <ty2> ; yields ty2
5909 The '``fpext``' extends a floating point ``value`` to a larger floating
5915 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
5916 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
5917 to. The source type must be smaller than the destination type.
5922 The '``fpext``' instruction extends the ``value`` from a smaller
5923 :ref:`floating point <t_floating>` type to a larger :ref:`floating
5924 point <t_floating>` type. The ``fpext`` cannot be used to make a
5925 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
5926 *no-op cast* for a floating point cast.
5931 .. code-block:: llvm
5933 %X = fpext float 3.125 to double ; yields double:3.125000e+00
5934 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
5936 '``fptoui .. to``' Instruction
5937 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5944 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
5949 The '``fptoui``' converts a floating point ``value`` to its unsigned
5950 integer equivalent of type ``ty2``.
5955 The '``fptoui``' instruction takes a value to cast, which must be a
5956 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5957 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5958 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5959 type with the same number of elements as ``ty``
5964 The '``fptoui``' instruction converts its :ref:`floating
5965 point <t_floating>` operand into the nearest (rounding towards zero)
5966 unsigned integer value. If the value cannot fit in ``ty2``, the results
5972 .. code-block:: llvm
5974 %X = fptoui double 123.0 to i32 ; yields i32:123
5975 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
5976 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
5978 '``fptosi .. to``' Instruction
5979 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5986 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
5991 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
5992 ``value`` to type ``ty2``.
5997 The '``fptosi``' instruction takes a value to cast, which must be a
5998 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5999 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
6000 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
6001 type with the same number of elements as ``ty``
6006 The '``fptosi``' instruction converts its :ref:`floating
6007 point <t_floating>` operand into the nearest (rounding towards zero)
6008 signed integer value. If the value cannot fit in ``ty2``, the results
6014 .. code-block:: llvm
6016 %X = fptosi double -123.0 to i32 ; yields i32:-123
6017 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
6018 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
6020 '``uitofp .. to``' Instruction
6021 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6028 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
6033 The '``uitofp``' instruction regards ``value`` as an unsigned integer
6034 and converts that value to the ``ty2`` type.
6039 The '``uitofp``' instruction takes a value to cast, which must be a
6040 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
6041 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
6042 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
6043 type with the same number of elements as ``ty``
6048 The '``uitofp``' instruction interprets its operand as an unsigned
6049 integer quantity and converts it to the corresponding floating point
6050 value. If the value cannot fit in the floating point value, the results
6056 .. code-block:: llvm
6058 %X = uitofp i32 257 to float ; yields float:257.0
6059 %Y = uitofp i8 -1 to double ; yields double:255.0
6061 '``sitofp .. to``' Instruction
6062 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6069 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
6074 The '``sitofp``' instruction regards ``value`` as a signed integer and
6075 converts that value to the ``ty2`` type.
6080 The '``sitofp``' instruction takes a value to cast, which must be a
6081 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
6082 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
6083 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
6084 type with the same number of elements as ``ty``
6089 The '``sitofp``' instruction interprets its operand as a signed integer
6090 quantity and converts it to the corresponding floating point value. If
6091 the value cannot fit in the floating point value, the results are
6097 .. code-block:: llvm
6099 %X = sitofp i32 257 to float ; yields float:257.0
6100 %Y = sitofp i8 -1 to double ; yields double:-1.0
6104 '``ptrtoint .. to``' Instruction
6105 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6112 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
6117 The '``ptrtoint``' instruction converts the pointer or a vector of
6118 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
6123 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
6124 a a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
6125 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
6126 a vector of integers type.
6131 The '``ptrtoint``' instruction converts ``value`` to integer type
6132 ``ty2`` by interpreting the pointer value as an integer and either
6133 truncating or zero extending that value to the size of the integer type.
6134 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
6135 ``value`` is larger than ``ty2`` then a truncation is done. If they are
6136 the same size, then nothing is done (*no-op cast*) other than a type
6142 .. code-block:: llvm
6144 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
6145 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
6146 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
6150 '``inttoptr .. to``' Instruction
6151 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6158 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
6163 The '``inttoptr``' instruction converts an integer ``value`` to a
6164 pointer type, ``ty2``.
6169 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
6170 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
6176 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
6177 applying either a zero extension or a truncation depending on the size
6178 of the integer ``value``. If ``value`` is larger than the size of a
6179 pointer then a truncation is done. If ``value`` is smaller than the size
6180 of a pointer then a zero extension is done. If they are the same size,
6181 nothing is done (*no-op cast*).
6186 .. code-block:: llvm
6188 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
6189 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
6190 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
6191 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
6195 '``bitcast .. to``' Instruction
6196 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6203 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
6208 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
6214 The '``bitcast``' instruction takes a value to cast, which must be a
6215 non-aggregate first class value, and a type to cast it to, which must
6216 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
6217 bit sizes of ``value`` and the destination type, ``ty2``, must be
6218 identical. If the source type is a pointer, the destination type must
6219 also be a pointer of the same size. This instruction supports bitwise
6220 conversion of vectors to integers and to vectors of other types (as
6221 long as they have the same size).
6226 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
6227 is always a *no-op cast* because no bits change with this
6228 conversion. The conversion is done as if the ``value`` had been stored
6229 to memory and read back as type ``ty2``. Pointer (or vector of
6230 pointers) types may only be converted to other pointer (or vector of
6231 pointers) types with the same address space through this instruction.
6232 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
6233 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
6238 .. code-block:: llvm
6240 %X = bitcast i8 255 to i8 ; yields i8 :-1
6241 %Y = bitcast i32* %x to sint* ; yields sint*:%x
6242 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
6243 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
6245 .. _i_addrspacecast:
6247 '``addrspacecast .. to``' Instruction
6248 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6255 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
6260 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
6261 address space ``n`` to type ``pty2`` in address space ``m``.
6266 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
6267 to cast and a pointer type to cast it to, which must have a different
6273 The '``addrspacecast``' instruction converts the pointer value
6274 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
6275 value modification, depending on the target and the address space
6276 pair. Pointer conversions within the same address space must be
6277 performed with the ``bitcast`` instruction. Note that if the address space
6278 conversion is legal then both result and operand refer to the same memory
6284 .. code-block:: llvm
6286 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
6287 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
6288 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
6295 The instructions in this category are the "miscellaneous" instructions,
6296 which defy better classification.
6300 '``icmp``' Instruction
6301 ^^^^^^^^^^^^^^^^^^^^^^
6308 <result> = icmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6313 The '``icmp``' instruction returns a boolean value or a vector of
6314 boolean values based on comparison of its two integer, integer vector,
6315 pointer, or pointer vector operands.
6320 The '``icmp``' instruction takes three operands. The first operand is
6321 the condition code indicating the kind of comparison to perform. It is
6322 not a value, just a keyword. The possible condition code are:
6325 #. ``ne``: not equal
6326 #. ``ugt``: unsigned greater than
6327 #. ``uge``: unsigned greater or equal
6328 #. ``ult``: unsigned less than
6329 #. ``ule``: unsigned less or equal
6330 #. ``sgt``: signed greater than
6331 #. ``sge``: signed greater or equal
6332 #. ``slt``: signed less than
6333 #. ``sle``: signed less or equal
6335 The remaining two arguments must be :ref:`integer <t_integer>` or
6336 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
6337 must also be identical types.
6342 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
6343 code given as ``cond``. The comparison performed always yields either an
6344 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
6346 #. ``eq``: yields ``true`` if the operands are equal, ``false``
6347 otherwise. No sign interpretation is necessary or performed.
6348 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
6349 otherwise. No sign interpretation is necessary or performed.
6350 #. ``ugt``: interprets the operands as unsigned values and yields
6351 ``true`` if ``op1`` is greater than ``op2``.
6352 #. ``uge``: interprets the operands as unsigned values and yields
6353 ``true`` if ``op1`` is greater than or equal to ``op2``.
6354 #. ``ult``: interprets the operands as unsigned values and yields
6355 ``true`` if ``op1`` is less than ``op2``.
6356 #. ``ule``: interprets the operands as unsigned values and yields
6357 ``true`` if ``op1`` is less than or equal to ``op2``.
6358 #. ``sgt``: interprets the operands as signed values and yields ``true``
6359 if ``op1`` is greater than ``op2``.
6360 #. ``sge``: interprets the operands as signed values and yields ``true``
6361 if ``op1`` is greater than or equal to ``op2``.
6362 #. ``slt``: interprets the operands as signed values and yields ``true``
6363 if ``op1`` is less than ``op2``.
6364 #. ``sle``: interprets the operands as signed values and yields ``true``
6365 if ``op1`` is less than or equal to ``op2``.
6367 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
6368 are compared as if they were integers.
6370 If the operands are integer vectors, then they are compared element by
6371 element. The result is an ``i1`` vector with the same number of elements
6372 as the values being compared. Otherwise, the result is an ``i1``.
6377 .. code-block:: llvm
6379 <result> = icmp eq i32 4, 5 ; yields: result=false
6380 <result> = icmp ne float* %X, %X ; yields: result=false
6381 <result> = icmp ult i16 4, 5 ; yields: result=true
6382 <result> = icmp sgt i16 4, 5 ; yields: result=false
6383 <result> = icmp ule i16 -4, 5 ; yields: result=false
6384 <result> = icmp sge i16 4, 5 ; yields: result=false
6386 Note that the code generator does not yet support vector types with the
6387 ``icmp`` instruction.
6391 '``fcmp``' Instruction
6392 ^^^^^^^^^^^^^^^^^^^^^^
6399 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6404 The '``fcmp``' instruction returns a boolean value or vector of boolean
6405 values based on comparison of its operands.
6407 If the operands are floating point scalars, then the result type is a
6408 boolean (:ref:`i1 <t_integer>`).
