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 = !{i32 42, null, !"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"] [, comdat [($name)]]
600 [, align <Alignment>]
602 For example, the following defines a global in a numbered address space
603 with an initializer, section, and alignment:
607 @G = addrspace(5) constant float 1.0, section "foo", align 4
609 The following example just declares a global variable
613 @G = external global i32
615 The following example defines a thread-local global with the
616 ``initialexec`` TLS model:
620 @G = thread_local(initialexec) global i32 0, align 4
622 .. _functionstructure:
627 LLVM function definitions consist of the "``define``" keyword, an
628 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
629 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
630 an optional :ref:`calling convention <callingconv>`,
631 an optional ``unnamed_addr`` attribute, a return type, an optional
632 :ref:`parameter attribute <paramattrs>` for the return type, a function
633 name, a (possibly empty) argument list (each with optional :ref:`parameter
634 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
635 an optional section, an optional alignment,
636 an optional :ref:`comdat <langref_comdats>`,
637 an optional :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
638 an optional :ref:`prologue <prologuedata>`, an opening
639 curly brace, a list of basic blocks, and a closing curly brace.
641 LLVM function declarations consist of the "``declare``" keyword, an
642 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
643 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
644 an optional :ref:`calling convention <callingconv>`,
645 an optional ``unnamed_addr`` attribute, a return type, an optional
646 :ref:`parameter attribute <paramattrs>` for the return type, a function
647 name, a possibly empty list of arguments, an optional alignment, an optional
648 :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
649 and an optional :ref:`prologue <prologuedata>`.
651 A function definition contains a list of basic blocks, forming the CFG (Control
652 Flow Graph) for the function. Each basic block may optionally start with a label
653 (giving the basic block a symbol table entry), contains a list of instructions,
654 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
655 function return). If an explicit label is not provided, a block is assigned an
656 implicit numbered label, using the next value from the same counter as used for
657 unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
658 entry block does not have an explicit label, it will be assigned label "%0",
659 then the first unnamed temporary in that block will be "%1", etc.
661 The first basic block in a function is special in two ways: it is
662 immediately executed on entrance to the function, and it is not allowed
663 to have predecessor basic blocks (i.e. there can not be any branches to
664 the entry block of a function). Because the block can have no
665 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
667 LLVM allows an explicit section to be specified for functions. If the
668 target supports it, it will emit functions to the section specified.
669 Additionally, the function can placed in a COMDAT.
671 An explicit alignment may be specified for a function. If not present,
672 or if the alignment is set to zero, the alignment of the function is set
673 by the target to whatever it feels convenient. If an explicit alignment
674 is specified, the function is forced to have at least that much
675 alignment. All alignments must be a power of 2.
677 If the ``unnamed_addr`` attribute is given, the address is know to not
678 be significant and two identical functions can be merged.
682 define [linkage] [visibility] [DLLStorageClass]
684 <ResultType> @<FunctionName> ([argument list])
685 [unnamed_addr] [fn Attrs] [section "name"] [comdat [($name)]]
686 [align N] [gc] [prefix Constant] [prologue Constant] { ... }
688 The argument list is a comma seperated sequence of arguments where each
689 argument is of the following form
693 <type> [parameter Attrs] [name]
701 Aliases, unlike function or variables, don't create any new data. They
702 are just a new symbol and metadata for an existing position.
704 Aliases have a name and an aliasee that is either a global value or a
707 Aliases may have an optional :ref:`linkage type <linkage>`, an optional
708 :ref:`visibility style <visibility>`, an optional :ref:`DLL storage class
709 <dllstorageclass>` and an optional :ref:`tls model <tls_model>`.
713 @<Name> = [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal] [unnamed_addr] alias <AliaseeTy> @<Aliasee>
715 The linkage must be one of ``private``, ``internal``, ``linkonce``, ``weak``,
716 ``linkonce_odr``, ``weak_odr``, ``external``. Note that some system linkers
717 might not correctly handle dropping a weak symbol that is aliased.
719 Alias that are not ``unnamed_addr`` are guaranteed to have the same address as
720 the aliasee expression. ``unnamed_addr`` ones are only guaranteed to point
723 Since aliases are only a second name, some restrictions apply, of which
724 some can only be checked when producing an object file:
726 * The expression defining the aliasee must be computable at assembly
727 time. Since it is just a name, no relocations can be used.
729 * No alias in the expression can be weak as the possibility of the
730 intermediate alias being overridden cannot be represented in an
733 * No global value in the expression can be a declaration, since that
734 would require a relocation, which is not possible.
741 Comdat IR provides access to COFF and ELF object file COMDAT functionality.
743 Comdats have a name which represents the COMDAT key. All global objects that
744 specify this key will only end up in the final object file if the linker chooses
745 that key over some other key. Aliases are placed in the same COMDAT that their
746 aliasee computes to, if any.
748 Comdats have a selection kind to provide input on how the linker should
749 choose between keys in two different object files.
753 $<Name> = comdat SelectionKind
755 The selection kind must be one of the following:
758 The linker may choose any COMDAT key, the choice is arbitrary.
760 The linker may choose any COMDAT key but the sections must contain the
763 The linker will choose the section containing the largest COMDAT key.
765 The linker requires that only section with this COMDAT key exist.
767 The linker may choose any COMDAT key but the sections must contain the
770 Note that the Mach-O platform doesn't support COMDATs and ELF only supports
771 ``any`` as a selection kind.
773 Here is an example of a COMDAT group where a function will only be selected if
774 the COMDAT key's section is the largest:
778 $foo = comdat largest
779 @foo = global i32 2, comdat($foo)
781 define void @bar() comdat($foo) {
785 As a syntactic sugar the ``$name`` can be omitted if the name is the same as
791 @foo = global i32 2, comdat
794 In a COFF object file, this will create a COMDAT section with selection kind
795 ``IMAGE_COMDAT_SELECT_LARGEST`` containing the contents of the ``@foo`` symbol
796 and another COMDAT section with selection kind
797 ``IMAGE_COMDAT_SELECT_ASSOCIATIVE`` which is associated with the first COMDAT
798 section and contains the contents of the ``@bar`` symbol.