6410 If the operands are floating point vectors, then the result type is a
6411 vector of boolean with the same number of elements as the operands being
6417 The '``fcmp``' instruction takes three operands. The first operand is
6418 the condition code indicating the kind of comparison to perform. It is
6419 not a value, just a keyword. The possible condition code are:
6421 #. ``false``: no comparison, always returns false
6422 #. ``oeq``: ordered and equal
6423 #. ``ogt``: ordered and greater than
6424 #. ``oge``: ordered and greater than or equal
6425 #. ``olt``: ordered and less than
6426 #. ``ole``: ordered and less than or equal
6427 #. ``one``: ordered and not equal
6428 #. ``ord``: ordered (no nans)
6429 #. ``ueq``: unordered or equal
6430 #. ``ugt``: unordered or greater than
6431 #. ``uge``: unordered or greater than or equal
6432 #. ``ult``: unordered or less than
6433 #. ``ule``: unordered or less than or equal
6434 #. ``une``: unordered or not equal
6435 #. ``uno``: unordered (either nans)
6436 #. ``true``: no comparison, always returns true
6438 *Ordered* means that neither operand is a QNAN while *unordered* means
6439 that either operand may be a QNAN.
6441 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
6442 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
6443 type. They must have identical types.
6448 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
6449 condition code given as ``cond``. If the operands are vectors, then the
6450 vectors are compared element by element. Each comparison performed
6451 always yields an :ref:`i1 <t_integer>` result, as follows:
6453 #. ``false``: always yields ``false``, regardless of operands.
6454 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
6455 is equal to ``op2``.
6456 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
6457 is greater than ``op2``.
6458 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
6459 is greater than or equal to ``op2``.
6460 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
6461 is less than ``op2``.
6462 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
6463 is less than or equal to ``op2``.
6464 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
6465 is not equal to ``op2``.
6466 #. ``ord``: yields ``true`` if both operands are not a QNAN.
6467 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
6469 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
6470 greater than ``op2``.
6471 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
6472 greater than or equal to ``op2``.
6473 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
6475 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
6476 less than or equal to ``op2``.
6477 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
6478 not equal to ``op2``.
6479 #. ``uno``: yields ``true`` if either operand is a QNAN.
6480 #. ``true``: always yields ``true``, regardless of operands.
6485 .. code-block:: llvm
6487 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
6488 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
6489 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
6490 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
6492 Note that the code generator does not yet support vector types with the
6493 ``fcmp`` instruction.
6497 '``phi``' Instruction
6498 ^^^^^^^^^^^^^^^^^^^^^
6505 <result> = phi <ty> [ <val0>, <label0>], ...
6510 The '``phi``' instruction is used to implement the φ node in the SSA
6511 graph representing the function.
6516 The type of the incoming values is specified with the first type field.
6517 After this, the '``phi``' instruction takes a list of pairs as
6518 arguments, with one pair for each predecessor basic block of the current
6519 block. Only values of :ref:`first class <t_firstclass>` type may be used as
6520 the value arguments to the PHI node. Only labels may be used as the
6523 There must be no non-phi instructions between the start of a basic block
6524 and the PHI instructions: i.e. PHI instructions must be first in a basic
6527 For the purposes of the SSA form, the use of each incoming value is
6528 deemed to occur on the edge from the corresponding predecessor block to
6529 the current block (but after any definition of an '``invoke``'
6530 instruction's return value on the same edge).
6535 At runtime, the '``phi``' instruction logically takes on the value
6536 specified by the pair corresponding to the predecessor basic block that
6537 executed just prior to the current block.
6542 .. code-block:: llvm
6544 Loop: ; Infinite loop that counts from 0 on up...
6545 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
6546 %nextindvar = add i32 %indvar, 1
6551 '``select``' Instruction
6552 ^^^^^^^^^^^^^^^^^^^^^^^^
6559 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
6561 selty is either i1 or {<N x i1>}
6566 The '``select``' instruction is used to choose one value based on a
6567 condition, without IR-level branching.
6572 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
6573 values indicating the condition, and two values of the same :ref:`first
6574 class <t_firstclass>` type. If the val1/val2 are vectors and the
6575 condition is a scalar, then entire vectors are selected, not individual
6581 If the condition is an i1 and it evaluates to 1, the instruction returns
6582 the first value argument; otherwise, it returns the second value
6585 If the condition is a vector of i1, then the value arguments must be
6586 vectors of the same size, and the selection is done element by element.
6591 .. code-block:: llvm
6593 %X = select i1 true, i8 17, i8 42 ; yields i8:17
6597 '``call``' Instruction
6598 ^^^^^^^^^^^^^^^^^^^^^^
6605 <result> = [tail | musttail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
6610 The '``call``' instruction represents a simple function call.
6615 This instruction requires several arguments:
6617 #. The optional ``tail`` and ``musttail`` markers indicate that the optimizers
6618 should perform tail call optimization. The ``tail`` marker is a hint that
6619 `can be ignored <CodeGenerator.html#sibcallopt>`_. The ``musttail`` marker
6620 means that the call must be tail call optimized in order for the program to
6621 be correct. The ``musttail`` marker provides these guarantees:
6623 #. The call will not cause unbounded stack growth if it is part of a
6624 recursive cycle in the call graph.
6625 #. Arguments with the :ref:`inalloca <attr_inalloca>` attribute are
6628 Both markers imply that the callee does not access allocas or varargs from
6629 the caller. Calls marked ``musttail`` must obey the following additional
6632 - The call must immediately precede a :ref:`ret <i_ret>` instruction,
6633 or a pointer bitcast followed by a ret instruction.
6634 - The ret instruction must return the (possibly bitcasted) value
6635 produced by the call or void.
6636 - The caller and callee prototypes must match. Pointer types of
6637 parameters or return types may differ in pointee type, but not
6639 - The calling conventions of the caller and callee must match.
6640 - All ABI-impacting function attributes, such as sret, byval, inreg,
6641 returned, and inalloca, must match.
6642 - The callee must be varargs iff the caller is varargs. Bitcasting a
6643 non-varargs function to the appropriate varargs type is legal so
6644 long as the non-varargs prefixes obey the other rules.
6646 Tail call optimization for calls marked ``tail`` is guaranteed to occur if
6647 the following conditions are met:
6649 - Caller and callee both have the calling convention ``fastcc``.
6650 - The call is in tail position (ret immediately follows call and ret
6651 uses value of call or is void).
6652 - Option ``-tailcallopt`` is enabled, or
6653 ``llvm::GuaranteedTailCallOpt`` is ``true``.
6654 - `Platform-specific constraints are
6655 met. <CodeGenerator.html#tailcallopt>`_
6657 #. The optional "cconv" marker indicates which :ref:`calling
6658 convention <callingconv>` the call should use. If none is
6659 specified, the call defaults to using C calling conventions. The
6660 calling convention of the call must match the calling convention of
6661 the target function, or else the behavior is undefined.
6662 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
6663 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
6665 #. '``ty``': the type of the call instruction itself which is also the
6666 type of the return value. Functions that return no value are marked
6668 #. '``fnty``': shall be the signature of the pointer to function value
6669 being invoked. The argument types must match the types implied by
6670 this signature. This type can be omitted if the function is not
6671 varargs and if the function type does not return a pointer to a
6673 #. '``fnptrval``': An LLVM value containing a pointer to a function to
6674 be invoked. In most cases, this is a direct function invocation, but
6675 indirect ``call``'s are just as possible, calling an arbitrary pointer
6677 #. '``function args``': argument list whose types match the function
6678 signature argument types and parameter attributes. All arguments must
6679 be of :ref:`first class <t_firstclass>` type. If the function signature
6680 indicates the function accepts a variable number of arguments, the
6681 extra arguments can be specified.
6682 #. The optional :ref:`function attributes <fnattrs>` list. Only
6683 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
6684 attributes are valid here.
6689 The '``call``' instruction is used to cause control flow to transfer to
6690 a specified function, with its incoming arguments bound to the specified
6691 values. Upon a '``ret``' instruction in the called function, control
6692 flow continues with the instruction after the function call, and the
6693 return value of the function is bound to the result argument.
6698 .. code-block:: llvm
6700 %retval = call i32 @test(i32 %argc)
6701 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
6702 %X = tail call i32 @foo() ; yields i32
6703 %Y = tail call fastcc i32 @foo() ; yields i32
6704 call void %foo(i8 97 signext)
6706 %struct.A = type { i32, i8 }
6707 %r = call %struct.A @foo() ; yields { i32, i8 }
6708 %gr = extractvalue %struct.A %r, 0 ; yields i32
6709 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
6710 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
6711 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
6713 llvm treats calls to some functions with names and arguments that match
6714 the standard C99 library as being the C99 library functions, and may
6715 perform optimizations or generate code for them under that assumption.
6716 This is something we'd like to change in the future to provide better
6717 support for freestanding environments and non-C-based languages.
6721 '``va_arg``' Instruction
6722 ^^^^^^^^^^^^^^^^^^^^^^^^
6729 <resultval> = va_arg <va_list*> <arglist>, <argty>
6734 The '``va_arg``' instruction is used to access arguments passed through
6735 the "variable argument" area of a function call. It is used to implement
6736 the ``va_arg`` macro in C.
6741 This instruction takes a ``va_list*`` value and the type of the
6742 argument. It returns a value of the specified argument type and
6743 increments the ``va_list`` to point to the next argument. The actual
6744 type of ``va_list`` is target specific.
6749 The '``va_arg``' instruction loads an argument of the specified type
6750 from the specified ``va_list`` and causes the ``va_list`` to point to
6751 the next argument. For more information, see the variable argument
6752 handling :ref:`Intrinsic Functions <int_varargs>`.
6754 It is legal for this instruction to be called in a function which does
6755 not take a variable number of arguments, for example, the ``vfprintf``
6758 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
6759 function <intrinsics>` because it takes a type as an argument.
6764 See the :ref:`variable argument processing <int_varargs>` section.
6766 Note that the code generator does not yet fully support va\_arg on many
6767 targets. Also, it does not currently support va\_arg with aggregate
6768 types on any target.
6772 '``landingpad``' Instruction
6773 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6780 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
6781 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
6783 <clause> := catch <type> <value>
6784 <clause> := filter <array constant type> <array constant>
6789 The '``landingpad``' instruction is used by `LLVM's exception handling
6790 system <ExceptionHandling.html#overview>`_ to specify that a basic block
6791 is a landing pad --- one where the exception lands, and corresponds to the
6792 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
6793 defines values supplied by the personality function (``pers_fn``) upon
6794 re-entry to the function. The ``resultval`` has the type ``resultty``.
6799 This instruction takes a ``pers_fn`` value. This is the personality
6800 function associated with the unwinding mechanism. The optional
6801 ``cleanup`` flag indicates that the landing pad block is a cleanup.
6803 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
6804 contains the global variable representing the "type" that may be caught
6805 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
6806 clause takes an array constant as its argument. Use
6807 "``[0 x i8**] undef``" for a filter which cannot throw. The
6808 '``landingpad``' instruction must contain *at least* one ``clause`` or
6809 the ``cleanup`` flag.
6814 The '``landingpad``' instruction defines the values which are set by the
6815 personality function (``pers_fn``) upon re-entry to the function, and
6816 therefore the "result type" of the ``landingpad`` instruction. As with
6817 calling conventions, how the personality function results are
6818 represented in LLVM IR is target specific.
6820 The clauses are applied in order from top to bottom. If two
6821 ``landingpad`` instructions are merged together through inlining, the
6822 clauses from the calling function are appended to the list of clauses.
6823 When the call stack is being unwound due to an exception being thrown,
6824 the exception is compared against each ``clause`` in turn. If it doesn't
6825 match any of the clauses, and the ``cleanup`` flag is not set, then
6826 unwinding continues further up the call stack.
6828 The ``landingpad`` instruction has several restrictions:
6830 - A landing pad block is a basic block which is the unwind destination
6831 of an '``invoke``' instruction.
6832 - A landing pad block must have a '``landingpad``' instruction as its
6833 first non-PHI instruction.
6834 - There can be only one '``landingpad``' instruction within the landing
6836 - A basic block that is not a landing pad block may not include a
6837 '``landingpad``' instruction.
6838 - All '``landingpad``' instructions in a function must have the same
6839 personality function.
6844 .. code-block:: llvm
6846 ;; A landing pad which can catch an integer.
6847 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6849 ;; A landing pad that is a cleanup.
6850 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6852 ;; A landing pad which can catch an integer and can only throw a double.
6853 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6855 filter [1 x i8**] [@_ZTId]
6862 LLVM supports the notion of an "intrinsic function". These functions
6863 have well known names and semantics and are required to follow certain
6864 restrictions. Overall, these intrinsics represent an extension mechanism
6865 for the LLVM language that does not require changing all of the
6866 transformations in LLVM when adding to the language (or the bitcode
6867 reader/writer, the parser, etc...).
6869 Intrinsic function names must all start with an "``llvm.``" prefix. This
6870 prefix is reserved in LLVM for intrinsic names; thus, function names may
6871 not begin with this prefix. Intrinsic functions must always be external
6872 functions: you cannot define the body of intrinsic functions. Intrinsic
6873 functions may only be used in call or invoke instructions: it is illegal
6874 to take the address of an intrinsic function. Additionally, because
6875 intrinsic functions are part of the LLVM language, it is required if any
6876 are added that they be documented here.
6878 Some intrinsic functions can be overloaded, i.e., the intrinsic
6879 represents a family of functions that perform the same operation but on
6880 different data types. Because LLVM can represent over 8 million
6881 different integer types, overloading is used commonly to allow an
6882 intrinsic function to operate on any integer type. One or more of the
6883 argument types or the result type can be overloaded to accept any
6884 integer type. Argument types may also be defined as exactly matching a
6885 previous argument's type or the result type. This allows an intrinsic
6886 function which accepts multiple arguments, but needs all of them to be
6887 of the same type, to only be overloaded with respect to a single
6888 argument or the result.
6890 Overloaded intrinsics will have the names of its overloaded argument
6891 types encoded into its function name, each preceded by a period. Only
6892 those types which are overloaded result in a name suffix. Arguments
6893 whose type is matched against another type do not. For example, the
6894 ``llvm.ctpop`` function can take an integer of any width and returns an
6895 integer of exactly the same integer width. This leads to a family of
6896 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
6897 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
6898 overloaded, and only one type suffix is required. Because the argument's
6899 type is matched against the return type, it does not require its own
6902 To learn how to add an intrinsic function, please see the `Extending
6903 LLVM Guide <ExtendingLLVM.html>`_.
6907 Variable Argument Handling Intrinsics
6908 -------------------------------------
6910 Variable argument support is defined in LLVM with the
6911 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
6912 functions. These functions are related to the similarly named macros
6913 defined in the ``<stdarg.h>`` header file.
6915 All of these functions operate on arguments that use a target-specific
6916 value type "``va_list``". The LLVM assembly language reference manual
6917 does not define what this type is, so all transformations should be
6918 prepared to handle these functions regardless of the type used.
6920 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
6921 variable argument handling intrinsic functions are used.
6923 .. code-block:: llvm
6925 ; This struct is different for every platform. For most platforms,
6926 ; it is merely an i8*.
6927 %struct.va_list = type { i8* }
6929 ; For Unix x86_64 platforms, va_list is the following struct:
6930 ; %struct.va_list = type { i32, i32, i8*, i8* }
6932 define i32 @test(i32 %X, ...) {
6933 ; Initialize variable argument processing
6934 %ap = alloca %struct.va_list
6935 %ap2 = bitcast %struct.va_list* %ap to i8*
6936 call void @llvm.va_start(i8* %ap2)
6938 ; Read a single integer argument
6939 %tmp = va_arg i8* %ap2, i32
6941 ; Demonstrate usage of llvm.va_copy and llvm.va_end
6943 %aq2 = bitcast i8** %aq to i8*
6944 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
6945 call void @llvm.va_end(i8* %aq2)
6947 ; Stop processing of arguments.
6948 call void @llvm.va_end(i8* %ap2)
6952 declare void @llvm.va_start(i8*)
6953 declare void @llvm.va_copy(i8*, i8*)
6954 declare void @llvm.va_end(i8*)
6958 '``llvm.va_start``' Intrinsic
6959 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6966 declare void @llvm.va_start(i8* <arglist>)
6971 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
6972 subsequent use by ``va_arg``.
6977 The argument is a pointer to a ``va_list`` element to initialize.
6982 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
6983 available in C. In a target-dependent way, it initializes the
6984 ``va_list`` element to which the argument points, so that the next call
6985 to ``va_arg`` will produce the first variable argument passed to the
6986 function. Unlike the C ``va_start`` macro, this intrinsic does not need
6987 to know the last argument of the function as the compiler can figure
6990 '``llvm.va_end``' Intrinsic
6991 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6998 declare void @llvm.va_end(i8* <arglist>)
7003 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
7004 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
7009 The argument is a pointer to a ``va_list`` to destroy.
7014 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
7015 available in C. In a target-dependent way, it destroys the ``va_list``
7016 element to which the argument points. Calls to
7017 :ref:`llvm.va_start <int_va_start>` and
7018 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
7023 '``llvm.va_copy``' Intrinsic
7024 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7031 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
7036 The '``llvm.va_copy``' intrinsic copies the current argument position
7037 from the source argument list to the destination argument list.
7042 The first argument is a pointer to a ``va_list`` element to initialize.
7043 The second argument is a pointer to a ``va_list`` element to copy from.
7048 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
7049 available in C. In a target-dependent way, it copies the source
7050 ``va_list`` element into the destination ``va_list`` element. This
7051 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
7052 arbitrarily complex and require, for example, memory allocation.
7054 Accurate Garbage Collection Intrinsics
7055 --------------------------------------
7057 LLVM support for `Accurate Garbage Collection <GarbageCollection.html>`_
7058 (GC) requires the implementation and generation of these intrinsics.
7059 These intrinsics allow identification of :ref:`GC roots on the
7060 stack <int_gcroot>`, as well as garbage collector implementations that
7061 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
7062 Front-ends for type-safe garbage collected languages should generate
7063 these intrinsics to make use of the LLVM garbage collectors. For more
7064 details, see `Accurate Garbage Collection with
7065 LLVM <GarbageCollection.html>`_.
7067 The garbage collection intrinsics only operate on objects in the generic
7068 address space (address space zero).
7072 '``llvm.gcroot``' Intrinsic
7073 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7080 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
7085 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
7086 the code generator, and allows some metadata to be associated with it.
7091 The first argument specifies the address of a stack object that contains
7092 the root pointer. The second pointer (which must be either a constant or
7093 a global value address) contains the meta-data to be associated with the
7099 At runtime, a call to this intrinsic stores a null pointer into the
7100 "ptrloc" location. At compile-time, the code generator generates
7101 information to allow the runtime to find the pointer at GC safe points.
7102 The '``llvm.gcroot``' intrinsic may only be used in a function which
7103 :ref:`specifies a GC algorithm <gc>`.
7107 '``llvm.gcread``' Intrinsic
7108 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7115 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
7120 The '``llvm.gcread``' intrinsic identifies reads of references from heap
7121 locations, allowing garbage collector implementations that require read
7127 The second argument is the address to read from, which should be an
7128 address allocated from the garbage collector. The first object is a
7129 pointer to the start of the referenced object, if needed by the language
7130 runtime (otherwise null).