800 There are some restrictions on the properties of the global object.
801 It, or an alias to it, must have the same name as the COMDAT group when
803 The contents and size of this object may be used during link-time to determine
804 which COMDAT groups get selected depending on the selection kind.
805 Because the name of the object must match the name of the COMDAT group, the
806 linkage of the global object must not be local; local symbols can get renamed
807 if a collision occurs in the symbol table.
809 The combined use of COMDATS and section attributes may yield surprising results.
816 @g1 = global i32 42, section "sec", comdat($foo)
817 @g2 = global i32 42, section "sec", comdat($bar)
819 From the object file perspective, this requires the creation of two sections
820 with the same name. This is necessary because both globals belong to different
821 COMDAT groups and COMDATs, at the object file level, are represented by
824 Note that certain IR constructs like global variables and functions may create
825 COMDATs in the object file in addition to any which are specified using COMDAT
826 IR. This arises, for example, when a global variable has linkonce_odr linkage.
828 .. _namedmetadatastructure:
833 Named metadata is a collection of metadata. :ref:`Metadata
834 nodes <metadata>` (but not metadata strings) are the only valid
835 operands for a named metadata.
839 ; Some unnamed metadata nodes, which are referenced by the named metadata.
844 !name = !{!0, !1, !2}
851 The return type and each parameter of a function type may have a set of
852 *parameter attributes* associated with them. Parameter attributes are
853 used to communicate additional information about the result or
854 parameters of a function. Parameter attributes are considered to be part
855 of the function, not of the function type, so functions with different
856 parameter attributes can have the same function type.
858 Parameter attributes are simple keywords that follow the type specified.
859 If multiple parameter attributes are needed, they are space separated.
864 declare i32 @printf(i8* noalias nocapture, ...)
865 declare i32 @atoi(i8 zeroext)
866 declare signext i8 @returns_signed_char()
868 Note that any attributes for the function result (``nounwind``,
869 ``readonly``) come immediately after the argument list.
871 Currently, only the following parameter attributes are defined:
874 This indicates to the code generator that the parameter or return
875 value should be zero-extended to the extent required by the target's
876 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
877 the caller (for a parameter) or the callee (for a return value).
879 This indicates to the code generator that the parameter or return
880 value should be sign-extended to the extent required by the target's
881 ABI (which is usually 32-bits) by the caller (for a parameter) or
882 the callee (for a return value).
884 This indicates that this parameter or return value should be treated
885 in a special target-dependent fashion during while emitting code for
886 a function call or return (usually, by putting it in a register as
887 opposed to memory, though some targets use it to distinguish between
888 two different kinds of registers). Use of this attribute is
891 This indicates that the pointer parameter should really be passed by
892 value to the function. The attribute implies that a hidden copy of
893 the pointee is made between the caller and the callee, so the callee
894 is unable to modify the value in the caller. This attribute is only
895 valid on LLVM pointer arguments. It is generally used to pass
896 structs and arrays by value, but is also valid on pointers to
897 scalars. The copy is considered to belong to the caller not the
898 callee (for example, ``readonly`` functions should not write to
899 ``byval`` parameters). This is not a valid attribute for return
902 The byval attribute also supports specifying an alignment with the
903 align attribute. It indicates the alignment of the stack slot to
904 form and the known alignment of the pointer specified to the call
905 site. If the alignment is not specified, then the code generator
906 makes a target-specific assumption.
912 The ``inalloca`` argument attribute allows the caller to take the
913 address of outgoing stack arguments. An ``inalloca`` argument must
914 be a pointer to stack memory produced by an ``alloca`` instruction.
915 The alloca, or argument allocation, must also be tagged with the
916 inalloca keyword. Only the last argument may have the ``inalloca``
917 attribute, and that argument is guaranteed to be passed in memory.
919 An argument allocation may be used by a call at most once because
920 the call may deallocate it. The ``inalloca`` attribute cannot be
921 used in conjunction with other attributes that affect argument
922 storage, like ``inreg``, ``nest``, ``sret``, or ``byval``. The
923 ``inalloca`` attribute also disables LLVM's implicit lowering of
924 large aggregate return values, which means that frontend authors
925 must lower them with ``sret`` pointers.
927 When the call site is reached, the argument allocation must have
928 been the most recent stack allocation that is still live, or the
929 results are undefined. It is possible to allocate additional stack
930 space after an argument allocation and before its call site, but it
931 must be cleared off with :ref:`llvm.stackrestore
934 See :doc:`InAlloca` for more information on how to use this
938 This indicates that the pointer parameter specifies the address of a
939 structure that is the return value of the function in the source
940 program. This pointer must be guaranteed by the caller to be valid:
941 loads and stores to the structure may be assumed by the callee
942 not to trap and to be properly aligned. This may only be applied to
943 the first parameter. This is not a valid attribute for return
947 This indicates that the pointer value may be assumed by the optimizer to
948 have the specified alignment.
950 Note that this attribute has additional semantics when combined with the
956 This indicates that objects accessed via pointer values
957 :ref:`based <pointeraliasing>` on the argument or return value are not also
958 accessed, during the execution of the function, via pointer values not
959 *based* on the argument or return value. The attribute on a return value
960 also has additional semantics described below. The caller shares the
961 responsibility with the callee for ensuring that these requirements are met.
962 For further details, please see the discussion of the NoAlias response in
963 :ref:`alias analysis <Must, May, or No>`.
965 Note that this definition of ``noalias`` is intentionally similar
966 to the definition of ``restrict`` in C99 for function arguments.
968 For function return values, C99's ``restrict`` is not meaningful,
969 while LLVM's ``noalias`` is. Furthermore, the semantics of the ``noalias``
970 attribute on return values are stronger than the semantics of the attribute
971 when used on function arguments. On function return values, the ``noalias``
972 attribute indicates that the function acts like a system memory allocation
973 function, returning a pointer to allocated storage disjoint from the
974 storage for any other object accessible to the caller.
977 This indicates that the callee does not make any copies of the
978 pointer that outlive the callee itself. This is not a valid
979 attribute for return values.
984 This indicates that the pointer parameter can be excised using the
985 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
986 attribute for return values and can only be applied to one parameter.