7135 The '``llvm.gcread``' intrinsic has the same semantics as a load
7136 instruction, but may be replaced with substantially more complex code by
7137 the garbage collector runtime, as needed. The '``llvm.gcread``'
7138 intrinsic may only be used in a function which :ref:`specifies a GC
7143 '``llvm.gcwrite``' Intrinsic
7144 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7151 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
7156 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
7157 locations, allowing garbage collector implementations that require write
7158 barriers (such as generational or reference counting collectors).
7163 The first argument is the reference to store, the second is the start of
7164 the object to store it to, and the third is the address of the field of
7165 Obj to store to. If the runtime does not require a pointer to the
7166 object, Obj may be null.
7171 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
7172 instruction, but may be replaced with substantially more complex code by
7173 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
7174 intrinsic may only be used in a function which :ref:`specifies a GC
7177 Code Generator Intrinsics
7178 -------------------------
7180 These intrinsics are provided by LLVM to expose special features that
7181 may only be implemented with code generator support.
7183 '``llvm.returnaddress``' Intrinsic
7184 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7191 declare i8 *@llvm.returnaddress(i32 <level>)
7196 The '``llvm.returnaddress``' intrinsic attempts to compute a
7197 target-specific value indicating the return address of the current
7198 function or one of its callers.
7203 The argument to this intrinsic indicates which function to return the
7204 address for. Zero indicates the calling function, one indicates its
7205 caller, etc. The argument is **required** to be a constant integer
7211 The '``llvm.returnaddress``' intrinsic either returns a pointer
7212 indicating the return address of the specified call frame, or zero if it
7213 cannot be identified. The value returned by this intrinsic is likely to
7214 be incorrect or 0 for arguments other than zero, so it should only be
7215 used for debugging purposes.
7217 Note that calling this intrinsic does not prevent function inlining or
7218 other aggressive transformations, so the value returned may not be that
7219 of the obvious source-language caller.
7221 '``llvm.frameaddress``' Intrinsic
7222 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7229 declare i8* @llvm.frameaddress(i32 <level>)
7234 The '``llvm.frameaddress``' intrinsic attempts to return the
7235 target-specific frame pointer value for the specified stack frame.
7240 The argument to this intrinsic indicates which function to return the
7241 frame pointer for. Zero indicates the calling function, one indicates
7242 its caller, etc. The argument is **required** to be a constant integer
7248 The '``llvm.frameaddress``' intrinsic either returns a pointer
7249 indicating the frame address of the specified call frame, or zero if it
7250 cannot be identified. The value returned by this intrinsic is likely to
7251 be incorrect or 0 for arguments other than zero, so it should only be
7252 used for debugging purposes.
7254 Note that calling this intrinsic does not prevent function inlining or
7255 other aggressive transformations, so the value returned may not be that
7256 of the obvious source-language caller.
7258 .. _int_read_register:
7259 .. _int_write_register:
7261 '``llvm.read_register``' and '``llvm.write_register``' Intrinsics
7262 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7269 declare i32 @llvm.read_register.i32(metadata)
7270 declare i64 @llvm.read_register.i64(metadata)
7271 declare void @llvm.write_register.i32(metadata, i32 @value)
7272 declare void @llvm.write_register.i64(metadata, i64 @value)
7273 !0 = metadata !{metadata !"sp\00"}
7278 The '``llvm.read_register``' and '``llvm.write_register``' intrinsics
7279 provides access to the named register. The register must be valid on
7280 the architecture being compiled to. The type needs to be compatible
7281 with the register being read.
7286 The '``llvm.read_register``' intrinsic returns the current value of the
7287 register, where possible. The '``llvm.write_register``' intrinsic sets
7288 the current value of the register, where possible.
7290 This is useful to implement named register global variables that need
7291 to always be mapped to a specific register, as is common practice on
7292 bare-metal programs including OS kernels.
7294 The compiler doesn't check for register availability or use of the used
7295 register in surrounding code, including inline assembly. Because of that,
7296 allocatable registers are not supported.
7298 Warning: So far it only works with the stack pointer on selected
7299 architectures (ARM, AArch64, PowerPC and x86_64). Significant amount of
7300 work is needed to support other registers and even more so, allocatable
7305 '``llvm.stacksave``' Intrinsic
7306 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7313 declare i8* @llvm.stacksave()
7318 The '``llvm.stacksave``' intrinsic is used to remember the current state
7319 of the function stack, for use with
7320 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
7321 implementing language features like scoped automatic variable sized
7327 This intrinsic returns a opaque pointer value that can be passed to
7328 :ref:`llvm.stackrestore <int_stackrestore>`. When an
7329 ``llvm.stackrestore`` intrinsic is executed with a value saved from
7330 ``llvm.stacksave``, it effectively restores the state of the stack to
7331 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
7332 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
7333 were allocated after the ``llvm.stacksave`` was executed.
7335 .. _int_stackrestore:
7337 '``llvm.stackrestore``' Intrinsic
7338 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7345 declare void @llvm.stackrestore(i8* %ptr)
7350 The '``llvm.stackrestore``' intrinsic is used to restore the state of
7351 the function stack to the state it was in when the corresponding
7352 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
7353 useful for implementing language features like scoped automatic variable
7354 sized arrays in C99.
7359 See the description for :ref:`llvm.stacksave <int_stacksave>`.
7361 '``llvm.prefetch``' Intrinsic
7362 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7369 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
7374 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
7375 insert a prefetch instruction if supported; otherwise, it is a noop.
7376 Prefetches have no effect on the behavior of the program but can change
7377 its performance characteristics.
7382 ``address`` is the address to be prefetched, ``rw`` is the specifier
7383 determining if the fetch should be for a read (0) or write (1), and
7384 ``locality`` is a temporal locality specifier ranging from (0) - no
7385 locality, to (3) - extremely local keep in cache. The ``cache type``
7386 specifies whether the prefetch is performed on the data (1) or
7387 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
7388 arguments must be constant integers.
7393 This intrinsic does not modify the behavior of the program. In
7394 particular, prefetches cannot trap and do not produce a value. On
7395 targets that support this intrinsic, the prefetch can provide hints to
7396 the processor cache for better performance.
7398 '``llvm.pcmarker``' Intrinsic
7399 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7406 declare void @llvm.pcmarker(i32 <id>)
7411 The '``llvm.pcmarker``' intrinsic is a method to export a Program
7412 Counter (PC) in a region of code to simulators and other tools. The
7413 method is target specific, but it is expected that the marker will use
7414 exported symbols to transmit the PC of the marker. The marker makes no
7415 guarantees that it will remain with any specific instruction after
7416 optimizations. It is possible that the presence of a marker will inhibit
7417 optimizations. The intended use is to be inserted after optimizations to
7418 allow correlations of simulation runs.
7423 ``id`` is a numerical id identifying the marker.
7428 This intrinsic does not modify the behavior of the program. Backends
7429 that do not support this intrinsic may ignore it.
7431 '``llvm.readcyclecounter``' Intrinsic
7432 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7439 declare i64 @llvm.readcyclecounter()
7444 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
7445 counter register (or similar low latency, high accuracy clocks) on those
7446 targets that support it. On X86, it should map to RDTSC. On Alpha, it
7447 should map to RPCC. As the backing counters overflow quickly (on the
7448 order of 9 seconds on alpha), this should only be used for small
7454 When directly supported, reading the cycle counter should not modify any
7455 memory. Implementations are allowed to either return a application
7456 specific value or a system wide value. On backends without support, this
7457 is lowered to a constant 0.
7459 Note that runtime support may be conditional on the privilege-level code is
7460 running at and the host platform.
7462 '``llvm.clear_cache``' Intrinsic
7463 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7470 declare void @llvm.clear_cache(i8*, i8*)
7475 The '``llvm.clear_cache``' intrinsic ensures visibility of modifications
7476 in the specified range to the execution unit of the processor. On
7477 targets with non-unified instruction and data cache, the implementation
7478 flushes the instruction cache.
7483 On platforms with coherent instruction and data caches (e.g. x86), this
7484 intrinsic is a nop. On platforms with non-coherent instruction and data
7485 cache (e.g. ARM, MIPS), the intrinsic is lowered either to appropriate
7486 instructions or a system call, if cache flushing requires special
7489 The default behavior is to emit a call to ``__clear_cache`` from the run
7492 This instrinsic does *not* empty the instruction pipeline. Modifications
7493 of the current function are outside the scope of the intrinsic.
7495 Standard C Library Intrinsics
7496 -----------------------------
7498 LLVM provides intrinsics for a few important standard C library
7499 functions. These intrinsics allow source-language front-ends to pass
7500 information about the alignment of the pointer arguments to the code
7501 generator, providing opportunity for more efficient code generation.
7505 '``llvm.memcpy``' Intrinsic
7506 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7511 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
7512 integer bit width and for different address spaces. Not all targets
7513 support all bit widths however.
7517 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7518 i32 <len>, i32 <align>, i1 <isvolatile>)
7519 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7520 i64 <len>, i32 <align>, i1 <isvolatile>)
7525 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7526 source location to the destination location.
7528 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
7529 intrinsics do not return a value, takes extra alignment/isvolatile
7530 arguments and the pointers can be in specified address spaces.
7535 The first argument is a pointer to the destination, the second is a
7536 pointer to the source. The third argument is an integer argument
7537 specifying the number of bytes to copy, the fourth argument is the
7538 alignment of the source and destination locations, and the fifth is a
7539 boolean indicating a volatile access.
7541 If the call to this intrinsic has an alignment value that is not 0 or 1,
7542 then the caller guarantees that both the source and destination pointers
7543 are aligned to that boundary.
7545 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
7546 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7547 very cleanly specified and it is unwise to depend on it.
7552 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7553 source location to the destination location, which are not allowed to
7554 overlap. It copies "len" bytes of memory over. If the argument is known
7555 to be aligned to some boundary, this can be specified as the fourth
7556 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
7558 '``llvm.memmove``' Intrinsic
7559 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7564 This is an overloaded intrinsic. You can use llvm.memmove on any integer
7565 bit width and for different address space. Not all targets support all
7570 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7571 i32 <len>, i32 <align>, i1 <isvolatile>)
7572 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7573 i64 <len>, i32 <align>, i1 <isvolatile>)
7578 The '``llvm.memmove.*``' intrinsics move a block of memory from the
7579 source location to the destination location. It is similar to the
7580 '``llvm.memcpy``' intrinsic but allows the two memory locations to
7583 Note that, unlike the standard libc function, the ``llvm.memmove.*``
7584 intrinsics do not return a value, takes extra alignment/isvolatile
7585 arguments and the pointers can be in specified address spaces.
7590 The first argument is a pointer to the destination, the second is a
7591 pointer to the source. The third argument is an integer argument
7592 specifying the number of bytes to copy, the fourth argument is the
7593 alignment of the source and destination locations, and the fifth is a
7594 boolean indicating a volatile access.
7596 If the call to this intrinsic has an alignment value that is not 0 or 1,
7597 then the caller guarantees that the source and destination pointers are
7598 aligned to that boundary.
7600 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
7601 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
7602 not very cleanly specified and it is unwise to depend on it.
7607 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
7608 source location to the destination location, which may overlap. It
7609 copies "len" bytes of memory over. If the argument is known to be
7610 aligned to some boundary, this can be specified as the fourth argument,
7611 otherwise it should be set to 0 or 1 (both meaning no alignment).
7613 '``llvm.memset.*``' Intrinsics
7614 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7619 This is an overloaded intrinsic. You can use llvm.memset on any integer
7620 bit width and for different address spaces. However, not all targets
7621 support all bit widths.
7625 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
7626 i32 <len>, i32 <align>, i1 <isvolatile>)
7627 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
7628 i64 <len>, i32 <align>, i1 <isvolatile>)
7633 The '``llvm.memset.*``' intrinsics fill a block of memory with a
7634 particular byte value.
7636 Note that, unlike the standard libc function, the ``llvm.memset``
7637 intrinsic does not return a value and takes extra alignment/volatile
7638 arguments. Also, the destination can be in an arbitrary address space.
7643 The first argument is a pointer to the destination to fill, the second
7644 is the byte value with which to fill it, the third argument is an
7645 integer argument specifying the number of bytes to fill, and the fourth
7646 argument is the known alignment of the destination location.
7648 If the call to this intrinsic has an alignment value that is not 0 or 1,
7649 then the caller guarantees that the destination pointer is aligned to
7652 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
7653 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7654 very cleanly specified and it is unwise to depend on it.
7659 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
7660 at the destination location. If the argument is known to be aligned to
7661 some boundary, this can be specified as the fourth argument, otherwise
7662 it should be set to 0 or 1 (both meaning no alignment).
7664 '``llvm.sqrt.*``' Intrinsic
7665 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7670 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
7671 floating point or vector of floating point type. Not all targets support
7676 declare float @llvm.sqrt.f32(float %Val)
7677 declare double @llvm.sqrt.f64(double %Val)
7678 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
7679 declare fp128 @llvm.sqrt.f128(fp128 %Val)
7680 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
7685 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
7686 returning the same value as the libm '``sqrt``' functions would. Unlike
7687 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
7688 negative numbers other than -0.0 (which allows for better optimization,
7689 because there is no need to worry about errno being set).
7690 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
7695 The argument and return value are floating point numbers of the same
7701 This function returns the sqrt of the specified operand if it is a
7702 nonnegative floating point number.
7704 '``llvm.powi.*``' Intrinsic
7705 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7710 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
7711 floating point or vector of floating point type. Not all targets support
7716 declare float @llvm.powi.f32(float %Val, i32 %power)
7717 declare double @llvm.powi.f64(double %Val, i32 %power)
7718 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7719 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7720 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
7725 The '``llvm.powi.*``' intrinsics return the first operand raised to the
7726 specified (positive or negative) power. The order of evaluation of
7727 multiplications is not defined. When a vector of floating point type is
7728 used, the second argument remains a scalar integer value.
7733 The second argument is an integer power, and the first is a value to
7734 raise to that power.
7739 This function returns the first value raised to the second power with an
7740 unspecified sequence of rounding operations.
7742 '``llvm.sin.*``' Intrinsic
7743 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7748 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
7749 floating point or vector of floating point type. Not all targets support
7754 declare float @llvm.sin.f32(float %Val)
7755 declare double @llvm.sin.f64(double %Val)
7756 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7757 declare fp128 @llvm.sin.f128(fp128 %Val)
7758 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7763 The '``llvm.sin.*``' intrinsics return the sine of the operand.
7768 The argument and return value are floating point numbers of the same
7774 This function returns the sine of the specified operand, returning the
7775 same values as the libm ``sin`` functions would, and handles error
7776 conditions in the same way.
7778 '``llvm.cos.*``' Intrinsic
7779 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7784 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
7785 floating point or vector of floating point type. Not all targets support
7790 declare float @llvm.cos.f32(float %Val)
7791 declare double @llvm.cos.f64(double %Val)
7792 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7793 declare fp128 @llvm.cos.f128(fp128 %Val)
7794 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7799 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
7804 The argument and return value are floating point numbers of the same
7810 This function returns the cosine of the specified operand, returning the
7811 same values as the libm ``cos`` functions would, and handles error
7812 conditions in the same way.
7814 '``llvm.pow.*``' Intrinsic
7815 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7820 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
7821 floating point or vector of floating point type. Not all targets support
7826 declare float @llvm.pow.f32(float %Val, float %Power)
7827 declare double @llvm.pow.f64(double %Val, double %Power)
7828 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7829 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7830 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7835 The '``llvm.pow.*``' intrinsics return the first operand raised to the
7836 specified (positive or negative) power.
7841 The second argument is a floating point power, and the first is a value
7842 to raise to that power.
7847 This function returns the first value raised to the second power,
7848 returning the same values as the libm ``pow`` functions would, and
7849 handles error conditions in the same way.
7851 '``llvm.exp.*``' Intrinsic
7852 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7857 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
7858 floating point or vector of floating point type. Not all targets support
7863 declare float @llvm.exp.f32(float %Val)
7864 declare double @llvm.exp.f64(double %Val)
7865 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7866 declare fp128 @llvm.exp.f128(fp128 %Val)
7867 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7872 The '``llvm.exp.*``' intrinsics perform the exp function.
7877 The argument and return value are floating point numbers of the same
7883 This function returns the same values as the libm ``exp`` functions
7884 would, and handles error conditions in the same way.
7886 '``llvm.exp2.*``' Intrinsic
7887 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7892 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
7893 floating point or vector of floating point type. Not all targets support
7898 declare float @llvm.exp2.f32(float %Val)
7899 declare double @llvm.exp2.f64(double %Val)
7900 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
7901 declare fp128 @llvm.exp2.f128(fp128 %Val)
7902 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
7907 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
7912 The argument and return value are floating point numbers of the same
7918 This function returns the same values as the libm ``exp2`` functions
7919 would, and handles error conditions in the same way.
7921 '``llvm.log.*``' Intrinsic
7922 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7927 This is an overloaded intrinsic. You can use ``llvm.log`` on any
7928 floating point or vector of floating point type. Not all targets support
7933 declare float @llvm.log.f32(float %Val)
7934 declare double @llvm.log.f64(double %Val)
7935 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7936 declare fp128 @llvm.log.f128(fp128 %Val)
7937 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7942 The '``llvm.log.*``' intrinsics perform the log function.
7947 The argument and return value are floating point numbers of the same
7953 This function returns the same values as the libm ``log`` functions
7954 would, and handles error conditions in the same way.
7956 '``llvm.log10.*``' Intrinsic
7957 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7962 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
7963 floating point or vector of floating point type. Not all targets support
7968 declare float @llvm.log10.f32(float %Val)
7969 declare double @llvm.log10.f64(double %Val)
7970 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
7971 declare fp128 @llvm.log10.f128(fp128 %Val)
7972 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
7977 The '``llvm.log10.*``' intrinsics perform the log10 function.
7982 The argument and return value are floating point numbers of the same
7988 This function returns the same values as the libm ``log10`` functions
7989 would, and handles error conditions in the same way.
7991 '``llvm.log2.*``' Intrinsic
7992 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7997 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
7998 floating point or vector of floating point type. Not all targets support
8003 declare float @llvm.log2.f32(float %Val)
8004 declare double @llvm.log2.f64(double %Val)
8005 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
8006 declare fp128 @llvm.log2.f128(fp128 %Val)
8007 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
8012 The '``llvm.log2.*``' intrinsics perform the log2 function.
8017 The argument and return value are floating point numbers of the same
8023 This function returns the same values as the libm ``log2`` functions
8024 would, and handles error conditions in the same way.
8026 '``llvm.fma.*``' Intrinsic
8027 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8032 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
8033 floating point or vector of floating point type. Not all targets support
8038 declare float @llvm.fma.f32(float %a, float %b, float %c)
8039 declare double @llvm.fma.f64(double %a, double %b, double %c)
8040 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
8041 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
8042 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
8047 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
8053 The argument and return value are floating point numbers of the same
8059 This function returns the same values as the libm ``fma`` functions
8060 would, and does not set errno.
8062 '``llvm.fabs.*``' Intrinsic
8063 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8068 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
8069 floating point or vector of floating point type. Not all targets support
8074 declare float @llvm.fabs.f32(float %Val)
8075 declare double @llvm.fabs.f64(double %Val)
8076 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
8077 declare fp128 @llvm.fabs.f128(fp128 %Val)
8078 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
8083 The '``llvm.fabs.*``' intrinsics return the absolute value of the
8089 The argument and return value are floating point numbers of the same
8095 This function returns the same values as the libm ``fabs`` functions
8096 would, and handles error conditions in the same way.