989 This indicates that the function always returns the argument as its return
990 value. This is an optimization hint to the code generator when generating
991 the caller, allowing tail call optimization and omission of register saves
992 and restores in some cases; it is not checked or enforced when generating
993 the callee. The parameter and the function return type must be valid
994 operands for the :ref:`bitcast instruction <i_bitcast>`. This is not a
995 valid attribute for return values and can only be applied to one parameter.
998 This indicates that the parameter or return pointer is not null. This
999 attribute may only be applied to pointer typed parameters. This is not
1000 checked or enforced by LLVM, the caller must ensure that the pointer
1001 passed in is non-null, or the callee must ensure that the returned pointer
1004 ``dereferenceable(<n>)``
1005 This indicates that the parameter or return pointer is dereferenceable. This
1006 attribute may only be applied to pointer typed parameters. A pointer that
1007 is dereferenceable can be loaded from speculatively without a risk of
1008 trapping. The number of bytes known to be dereferenceable must be provided
1009 in parentheses. It is legal for the number of bytes to be less than the
1010 size of the pointee type. The ``nonnull`` attribute does not imply
1011 dereferenceability (consider a pointer to one element past the end of an
1012 array), however ``dereferenceable(<n>)`` does imply ``nonnull`` in
1013 ``addrspace(0)`` (which is the default address space).
1017 Garbage Collector Names
1018 -----------------------
1020 Each function may specify a garbage collector name, which is simply a
1023 .. code-block:: llvm
1025 define void @f() gc "name" { ... }
1027 The compiler declares the supported values of *name*. Specifying a
1028 collector will cause the compiler to alter its output in order to
1029 support the named garbage collection algorithm.
1036 Prefix data is data associated with a function which the code
1037 generator will emit immediately before the function's entrypoint.
1038 The purpose of this feature is to allow frontends to associate
1039 language-specific runtime metadata with specific functions and make it
1040 available through the function pointer while still allowing the
1041 function pointer to be called.
1043 To access the data for a given function, a program may bitcast the
1044 function pointer to a pointer to the constant's type and dereference
1045 index -1. This implies that the IR symbol points just past the end of
1046 the prefix data. For instance, take the example of a function annotated
1047 with a single ``i32``,
1049 .. code-block:: llvm
1051 define void @f() prefix i32 123 { ... }
1053 The prefix data can be referenced as,
1055 .. code-block:: llvm
1057 %0 = bitcast *void () @f to *i32
1058 %a = getelementptr inbounds *i32 %0, i32 -1
1061 Prefix data is laid out as if it were an initializer for a global variable
1062 of the prefix data's type. The function will be placed such that the
1063 beginning of the prefix data is aligned. This means that if the size
1064 of the prefix data is not a multiple of the alignment size, the
1065 function's entrypoint will not be aligned. If alignment of the
1066 function's entrypoint is desired, padding must be added to the prefix
1069 A function may have prefix data but no body. This has similar semantics
1070 to the ``available_externally`` linkage in that the data may be used by the
1071 optimizers but will not be emitted in the object file.
1078 The ``prologue`` attribute allows arbitrary code (encoded as bytes) to
1079 be inserted prior to the function body. This can be used for enabling
1080 function hot-patching and instrumentation.
1082 To maintain the semantics of ordinary function calls, the prologue data must
1083 have a particular format. Specifically, it must begin with a sequence of
1084 bytes which decode to a sequence of machine instructions, valid for the
1085 module's target, which transfer control to the point immediately succeeding
1086 the prologue data, without performing any other visible action. This allows
1087 the inliner and other passes to reason about the semantics of the function
1088 definition without needing to reason about the prologue data. Obviously this
1089 makes the format of the prologue data highly target dependent.
1091 A trivial example of valid prologue data for the x86 architecture is ``i8 144``,
1092 which encodes the ``nop`` instruction:
1094 .. code-block:: llvm
1096 define void @f() prologue i8 144 { ... }
1098 Generally prologue data can be formed by encoding a relative branch instruction
1099 which skips the metadata, as in this example of valid prologue data for the
1100 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
1102 .. code-block:: llvm
1104 %0 = type <{ i8, i8, i8* }>
1106 define void @f() prologue %0 <{ i8 235, i8 8, i8* @md}> { ... }
1108 A function may have prologue data but no body. This has similar semantics
1109 to the ``available_externally`` linkage in that the data may be used by the
1110 optimizers but will not be emitted in the object file.
1117 Attribute groups are groups of attributes that are referenced by objects within
1118 the IR. They are important for keeping ``.ll`` files readable, because a lot of
1119 functions will use the same set of attributes. In the degenerative case of a
1120 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
1121 group will capture the important command line flags used to build that file.
1123 An attribute group is a module-level object. To use an attribute group, an
1124 object references the attribute group's ID (e.g. ``#37``). An object may refer
1125 to more than one attribute group. In that situation, the attributes from the
1126 different groups are merged.
1128 Here is an example of attribute groups for a function that should always be
1129 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
1131 .. code-block:: llvm
1133 ; Target-independent attributes:
1134 attributes #0 = { alwaysinline alignstack=4 }
1136 ; Target-dependent attributes:
1137 attributes #1 = { "no-sse" }
1139 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
1140 define void @f() #0 #1 { ... }
1147 Function attributes are set to communicate additional information about
1148 a function. Function attributes are considered to be part of the
1149 function, not of the function type, so functions with different function
1150 attributes can have the same function type.
1152 Function attributes are simple keywords that follow the type specified.
1153 If multiple attributes are needed, they are space separated. For
1156 .. code-block:: llvm
1158 define void @f() noinline { ... }
1159 define void @f() alwaysinline { ... }
1160 define void @f() alwaysinline optsize { ... }
1161 define void @f() optsize { ... }
1164 This attribute indicates that, when emitting the prologue and
1165 epilogue, the backend should forcibly align the stack pointer.
1166 Specify the desired alignment, which must be a power of two, in
1169 This attribute indicates that the inliner should attempt to inline
1170 this function into callers whenever possible, ignoring any active
1171 inlining size threshold for this caller.