8098 '``llvm.minnum.*``' Intrinsic
8099 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8104 This is an overloaded intrinsic. You can use ``llvm.minnum`` on any
8105 floating point or vector of floating point type. Not all targets support
8110 declare float @llvm.minnum.f32(float %Val0, float %Val1)
8111 declare double @llvm.minnum.f64(double %Val0, double %Val1)
8112 declare x86_fp80 @llvm.minnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
8113 declare fp128 @llvm.minnum.f128(fp128 %Val0, fp128 %Val1)
8114 declare ppc_fp128 @llvm.minnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
8119 The '``llvm.minnum.*``' intrinsics return the minimum of the two
8126 The arguments and return value are floating point numbers of the same
8132 Follows the IEEE-754 semantics for minNum, which also match for libm's
8135 If either operand is a NaN, returns the other non-NaN operand. Returns
8136 NaN only if both operands are NaN. If the operands compare equal,
8137 returns a value that compares equal to both operands. This means that
8138 fmin(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
8140 '``llvm.maxnum.*``' Intrinsic
8141 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8146 This is an overloaded intrinsic. You can use ``llvm.maxnum`` on any
8147 floating point or vector of floating point type. Not all targets support
8152 declare float @llvm.maxnum.f32(float %Val0, float %Val1l)
8153 declare double @llvm.maxnum.f64(double %Val0, double %Val1)
8154 declare x86_fp80 @llvm.maxnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
8155 declare fp128 @llvm.maxnum.f128(fp128 %Val0, fp128 %Val1)
8156 declare ppc_fp128 @llvm.maxnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
8161 The '``llvm.maxnum.*``' intrinsics return the maximum of the two
8168 The arguments and return value are floating point numbers of the same
8173 Follows the IEEE-754 semantics for maxNum, which also match for libm's
8176 If either operand is a NaN, returns the other non-NaN operand. Returns
8177 NaN only if both operands are NaN. If the operands compare equal,
8178 returns a value that compares equal to both operands. This means that
8179 fmax(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
8181 '``llvm.copysign.*``' Intrinsic
8182 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8187 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
8188 floating point or vector of floating point type. Not all targets support
8193 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
8194 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
8195 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
8196 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
8197 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
8202 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
8203 first operand and the sign of the second operand.
8208 The arguments and return value are floating point numbers of the same
8214 This function returns the same values as the libm ``copysign``
8215 functions would, and handles error conditions in the same way.
8217 '``llvm.floor.*``' Intrinsic
8218 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8223 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
8224 floating point or vector of floating point type. Not all targets support
8229 declare float @llvm.floor.f32(float %Val)
8230 declare double @llvm.floor.f64(double %Val)
8231 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
8232 declare fp128 @llvm.floor.f128(fp128 %Val)
8233 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
8238 The '``llvm.floor.*``' intrinsics return the floor of the operand.
8243 The argument and return value are floating point numbers of the same
8249 This function returns the same values as the libm ``floor`` functions
8250 would, and handles error conditions in the same way.
8252 '``llvm.ceil.*``' Intrinsic
8253 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8258 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
8259 floating point or vector of floating point type. Not all targets support
8264 declare float @llvm.ceil.f32(float %Val)
8265 declare double @llvm.ceil.f64(double %Val)
8266 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
8267 declare fp128 @llvm.ceil.f128(fp128 %Val)
8268 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
8273 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
8278 The argument and return value are floating point numbers of the same
8284 This function returns the same values as the libm ``ceil`` functions
8285 would, and handles error conditions in the same way.
8287 '``llvm.trunc.*``' Intrinsic
8288 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8293 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
8294 floating point or vector of floating point type. Not all targets support
8299 declare float @llvm.trunc.f32(float %Val)
8300 declare double @llvm.trunc.f64(double %Val)
8301 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
8302 declare fp128 @llvm.trunc.f128(fp128 %Val)
8303 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
8308 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
8309 nearest integer not larger in magnitude than the operand.
8314 The argument and return value are floating point numbers of the same
8320 This function returns the same values as the libm ``trunc`` functions
8321 would, and handles error conditions in the same way.
8323 '``llvm.rint.*``' Intrinsic
8324 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8329 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
8330 floating point or vector of floating point type. Not all targets support
8335 declare float @llvm.rint.f32(float %Val)
8336 declare double @llvm.rint.f64(double %Val)
8337 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
8338 declare fp128 @llvm.rint.f128(fp128 %Val)
8339 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
8344 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
8345 nearest integer. It may raise an inexact floating-point exception if the
8346 operand isn't an integer.
8351 The argument and return value are floating point numbers of the same
8357 This function returns the same values as the libm ``rint`` functions
8358 would, and handles error conditions in the same way.
8360 '``llvm.nearbyint.*``' Intrinsic
8361 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8366 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
8367 floating point or vector of floating point type. Not all targets support
8372 declare float @llvm.nearbyint.f32(float %Val)
8373 declare double @llvm.nearbyint.f64(double %Val)
8374 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
8375 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
8376 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
8381 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
8387 The argument and return value are floating point numbers of the same
8393 This function returns the same values as the libm ``nearbyint``
8394 functions would, and handles error conditions in the same way.
8396 '``llvm.round.*``' Intrinsic
8397 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8402 This is an overloaded intrinsic. You can use ``llvm.round`` on any
8403 floating point or vector of floating point type. Not all targets support
8408 declare float @llvm.round.f32(float %Val)
8409 declare double @llvm.round.f64(double %Val)
8410 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
8411 declare fp128 @llvm.round.f128(fp128 %Val)
8412 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
8417 The '``llvm.round.*``' intrinsics returns the operand rounded to the
8423 The argument and return value are floating point numbers of the same
8429 This function returns the same values as the libm ``round``
8430 functions would, and handles error conditions in the same way.
8432 Bit Manipulation Intrinsics
8433 ---------------------------
8435 LLVM provides intrinsics for a few important bit manipulation
8436 operations. These allow efficient code generation for some algorithms.
8438 '``llvm.bswap.*``' Intrinsics
8439 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8444 This is an overloaded intrinsic function. You can use bswap on any
8445 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
8449 declare i16 @llvm.bswap.i16(i16 <id>)
8450 declare i32 @llvm.bswap.i32(i32 <id>)
8451 declare i64 @llvm.bswap.i64(i64 <id>)
8456 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
8457 values with an even number of bytes (positive multiple of 16 bits).
8458 These are useful for performing operations on data that is not in the
8459 target's native byte order.
8464 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
8465 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
8466 intrinsic returns an i32 value that has the four bytes of the input i32
8467 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
8468 returned i32 will have its bytes in 3, 2, 1, 0 order. The
8469 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
8470 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
8473 '``llvm.ctpop.*``' Intrinsic
8474 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8479 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
8480 bit width, or on any vector with integer elements. Not all targets
8481 support all bit widths or vector types, however.
8485 declare i8 @llvm.ctpop.i8(i8 <src>)
8486 declare i16 @llvm.ctpop.i16(i16 <src>)
8487 declare i32 @llvm.ctpop.i32(i32 <src>)
8488 declare i64 @llvm.ctpop.i64(i64 <src>)
8489 declare i256 @llvm.ctpop.i256(i256 <src>)
8490 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
8495 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
8501 The only argument is the value to be counted. The argument may be of any
8502 integer type, or a vector with integer elements. The return type must
8503 match the argument type.
8508 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
8509 each element of a vector.
8511 '``llvm.ctlz.*``' Intrinsic
8512 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8517 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
8518 integer bit width, or any vector whose elements are integers. Not all
8519 targets support all bit widths or vector types, however.
8523 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
8524 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
8525 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
8526 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
8527 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
8528 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
8533 The '``llvm.ctlz``' family of intrinsic functions counts the number of
8534 leading zeros in a variable.
8539 The first argument is the value to be counted. This argument may be of
8540 any integer type, or a vectory with integer element type. The return
8541 type must match the first argument type.
8543 The second argument must be a constant and is a flag to indicate whether
8544 the intrinsic should ensure that a zero as the first argument produces a
8545 defined result. Historically some architectures did not provide a
8546 defined result for zero values as efficiently, and many algorithms are
8547 now predicated on avoiding zero-value inputs.
8552 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
8553 zeros in a variable, or within each element of the vector. If
8554 ``src == 0`` then the result is the size in bits of the type of ``src``
8555 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
8556 ``llvm.ctlz(i32 2) = 30``.
8558 '``llvm.cttz.*``' Intrinsic
8559 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8564 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
8565 integer bit width, or any vector of integer elements. Not all targets
8566 support all bit widths or vector types, however.
8570 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
8571 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
8572 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
8573 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
8574 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
8575 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
8580 The '``llvm.cttz``' family of intrinsic functions counts the number of
8586 The first argument is the value to be counted. This argument may be of
8587 any integer type, or a vectory with integer element type. The return
8588 type must match the first argument type.
8590 The second argument must be a constant and is a flag to indicate whether
8591 the intrinsic should ensure that a zero as the first argument produces a
8592 defined result. Historically some architectures did not provide a
8593 defined result for zero values as efficiently, and many algorithms are
8594 now predicated on avoiding zero-value inputs.
8599 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
8600 zeros in a variable, or within each element of a vector. If ``src == 0``
8601 then the result is the size in bits of the type of ``src`` if
8602 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
8603 ``llvm.cttz(2) = 1``.
8605 Arithmetic with Overflow Intrinsics
8606 -----------------------------------
8608 LLVM provides intrinsics for some arithmetic with overflow operations.
8610 '``llvm.sadd.with.overflow.*``' Intrinsics
8611 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8616 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
8617 on any integer bit width.