1173 This indicates that the callee function at a call site should be
1174 recognized as a built-in function, even though the function's declaration
1175 uses the ``nobuiltin`` attribute. This is only valid at call sites for
1176 direct calls to functions that are declared with the ``nobuiltin``
1179 This attribute indicates that this function is rarely called. When
1180 computing edge weights, basic blocks post-dominated by a cold
1181 function call are also considered to be cold; and, thus, given low
1184 This attribute indicates that the source code contained a hint that
1185 inlining this function is desirable (such as the "inline" keyword in
1186 C/C++). It is just a hint; it imposes no requirements on the
1189 This attribute indicates that the function should be added to a
1190 jump-instruction table at code-generation time, and that all address-taken
1191 references to this function should be replaced with a reference to the
1192 appropriate jump-instruction-table function pointer. Note that this creates
1193 a new pointer for the original function, which means that code that depends
1194 on function-pointer identity can break. So, any function annotated with
1195 ``jumptable`` must also be ``unnamed_addr``.
1197 This attribute suggests that optimization passes and code generator
1198 passes make choices that keep the code size of this function as small
1199 as possible and perform optimizations that may sacrifice runtime
1200 performance in order to minimize the size of the generated code.
1202 This attribute disables prologue / epilogue emission for the
1203 function. This can have very system-specific consequences.
1205 This indicates that the callee function at a call site is not recognized as
1206 a built-in function. LLVM will retain the original call and not replace it
1207 with equivalent code based on the semantics of the built-in function, unless
1208 the call site uses the ``builtin`` attribute. This is valid at call sites
1209 and on function declarations and definitions.
1211 This attribute indicates that calls to the function cannot be
1212 duplicated. A call to a ``noduplicate`` function may be moved
1213 within its parent function, but may not be duplicated within
1214 its parent function.
1216 A function containing a ``noduplicate`` call may still
1217 be an inlining candidate, provided that the call is not
1218 duplicated by inlining. That implies that the function has
1219 internal linkage and only has one call site, so the original
1220 call is dead after inlining.
1222 This attributes disables implicit floating point instructions.
1224 This attribute indicates that the inliner should never inline this
1225 function in any situation. This attribute may not be used together
1226 with the ``alwaysinline`` attribute.
1228 This attribute suppresses lazy symbol binding for the function. This
1229 may make calls to the function faster, at the cost of extra program
1230 startup time if the function is not called during program startup.
1232 This attribute indicates that the code generator should not use a
1233 red zone, even if the target-specific ABI normally permits it.
1235 This function attribute indicates that the function never returns
1236 normally. This produces undefined behavior at runtime if the
1237 function ever does dynamically return.
1239 This function attribute indicates that the function never returns
1240 with an unwind or exceptional control flow. If the function does
1241 unwind, its runtime behavior is undefined.
1243 This function attribute indicates that the function is not optimized
1244 by any optimization or code generator passes with the
1245 exception of interprocedural optimization passes.
1246 This attribute cannot be used together with the ``alwaysinline``
1247 attribute; this attribute is also incompatible
1248 with the ``minsize`` attribute and the ``optsize`` attribute.
1250 This attribute requires the ``noinline`` attribute to be specified on
1251 the function as well, so the function is never inlined into any caller.
1252 Only functions with the ``alwaysinline`` attribute are valid
1253 candidates for inlining into the body of this function.
1255 This attribute suggests that optimization passes and code generator
1256 passes make choices that keep the code size of this function low,
1257 and otherwise do optimizations specifically to reduce code size as
1258 long as they do not significantly impact runtime performance.
1260 On a function, this attribute indicates that the function computes its
1261 result (or decides to unwind an exception) based strictly on its arguments,
1262 without dereferencing any pointer arguments or otherwise accessing
1263 any mutable state (e.g. memory, control registers, etc) visible to
1264 caller functions. It does not write through any pointer arguments
1265 (including ``byval`` arguments) and never changes any state visible
1266 to callers. This means that it cannot unwind exceptions by calling
1267 the ``C++`` exception throwing methods.
1269 On an argument, this attribute indicates that the function does not
1270 dereference that pointer argument, even though it may read or write the
1271 memory that the pointer points to if accessed through other pointers.
1273 On a function, this attribute indicates that the function does not write
1274 through any pointer arguments (including ``byval`` arguments) or otherwise
1275 modify any state (e.g. memory, control registers, etc) visible to
1276 caller functions. It may dereference pointer arguments and read
1277 state that may be set in the caller. A readonly function always
1278 returns the same value (or unwinds an exception identically) when
1279 called with the same set of arguments and global state. It cannot
1280 unwind an exception by calling the ``C++`` exception throwing
1283 On an argument, this attribute indicates that the function does not write
1284 through this pointer argument, even though it may write to the memory that
1285 the pointer points to.
1287 This attribute indicates that this function can return twice. The C
1288 ``setjmp`` is an example of such a function. The compiler disables
1289 some optimizations (like tail calls) in the caller of these
1291 ``sanitize_address``
1292 This attribute indicates that AddressSanitizer checks
1293 (dynamic address safety analysis) are enabled for this function.
1295 This attribute indicates that MemorySanitizer checks (dynamic detection
1296 of accesses to uninitialized memory) are enabled for this function.
1298 This attribute indicates that ThreadSanitizer checks
1299 (dynamic thread safety analysis) are enabled for this function.
1301 This attribute indicates that the function should emit a stack
1302 smashing protector. It is in the form of a "canary" --- a random value
1303 placed on the stack before the local variables that's checked upon
1304 return from the function to see if it has been overwritten. A
1305 heuristic is used to determine if a function needs stack protectors
1306 or not. The heuristic used will enable protectors for functions with:
1308 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1309 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1310 - Calls to alloca() with variable sizes or constant sizes greater than
1311 ``ssp-buffer-size``.
1313 Variables that are identified as requiring a protector will be arranged
1314 on the stack such that they are adjacent to the stack protector guard.
1316 If a function that has an ``ssp`` attribute is inlined into a
1317 function that doesn't have an ``ssp`` attribute, then the resulting
1318 function will have an ``ssp`` attribute.
1320 This attribute indicates that the function should *always* emit a
1321 stack smashing protector. This overrides the ``ssp`` function
1324 Variables that are identified as requiring a protector will be arranged
1325 on the stack such that they are adjacent to the stack protector guard.
1326 The specific layout rules are:
1328 #. Large arrays and structures containing large arrays
1329 (``>= ssp-buffer-size``) are closest to the stack protector.