8621 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
8622 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8623 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
8628 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8629 a signed addition of the two arguments, and indicate whether an overflow
8630 occurred during the signed summation.
8635 The arguments (%a and %b) and the first element of the result structure
8636 may be of integer types of any bit width, but they must have the same
8637 bit width. The second element of the result structure must be of type
8638 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8644 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8645 a signed addition of the two variables. They return a structure --- the
8646 first element of which is the signed summation, and the second element
8647 of which is a bit specifying if the signed summation resulted in an
8653 .. code-block:: llvm
8655 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8656 %sum = extractvalue {i32, i1} %res, 0
8657 %obit = extractvalue {i32, i1} %res, 1
8658 br i1 %obit, label %overflow, label %normal
8660 '``llvm.uadd.with.overflow.*``' Intrinsics
8661 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8666 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
8667 on any integer bit width.
8671 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
8672 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8673 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
8678 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8679 an unsigned addition of the two arguments, and indicate whether a carry
8680 occurred during the unsigned summation.
8685 The arguments (%a and %b) and the first element of the result structure
8686 may be of integer types of any bit width, but they must have the same
8687 bit width. The second element of the result structure must be of type
8688 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8694 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8695 an unsigned addition of the two arguments. They return a structure --- the
8696 first element of which is the sum, and the second element of which is a
8697 bit specifying if the unsigned summation resulted in a carry.
8702 .. code-block:: llvm
8704 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8705 %sum = extractvalue {i32, i1} %res, 0
8706 %obit = extractvalue {i32, i1} %res, 1
8707 br i1 %obit, label %carry, label %normal
8709 '``llvm.ssub.with.overflow.*``' Intrinsics
8710 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8715 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
8716 on any integer bit width.
8720 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
8721 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8722 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
8727 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8728 a signed subtraction of the two arguments, and indicate whether an
8729 overflow occurred during the signed subtraction.
8734 The arguments (%a and %b) and the first element of the result structure
8735 may be of integer types of any bit width, but they must have the same
8736 bit width. The second element of the result structure must be of type
8737 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8743 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8744 a signed subtraction of the two arguments. They return a structure --- the
8745 first element of which is the subtraction, and the second element of
8746 which is a bit specifying if the signed subtraction resulted in an
8752 .. code-block:: llvm
8754 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8755 %sum = extractvalue {i32, i1} %res, 0
8756 %obit = extractvalue {i32, i1} %res, 1
8757 br i1 %obit, label %overflow, label %normal
8759 '``llvm.usub.with.overflow.*``' Intrinsics
8760 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8765 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
8766 on any integer bit width.
8770 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
8771 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8772 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
8777 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8778 an unsigned subtraction of the two arguments, and indicate whether an
8779 overflow occurred during the unsigned subtraction.
8784 The arguments (%a and %b) and the first element of the result structure
8785 may be of integer types of any bit width, but they must have the same
8786 bit width. The second element of the result structure must be of type
8787 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8793 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8794 an unsigned subtraction of the two arguments. They return a structure ---
8795 the first element of which is the subtraction, and the second element of
8796 which is a bit specifying if the unsigned subtraction resulted in an
8802 .. code-block:: llvm
8804 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8805 %sum = extractvalue {i32, i1} %res, 0
8806 %obit = extractvalue {i32, i1} %res, 1
8807 br i1 %obit, label %overflow, label %normal
8809 '``llvm.smul.with.overflow.*``' Intrinsics
8810 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8815 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
8816 on any integer bit width.
8820 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
8821 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8822 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
8827 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8828 a signed multiplication of the two arguments, and indicate whether an
8829 overflow occurred during the signed multiplication.
8834 The arguments (%a and %b) and the first element of the result structure
8835 may be of integer types of any bit width, but they must have the same
8836 bit width. The second element of the result structure must be of type
8837 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8843 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8844 a signed multiplication of the two arguments. They return a structure ---
8845 the first element of which is the multiplication, and the second element
8846 of which is a bit specifying if the signed multiplication resulted in an
8852 .. code-block:: llvm
8854 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8855 %sum = extractvalue {i32, i1} %res, 0
8856 %obit = extractvalue {i32, i1} %res, 1
8857 br i1 %obit, label %overflow, label %normal
8859 '``llvm.umul.with.overflow.*``' Intrinsics
8860 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8865 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
8866 on any integer bit width.
8870 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
8871 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8872 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
8877 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
8878 a unsigned multiplication of the two arguments, and indicate whether an
8879 overflow occurred during the unsigned multiplication.
8884 The arguments (%a and %b) and the first element of the result structure
8885 may be of integer types of any bit width, but they must have the same
8886 bit width. The second element of the result structure must be of type
8887 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8893 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
8894 an unsigned multiplication of the two arguments. They return a structure ---
8895 the first element of which is the multiplication, and the second
8896 element of which is a bit specifying if the unsigned multiplication
8897 resulted in an overflow.
8902 .. code-block:: llvm
8904 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8905 %sum = extractvalue {i32, i1} %res, 0
8906 %obit = extractvalue {i32, i1} %res, 1
8907 br i1 %obit, label %overflow, label %normal
8909 Specialised Arithmetic Intrinsics
8910 ---------------------------------
8912 '``llvm.fmuladd.*``' Intrinsic
8913 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8920 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
8921 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
8926 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
8927 expressions that can be fused if the code generator determines that (a) the
8928 target instruction set has support for a fused operation, and (b) that the
8929 fused operation is more efficient than the equivalent, separate pair of mul
8930 and add instructions.
8935 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
8936 multiplicands, a and b, and an addend c.
8945 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
8947 is equivalent to the expression a \* b + c, except that rounding will
8948 not be performed between the multiplication and addition steps if the
8949 code generator fuses the operations. Fusion is not guaranteed, even if
8950 the target platform supports it. If a fused multiply-add is required the
8951 corresponding llvm.fma.\* intrinsic function should be used
8952 instead. This never sets errno, just as '``llvm.fma.*``'.
8957 .. code-block:: llvm
8959 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields float:r2 = (a * b) + c
8961 Half Precision Floating Point Intrinsics
8962 ----------------------------------------
8964 For most target platforms, half precision floating point is a
8965 storage-only format. This means that it is a dense encoding (in memory)
8966 but does not support computation in the format.
8968 This means that code must first load the half-precision floating point
8969 value as an i16, then convert it to float with
8970 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
8971 then be performed on the float value (including extending to double
8972 etc). To store the value back to memory, it is first converted to float
8973 if needed, then converted to i16 with
8974 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
8977 .. _int_convert_to_fp16:
8979 '``llvm.convert.to.fp16``' Intrinsic
8980 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8987 declare i16 @llvm.convert.to.fp16.f32(float %a)
8988 declare i16 @llvm.convert.to.fp16.f64(double %a)
8993 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
8994 conventional floating point type to half precision floating point format.
8999 The intrinsic function contains single argument - the value to be
9005 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
9006 conventional floating point format to half precision floating point format. The
9007 return value is an ``i16`` which contains the converted number.
9012 .. code-block:: llvm
9014 %res = call i16 @llvm.convert.to.fp16.f32(float %a)
9015 store i16 %res, i16* @x, align 2
9017 .. _int_convert_from_fp16:
9019 '``llvm.convert.from.fp16``' Intrinsic
9020 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9027 declare float @llvm.convert.from.fp16.f32(i16 %a)
9028 declare double @llvm.convert.from.fp16.f64(i16 %a)
9033 The '``llvm.convert.from.fp16``' intrinsic function performs a
9034 conversion from half precision floating point format to single precision
9035 floating point format.
9040 The intrinsic function contains single argument - the value to be
9046 The '``llvm.convert.from.fp16``' intrinsic function performs a
9047 conversion from half single precision floating point format to single
9048 precision floating point format. The input half-float value is
9049 represented by an ``i16`` value.
9054 .. code-block:: llvm
9056 %a = load i16* @x, align 2
9057 %res = call float @llvm.convert.from.fp16(i16 %a)
9062 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
9063 prefix), are described in the `LLVM Source Level
9064 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
9067 Exception Handling Intrinsics
9068 -----------------------------
9070 The LLVM exception handling intrinsics (which all start with
9071 ``llvm.eh.`` prefix), are described in the `LLVM Exception
9072 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
9076 Trampoline Intrinsics
9077 ---------------------
9079 These intrinsics make it possible to excise one parameter, marked with
9080 the :ref:`nest <nest>` attribute, from a function. The result is a
9081 callable function pointer lacking the nest parameter - the caller does
9082 not need to provide a value for it. Instead, the value to use is stored
9083 in advance in a "trampoline", a block of memory usually allocated on the
9084 stack, which also contains code to splice the nest value into the
9085 argument list. This is used to implement the GCC nested function address
9088 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
9089 then the resulting function pointer has signature ``i32 (i32, i32)*``.
9090 It can be created as follows:
9092 .. code-block:: llvm
9094 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
9095 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
9096 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
9097 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
9098 %fp = bitcast i8* %p to i32 (i32, i32)*
9100 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
9101 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
9105 '``llvm.init.trampoline``' Intrinsic
9106 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9113 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
9118 This fills the memory pointed to by ``tramp`` with executable code,
9119 turning it into a trampoline.
9124 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
9125 pointers. The ``tramp`` argument must point to a sufficiently large and
9126 sufficiently aligned block of memory; this memory is written to by the
9127 intrinsic. Note that the size and the alignment are target-specific -
9128 LLVM currently provides no portable way of determining them, so a
9129 front-end that generates this intrinsic needs to have some
9130 target-specific knowledge. The ``func`` argument must hold a function
9131 bitcast to an ``i8*``.