1330 #. Small arrays and structures containing small arrays
1331 (``< ssp-buffer-size``) are 2nd closest to the protector.
1332 #. Variables that have had their address taken are 3rd closest to the
1335 If a function that has an ``sspreq`` attribute is inlined into a
1336 function that doesn't have an ``sspreq`` attribute or which has an
1337 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1338 an ``sspreq`` attribute.
1340 This attribute indicates that the function should emit a stack smashing
1341 protector. This attribute causes a strong heuristic to be used when
1342 determining if a function needs stack protectors. The strong heuristic
1343 will enable protectors for functions with:
1345 - Arrays of any size and type
1346 - Aggregates containing an array of any size and type.
1347 - Calls to alloca().
1348 - Local variables that have had their address taken.
1350 Variables that are identified as requiring a protector will be arranged
1351 on the stack such that they are adjacent to the stack protector guard.
1352 The specific layout rules are:
1354 #. Large arrays and structures containing large arrays
1355 (``>= ssp-buffer-size``) are closest to the stack protector.
1356 #. Small arrays and structures containing small arrays
1357 (``< ssp-buffer-size``) are 2nd closest to the protector.
1358 #. Variables that have had their address taken are 3rd closest to the
1361 This overrides the ``ssp`` function attribute.
1363 If a function that has an ``sspstrong`` attribute is inlined into a
1364 function that doesn't have an ``sspstrong`` attribute, then the
1365 resulting function will have an ``sspstrong`` attribute.
1367 This attribute indicates that the ABI being targeted requires that
1368 an unwind table entry be produce for this function even if we can
1369 show that no exceptions passes by it. This is normally the case for
1370 the ELF x86-64 abi, but it can be disabled for some compilation
1375 Module-Level Inline Assembly
1376 ----------------------------
1378 Modules may contain "module-level inline asm" blocks, which corresponds
1379 to the GCC "file scope inline asm" blocks. These blocks are internally
1380 concatenated by LLVM and treated as a single unit, but may be separated
1381 in the ``.ll`` file if desired. The syntax is very simple:
1383 .. code-block:: llvm
1385 module asm "inline asm code goes here"
1386 module asm "more can go here"
1388 The strings can contain any character by escaping non-printable
1389 characters. The escape sequence used is simply "\\xx" where "xx" is the
1390 two digit hex code for the number.
1392 The inline asm code is simply printed to the machine code .s file when
1393 assembly code is generated.
1395 .. _langref_datalayout:
1400 A module may specify a target specific data layout string that specifies
1401 how data is to be laid out in memory. The syntax for the data layout is
1404 .. code-block:: llvm
1406 target datalayout = "layout specification"
1408 The *layout specification* consists of a list of specifications
1409 separated by the minus sign character ('-'). Each specification starts
1410 with a letter and may include other information after the letter to
1411 define some aspect of the data layout. The specifications accepted are
1415 Specifies that the target lays out data in big-endian form. That is,
1416 the bits with the most significance have the lowest address
1419 Specifies that the target lays out data in little-endian form. That
1420 is, the bits with the least significance have the lowest address
1423 Specifies the natural alignment of the stack in bits. Alignment
1424 promotion of stack variables is limited to the natural stack
1425 alignment to avoid dynamic stack realignment. The stack alignment
1426 must be a multiple of 8-bits. If omitted, the natural stack
1427 alignment defaults to "unspecified", which does not prevent any
1428 alignment promotions.
1429 ``p[n]:<size>:<abi>:<pref>``
1430 This specifies the *size* of a pointer and its ``<abi>`` and
1431 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1432 bits. The address space, ``n`` is optional, and if not specified,
1433 denotes the default address space 0. The value of ``n`` must be
1434 in the range [1,2^23).
1435 ``i<size>:<abi>:<pref>``
1436 This specifies the alignment for an integer type of a given bit
1437 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1438 ``v<size>:<abi>:<pref>``
1439 This specifies the alignment for a vector type of a given bit
1441 ``f<size>:<abi>:<pref>``
1442 This specifies the alignment for a floating point type of a given bit
1443 ``<size>``. Only values of ``<size>`` that are supported by the target
1444 will work. 32 (float) and 64 (double) are supported on all targets; 80
1445 or 128 (different flavors of long double) are also supported on some
1448 This specifies the alignment for an object of aggregate type.
1450 If present, specifies that llvm names are mangled in the output. The
1453 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1454 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1455 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1456 symbols get a ``_`` prefix.
1457 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1458 functions also get a suffix based on the frame size.
1459 ``n<size1>:<size2>:<size3>...``
1460 This specifies a set of native integer widths for the target CPU in
1461 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1462 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1463 this set are considered to support most general arithmetic operations
1466 On every specification that takes a ``<abi>:<pref>``, specifying the
1467 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1468 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1470 When constructing the data layout for a given target, LLVM starts with a
1471 default set of specifications which are then (possibly) overridden by
1472 the specifications in the ``datalayout`` keyword. The default
1473 specifications are given in this list:
1475 - ``E`` - big endian
1476 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1477 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1478 same as the default address space.
1479 - ``S0`` - natural stack alignment is unspecified
1480 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1481 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1482 - ``i16:16:16`` - i16 is 16-bit aligned
1483 - ``i32:32:32`` - i32 is 32-bit aligned
1484 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1485 alignment of 64-bits
1486 - ``f16:16:16`` - half is 16-bit aligned
1487 - ``f32:32:32`` - float is 32-bit aligned
1488 - ``f64:64:64`` - double is 64-bit aligned
1489 - ``f128:128:128`` - quad is 128-bit aligned
1490 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1491 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1492 - ``a:0:64`` - aggregates are 64-bit aligned
1494 When LLVM is determining the alignment for a given type, it uses the
1497 #. If the type sought is an exact match for one of the specifications,
1498 that specification is used.
1499 #. If no match is found, and the type sought is an integer type, then
1500 the smallest integer type that is larger than the bitwidth of the
1501 sought type is used. If none of the specifications are larger than
1502 the bitwidth then the largest integer type is used. For example,
1503 given the default specifications above, the i7 type will use the
1504 alignment of i8 (next largest) while both i65 and i256 will use the
1505 alignment of i64 (largest specified).
1506 #. If no match is found, and the type sought is a vector type, then the
1507 largest vector type that is smaller than the sought vector type will
1508 be used as a fall back. This happens because <128 x double> can be
1509 implemented in terms of 64 <2 x double>, for example.
1511 The function of the data layout string may not be what you expect.
1512 Notably, this is not a specification from the frontend of what alignment
1513 the code generator should use.
1515 Instead, if specified, the target data layout is required to match what
1516 the ultimate *code generator* expects. This string is used by the
1517 mid-level optimizers to improve code, and this only works if it matches
1518 what the ultimate code generator uses. If you would like to generate IR
1519 that does not embed this target-specific detail into the IR, then you
1520 don't have to specify the string. This will disable some optimizations
1521 that require precise layout information, but this also prevents those
1522 optimizations from introducing target specificity into the IR.
1529 A module may specify a target triple string that describes the target
1530 host. The syntax for the target triple is simply:
1532 .. code-block:: llvm
1534 target triple = "x86_64-apple-macosx10.7.0"
1536 The *target triple* string consists of a series of identifiers delimited
1537 by the minus sign character ('-'). The canonical forms are:
1541 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1542 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1544 This information is passed along to the backend so that it generates
1545 code for the proper architecture. It's possible to override this on the
1546 command line with the ``-mtriple`` command line option.
1548 .. _pointeraliasing:
1550 Pointer Aliasing Rules
1551 ----------------------
1553 Any memory access must be done through a pointer value associated with
1554 an address range of the memory access, otherwise the behavior is
1555 undefined. Pointer values are associated with address ranges according
1556 to the following rules:
1558 - A pointer value is associated with the addresses associated with any
1559 value it is *based* on.
1560 - An address of a global variable is associated with the address range
1561 of the variable's storage.
1562 - The result value of an allocation instruction is associated with the
1563 address range of the allocated storage.
1564 - A null pointer in the default address-space is associated with no
1566 - An integer constant other than zero or a pointer value returned from
1567 a function not defined within LLVM may be associated with address
1568 ranges allocated through mechanisms other than those provided by
1569 LLVM. Such ranges shall not overlap with any ranges of addresses
1570 allocated by mechanisms provided by LLVM.
1572 A pointer value is *based* on another pointer value according to the
1575 - A pointer value formed from a ``getelementptr`` operation is *based*
1576 on the first operand of the ``getelementptr``.
1577 - The result value of a ``bitcast`` is *based* on the operand of the
1579 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1580 values that contribute (directly or indirectly) to the computation of
1581 the pointer's value.
1582 - The "*based* on" relationship is transitive.
1584 Note that this definition of *"based"* is intentionally similar to the
1585 definition of *"based"* in C99, though it is slightly weaker.
1587 LLVM IR does not associate types with memory. The result type of a
1588 ``load`` merely indicates the size and alignment of the memory from
1589 which to load, as well as the interpretation of the value. The first
1590 operand type of a ``store`` similarly only indicates the size and
1591 alignment of the store.
1593 Consequently, type-based alias analysis, aka TBAA, aka
1594 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1595 :ref:`Metadata <metadata>` may be used to encode additional information
1596 which specialized optimization passes may use to implement type-based
1601 Volatile Memory Accesses
1602 ------------------------
1604 Certain memory accesses, such as :ref:`load <i_load>`'s,
1605 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1606 marked ``volatile``. The optimizers must not change the number of
1607 volatile operations or change their order of execution relative to other
1608 volatile operations. The optimizers *may* change the order of volatile
1609 operations relative to non-volatile operations. This is not Java's
1610 "volatile" and has no cross-thread synchronization behavior.
1612 IR-level volatile loads and stores cannot safely be optimized into
1613 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1614 flagged volatile. Likewise, the backend should never split or merge
1615 target-legal volatile load/store instructions.
1617 .. admonition:: Rationale
1619 Platforms may rely on volatile loads and stores of natively supported
1620 data width to be executed as single instruction. For example, in C
1621 this holds for an l-value of volatile primitive type with native
1622 hardware support, but not necessarily for aggregate types. The
1623 frontend upholds these expectations, which are intentionally
1624 unspecified in the IR. The rules above ensure that IR transformation
1625 do not violate the frontend's contract with the language.
1629 Memory Model for Concurrent Operations
1630 --------------------------------------
1632 The LLVM IR does not define any way to start parallel threads of
1633 execution or to register signal handlers. Nonetheless, there are
1634 platform-specific ways to create them, and we define LLVM IR's behavior
1635 in their presence. This model is inspired by the C++0x memory model.
1637 For a more informal introduction to this model, see the :doc:`Atomics`.
1639 We define a *happens-before* partial order as the least partial order
1642 - Is a superset of single-thread program order, and
1643 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1644 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1645 techniques, like pthread locks, thread creation, thread joining,
1646 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1647 Constraints <ordering>`).
1649 Note that program order does not introduce *happens-before* edges
1650 between a thread and signals executing inside that thread.
1652 Every (defined) read operation (load instructions, memcpy, atomic
1653 loads/read-modify-writes, etc.) R reads a series of bytes written by
1654 (defined) write operations (store instructions, atomic
1655 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1656 section, initialized globals are considered to have a write of the
1657 initializer which is atomic and happens before any other read or write
1658 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1659 may see any write to the same byte, except:
1661 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1662 write\ :sub:`2` happens before R\ :sub:`byte`, then
1663 R\ :sub:`byte` does not see write\ :sub:`1`.
1664 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1665 R\ :sub:`byte` does not see write\ :sub:`3`.
1667 Given that definition, R\ :sub:`byte` is defined as follows:
1669 - If R is volatile, the result is target-dependent. (Volatile is
1670 supposed to give guarantees which can support ``sig_atomic_t`` in
1671 C/C++, and may be used for accesses to addresses that do not behave
1672 like normal memory. It does not generally provide cross-thread
1674 - Otherwise, if there is no write to the same byte that happens before
1675 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1676 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1677 R\ :sub:`byte` returns the value written by that write.
1678 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1679 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1680 Memory Ordering Constraints <ordering>` section for additional
1681 constraints on how the choice is made.
1682 - Otherwise R\ :sub:`byte` returns ``undef``.