9136 The block of memory pointed to by ``tramp`` is filled with target
9137 dependent code, turning it into a function. Then ``tramp`` needs to be
9138 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
9139 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
9140 function's signature is the same as that of ``func`` with any arguments
9141 marked with the ``nest`` attribute removed. At most one such ``nest``
9142 argument is allowed, and it must be of pointer type. Calling the new
9143 function is equivalent to calling ``func`` with the same argument list,
9144 but with ``nval`` used for the missing ``nest`` argument. If, after
9145 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
9146 modified, then the effect of any later call to the returned function
9147 pointer is undefined.
9151 '``llvm.adjust.trampoline``' Intrinsic
9152 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9159 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
9164 This performs any required machine-specific adjustment to the address of
9165 a trampoline (passed as ``tramp``).
9170 ``tramp`` must point to a block of memory which already has trampoline
9171 code filled in by a previous call to
9172 :ref:`llvm.init.trampoline <int_it>`.
9177 On some architectures the address of the code to be executed needs to be
9178 different than the address where the trampoline is actually stored. This
9179 intrinsic returns the executable address corresponding to ``tramp``
9180 after performing the required machine specific adjustments. The pointer
9181 returned can then be :ref:`bitcast and executed <int_trampoline>`.
9186 This class of intrinsics provides information about the lifetime of
9187 memory objects and ranges where variables are immutable.
9191 '``llvm.lifetime.start``' Intrinsic
9192 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9199 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
9204 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
9210 The first argument is a constant integer representing the size of the
9211 object, or -1 if it is variable sized. The second argument is a pointer
9217 This intrinsic indicates that before this point in the code, the value
9218 of the memory pointed to by ``ptr`` is dead. This means that it is known
9219 to never be used and has an undefined value. A load from the pointer
9220 that precedes this intrinsic can be replaced with ``'undef'``.
9224 '``llvm.lifetime.end``' Intrinsic
9225 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9232 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
9237 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
9243 The first argument is a constant integer representing the size of the
9244 object, or -1 if it is variable sized. The second argument is a pointer
9250 This intrinsic indicates that after this point in the code, the value of
9251 the memory pointed to by ``ptr`` is dead. This means that it is known to
9252 never be used and has an undefined value. Any stores into the memory
9253 object following this intrinsic may be removed as dead.
9255 '``llvm.invariant.start``' Intrinsic
9256 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9263 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
9268 The '``llvm.invariant.start``' intrinsic specifies that the contents of
9269 a memory object will not change.
9274 The first argument is a constant integer representing the size of the
9275 object, or -1 if it is variable sized. The second argument is a pointer
9281 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
9282 the return value, the referenced memory location is constant and
9285 '``llvm.invariant.end``' Intrinsic
9286 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9293 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
9298 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
9299 memory object are mutable.
9304 The first argument is the matching ``llvm.invariant.start`` intrinsic.
9305 The second argument is a constant integer representing the size of the
9306 object, or -1 if it is variable sized and the third argument is a
9307 pointer to the object.
9312 This intrinsic indicates that the memory is mutable again.
9317 This class of intrinsics is designed to be generic and has no specific
9320 '``llvm.var.annotation``' Intrinsic
9321 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9328 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9333 The '``llvm.var.annotation``' intrinsic.
9338 The first argument is a pointer to a value, the second is a pointer to a
9339 global string, the third is a pointer to a global string which is the
9340 source file name, and the last argument is the line number.
9345 This intrinsic allows annotation of local variables with arbitrary
9346 strings. This can be useful for special purpose optimizations that want
9347 to look for these annotations. These have no other defined use; they are
9348 ignored by code generation and optimization.
9350 '``llvm.ptr.annotation.*``' Intrinsic
9351 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9356 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
9357 pointer to an integer of any width. *NOTE* you must specify an address space for
9358 the pointer. The identifier for the default address space is the integer
9363 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9364 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
9365 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
9366 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
9367 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
9372 The '``llvm.ptr.annotation``' intrinsic.
9377 The first argument is a pointer to an integer value of arbitrary bitwidth
9378 (result of some expression), the second is a pointer to a global string, the
9379 third is a pointer to a global string which is the source file name, and the
9380 last argument is the line number. It returns the value of the first argument.
9385 This intrinsic allows annotation of a pointer to an integer with arbitrary
9386 strings. This can be useful for special purpose optimizations that want to look
9387 for these annotations. These have no other defined use; they are ignored by code
9388 generation and optimization.
9390 '``llvm.annotation.*``' Intrinsic
9391 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9396 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
9397 any integer bit width.
9401 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
9402 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
9403 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
9404 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
9405 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
9410 The '``llvm.annotation``' intrinsic.
9415 The first argument is an integer value (result of some expression), the
9416 second is a pointer to a global string, the third is a pointer to a
9417 global string which is the source file name, and the last argument is
9418 the line number. It returns the value of the first argument.
9423 This intrinsic allows annotations to be put on arbitrary expressions
9424 with arbitrary strings. This can be useful for special purpose
9425 optimizations that want to look for these annotations. These have no
9426 other defined use; they are ignored by code generation and optimization.
9428 '``llvm.trap``' Intrinsic
9429 ^^^^^^^^^^^^^^^^^^^^^^^^^
9436 declare void @llvm.trap() noreturn nounwind
9441 The '``llvm.trap``' intrinsic.
9451 This intrinsic is lowered to the target dependent trap instruction. If
9452 the target does not have a trap instruction, this intrinsic will be
9453 lowered to a call of the ``abort()`` function.
9455 '``llvm.debugtrap``' Intrinsic
9456 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9463 declare void @llvm.debugtrap() nounwind
9468 The '``llvm.debugtrap``' intrinsic.
9478 This intrinsic is lowered to code which is intended to cause an
9479 execution trap with the intention of requesting the attention of a
9482 '``llvm.stackprotector``' Intrinsic
9483 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9490 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
9495 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
9496 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
9497 is placed on the stack before local variables.
9502 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
9503 The first argument is the value loaded from the stack guard
9504 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
9505 enough space to hold the value of the guard.
9510 This intrinsic causes the prologue/epilogue inserter to force the position of
9511 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
9512 to ensure that if a local variable on the stack is overwritten, it will destroy
9513 the value of the guard. When the function exits, the guard on the stack is
9514 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
9515 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
9516 calling the ``__stack_chk_fail()`` function.
9518 '``llvm.stackprotectorcheck``' Intrinsic
9519 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9526 declare void @llvm.stackprotectorcheck(i8** <guard>)
9531 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
9532 created stack protector and if they are not equal calls the
9533 ``__stack_chk_fail()`` function.
9538 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
9539 the variable ``@__stack_chk_guard``.
9544 This intrinsic is provided to perform the stack protector check by comparing
9545 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
9546 values do not match call the ``__stack_chk_fail()`` function.
9548 The reason to provide this as an IR level intrinsic instead of implementing it
9549 via other IR operations is that in order to perform this operation at the IR
9550 level without an intrinsic, one would need to create additional basic blocks to
9551 handle the success/failure cases. This makes it difficult to stop the stack
9552 protector check from disrupting sibling tail calls in Codegen. With this
9553 intrinsic, we are able to generate the stack protector basic blocks late in
9554 codegen after the tail call decision has occurred.
9556 '``llvm.objectsize``' Intrinsic
9557 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9564 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
9565 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
9570 The ``llvm.objectsize`` intrinsic is designed to provide information to
9571 the optimizers to determine at compile time whether a) an operation
9572 (like memcpy) will overflow a buffer that corresponds to an object, or
9573 b) that a runtime check for overflow isn't necessary. An object in this
9574 context means an allocation of a specific class, structure, array, or
9580 The ``llvm.objectsize`` intrinsic takes two arguments. The first
9581 argument is a pointer to or into the ``object``. The second argument is
9582 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
9583 or -1 (if false) when the object size is unknown. The second argument
9584 only accepts constants.
9589 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
9590 the size of the object concerned. If the size cannot be determined at
9591 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
9592 on the ``min`` argument).
9594 '``llvm.expect``' Intrinsic
9595 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9600 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
9605 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
9606 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
9607 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
9612 The ``llvm.expect`` intrinsic provides information about expected (the
9613 most probable) value of ``val``, which can be used by optimizers.
9618 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
9619 a value. The second argument is an expected value, this needs to be a
9620 constant value, variables are not allowed.
9625 This intrinsic is lowered to the ``val``.
9627 '``llvm.assume``' Intrinsic
9628 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9635 declare void @llvm.assume(i1 %cond)
9640 The ``llvm.assume`` allows the optimizer to assume that the provided
9641 condition is true. This information can then be used in simplifying other parts
9647 The condition which the optimizer may assume is always true.
9652 The intrinsic allows the optimizer to assume that the provided condition is
9653 always true whenever the control flow reaches the intrinsic call. No code is
9654 generated for this intrinsic, and instructions that contribute only to the
9655 provided condition are not used for code generation. If the condition is
9656 violated during execution, the behavior is undefined.
9658 Please note that optimizer might limit the transformations performed on values
9659 used by the ``llvm.assume`` intrinsic in order to preserve the instructions
9660 only used to form the intrinsic's input argument. This might prove undesirable
9661 if the extra information provided by the ``llvm.assume`` intrinsic does cause
9662 sufficient overall improvement in code quality. For this reason,
9663 ``llvm.assume`` should not be used to document basic mathematical invariants
9664 that the optimizer can otherwise deduce or facts that are of little use to the
9667 '``llvm.donothing``' Intrinsic
9668 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9675 declare void @llvm.donothing() nounwind readnone
9680 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's one of only
9681 two intrinsics (besides ``llvm.experimental.patchpoint``) that can be called
9682 with an invoke instruction.
9692 This intrinsic does nothing, and it's removed by optimizers and ignored
9695 Stack Map Intrinsics
9696 --------------------
9698 LLVM provides experimental intrinsics to support runtime patching
9699 mechanisms commonly desired in dynamic language JITs. These intrinsics
9700 are described in :doc:`StackMaps`.