1684 R returns the value composed of the series of bytes it read. This
1685 implies that some bytes within the value may be ``undef`` **without**
1686 the entire value being ``undef``. Note that this only defines the
1687 semantics of the operation; it doesn't mean that targets will emit more
1688 than one instruction to read the series of bytes.
1690 Note that in cases where none of the atomic intrinsics are used, this
1691 model places only one restriction on IR transformations on top of what
1692 is required for single-threaded execution: introducing a store to a byte
1693 which might not otherwise be stored is not allowed in general.
1694 (Specifically, in the case where another thread might write to and read
1695 from an address, introducing a store can change a load that may see
1696 exactly one write into a load that may see multiple writes.)
1700 Atomic Memory Ordering Constraints
1701 ----------------------------------
1703 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1704 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1705 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1706 ordering parameters that determine which other atomic instructions on
1707 the same address they *synchronize with*. These semantics are borrowed
1708 from Java and C++0x, but are somewhat more colloquial. If these
1709 descriptions aren't precise enough, check those specs (see spec
1710 references in the :doc:`atomics guide <Atomics>`).
1711 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1712 differently since they don't take an address. See that instruction's
1713 documentation for details.
1715 For a simpler introduction to the ordering constraints, see the
1719 The set of values that can be read is governed by the happens-before
1720 partial order. A value cannot be read unless some operation wrote
1721 it. This is intended to provide a guarantee strong enough to model
1722 Java's non-volatile shared variables. This ordering cannot be
1723 specified for read-modify-write operations; it is not strong enough
1724 to make them atomic in any interesting way.
1726 In addition to the guarantees of ``unordered``, there is a single
1727 total order for modifications by ``monotonic`` operations on each
1728 address. All modification orders must be compatible with the
1729 happens-before order. There is no guarantee that the modification
1730 orders can be combined to a global total order for the whole program
1731 (and this often will not be possible). The read in an atomic
1732 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1733 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1734 order immediately before the value it writes. If one atomic read
1735 happens before another atomic read of the same address, the later
1736 read must see the same value or a later value in the address's
1737 modification order. This disallows reordering of ``monotonic`` (or
1738 stronger) operations on the same address. If an address is written
1739 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1740 read that address repeatedly, the other threads must eventually see
1741 the write. This corresponds to the C++0x/C1x
1742 ``memory_order_relaxed``.
1744 In addition to the guarantees of ``monotonic``, a
1745 *synchronizes-with* edge may be formed with a ``release`` operation.
1746 This is intended to model C++'s ``memory_order_acquire``.
1748 In addition to the guarantees of ``monotonic``, if this operation
1749 writes a value which is subsequently read by an ``acquire``
1750 operation, it *synchronizes-with* that operation. (This isn't a
1751 complete description; see the C++0x definition of a release
1752 sequence.) This corresponds to the C++0x/C1x
1753 ``memory_order_release``.
1754 ``acq_rel`` (acquire+release)
1755 Acts as both an ``acquire`` and ``release`` operation on its
1756 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1757 ``seq_cst`` (sequentially consistent)
1758 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1759 operation that only reads, ``release`` for an operation that only
1760 writes), there is a global total order on all
1761 sequentially-consistent operations on all addresses, which is
1762 consistent with the *happens-before* partial order and with the
1763 modification orders of all the affected addresses. Each
1764 sequentially-consistent read sees the last preceding write to the
1765 same address in this global order. This corresponds to the C++0x/C1x
1766 ``memory_order_seq_cst`` and Java volatile.
1770 If an atomic operation is marked ``singlethread``, it only *synchronizes
1771 with* or participates in modification and seq\_cst total orderings with
1772 other operations running in the same thread (for example, in signal
1780 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1781 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1782 :ref:`frem <i_frem>`) have the following flags that can set to enable
1783 otherwise unsafe floating point operations
1786 No NaNs - Allow optimizations to assume the arguments and result are not
1787 NaN. Such optimizations are required to retain defined behavior over
1788 NaNs, but the value of the result is undefined.
1791 No Infs - Allow optimizations to assume the arguments and result are not
1792 +/-Inf. Such optimizations are required to retain defined behavior over
1793 +/-Inf, but the value of the result is undefined.
1796 No Signed Zeros - Allow optimizations to treat the sign of a zero
1797 argument or result as insignificant.
1800 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1801 argument rather than perform division.
1804 Fast - Allow algebraically equivalent transformations that may
1805 dramatically change results in floating point (e.g. reassociate). This
1806 flag implies all the others.
1810 Use-list Order Directives
1811 -------------------------
1813 Use-list directives encode the in-memory order of each use-list, allowing the
1814 order to be recreated. ``<order-indexes>`` is a comma-separated list of
1815 indexes that are assigned to the referenced value's uses. The referenced
1816 value's use-list is immediately sorted by these indexes.
1818 Use-list directives may appear at function scope or global scope. They are not
1819 instructions, and have no effect on the semantics of the IR. When they're at
1820 function scope, they must appear after the terminator of the final basic block.
1822 If basic blocks have their address taken via ``blockaddress()`` expressions,
1823 ``uselistorder_bb`` can be used to reorder their use-lists from outside their
1830 uselistorder <ty> <value>, { <order-indexes> }
1831 uselistorder_bb @function, %block { <order-indexes> }
1837 define void @foo(i32 %arg1, i32 %arg2) {
1839 ; ... instructions ...
1841 ; ... instructions ...
1843 ; At function scope.
1844 uselistorder i32 %arg1, { 1, 0, 2 }
1845 uselistorder label %bb, { 1, 0 }
1849 uselistorder i32* @global, { 1, 2, 0 }
1850 uselistorder i32 7, { 1, 0 }
1851 uselistorder i32 (i32) @bar, { 1, 0 }
1852 uselistorder_bb @foo, %bb, { 5, 1, 3, 2, 0, 4 }
1859 The LLVM type system is one of the most important features of the
1860 intermediate representation. Being typed enables a number of
1861 optimizations to be performed on the intermediate representation
1862 directly, without having to do extra analyses on the side before the
1863 transformation. A strong type system makes it easier to read the
1864 generated code and enables novel analyses and transformations that are
1865 not feasible to perform on normal three address code representations.
1875 The void type does not represent any value and has no size.
1893 The function type can be thought of as a function signature. It consists of a
1894 return type and a list of formal parameter types. The return type of a function
1895 type is a void type or first class type --- except for :ref:`label <t_label>`
1896 and :ref:`metadata <t_metadata>` types.
1902 <returntype> (<parameter list>)
1904 ...where '``<parameter list>``' is a comma-separated list of type
1905 specifiers. Optionally, the parameter list may include a type ``...``, which
1906 indicates that the function takes a variable number of arguments. Variable
1907 argument functions can access their arguments with the :ref:`variable argument
1908 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
1909 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
1913 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1914 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1915 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1916 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1917 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1918 | ``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. |
1919 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1920 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1921 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1928 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1929 Values of these types are the only ones which can be produced by
1937 These are the types that are valid in registers from CodeGen's perspective.
1946 The integer type is a very simple type that simply specifies an
1947 arbitrary bit width for the integer type desired. Any bit width from 1
1948 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1956 The number of bits the integer will occupy is specified by the ``N``
1962 +----------------+------------------------------------------------+
1963 | ``i1`` | a single-bit integer. |
1964 +----------------+------------------------------------------------+
1965 | ``i32`` | a 32-bit integer. |
1966 +----------------+------------------------------------------------+
1967 | ``i1942652`` | a really big integer of over 1 million bits. |
1968 +----------------+------------------------------------------------+
1972 Floating Point Types
1973 """"""""""""""""""""
1982 - 16-bit floating point value
1985 - 32-bit floating point value
1988 - 64-bit floating point value
1991 - 128-bit floating point value (112-bit mantissa)
1994 - 80-bit floating point value (X87)
1997 - 128-bit floating point value (two 64-bits)
2004 The x86_mmx type represents a value held in an MMX register on an x86
2005 machine. The operations allowed on it are quite limited: parameters and
2006 return values, load and store, and bitcast. User-specified MMX
2007 instructions are represented as intrinsic or asm calls with arguments
2008 and/or results of this type. There are no arrays, vectors or constants
2025 The pointer type is used to specify memory locations. Pointers are
2026 commonly used to reference objects in memory.
2028 Pointer types may have an optional address space attribute defining the
2029 numbered address space where the pointed-to object resides. The default
2030 address space is number zero. The semantics of non-zero address spaces
2031 are target-specific.
2033 Note that LLVM does not permit pointers to void (``void*``) nor does it
2034 permit pointers to labels (``label*``). Use ``i8*`` instead.
2044 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2045 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
2046 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2047 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
2048 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2049 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
2050 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2059 A vector type is a simple derived type that represents a vector of
2060 elements. Vector types are used when multiple primitive data are
2061 operated in parallel using a single instruction (SIMD). A vector type
2062 requires a size (number of elements) and an underlying primitive data
2063 type. Vector types are considered :ref:`first class <t_firstclass>`.
2069 < <# elements> x <elementtype> >
2071 The number of elements is a constant integer value larger than 0;
2072 elementtype may be any integer, floating point or pointer type. Vectors
2073 of size zero are not allowed.
2077 +-------------------+--------------------------------------------------+
2078 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
2079 +-------------------+--------------------------------------------------+
2080 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
2081 +-------------------+--------------------------------------------------+
2082 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
2083 +-------------------+--------------------------------------------------+
2084 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
2085 +-------------------+--------------------------------------------------+
2094 The label type represents code labels.
2109 The metadata type represents embedded metadata. No derived types may be
2110 created from metadata except for :ref:`function <t_function>` arguments.
2123 Aggregate Types are a subset of derived types that can contain multiple
2124 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
2125 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
2135 The array type is a very simple derived type that arranges elements
2136 sequentially in memory. The array type requires a size (number of
2137 elements) and an underlying data type.
2143 [<# elements> x <elementtype>]
2145 The number of elements is a constant integer value; ``elementtype`` may
2146 be any type with a size.
2150 +------------------+--------------------------------------+
2151 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
2152 +------------------+--------------------------------------+
2153 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
2154 +------------------+--------------------------------------+
2155 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
2156 +------------------+--------------------------------------+
2158 Here are some examples of multidimensional arrays:
2160 +-----------------------------+----------------------------------------------------------+
2161 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
2162 +-----------------------------+----------------------------------------------------------+
2163 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
2164 +-----------------------------+----------------------------------------------------------+
2165 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
2166 +-----------------------------+----------------------------------------------------------+
2168 There is no restriction on indexing beyond the end of the array implied
2169 by a static type (though there are restrictions on indexing beyond the
2170 bounds of an allocated object in some cases). This means that
2171 single-dimension 'variable sized array' addressing can be implemented in
2172 LLVM with a zero length array type. An implementation of 'pascal style
2173 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
2183 The structure type is used to represent a collection of data members
2184 together in memory. The elements of a structure may be any type that has
2187 Structures in memory are accessed using '``load``' and '``store``' by
2188 getting a pointer to a field with the '``getelementptr``' instruction.
2189 Structures in registers are accessed using the '``extractvalue``' and
2190 '``insertvalue``' instructions.
2192 Structures may optionally be "packed" structures, which indicate that
2193 the alignment of the struct is one byte, and that there is no padding
2194 between the elements. In non-packed structs, padding between field types
2195 is inserted as defined by the DataLayout string in the module, which is
2196 required to match what the underlying code generator expects.
2198 Structures can either be "literal" or "identified". A literal structure
2199 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
2200 identified types are always defined at the top level with a name.
2201 Literal types are uniqued by their contents and can never be recursive
2202 or opaque since there is no way to write one. Identified types can be
2203 recursive, can be opaqued, and are never uniqued.
2209 %T1 = type { <type list> } ; Identified normal struct type
2210 %T2 = type <{ <type list> }> ; Identified packed struct type
2214 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2215 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
2216 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2217 | ``{ 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``. |
2218 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2219 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
2220 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2224 Opaque Structure Types
2225 """"""""""""""""""""""
2229 Opaque structure types are used to represent named structure types that
2230 do not have a body specified. This corresponds (for example) to the C
2231 notion of a forward declared structure.
2242 +--------------+-------------------+
2243 | ``opaque`` | An opaque type. |
2244 +--------------+-------------------+
2251 LLVM has several different basic types of constants. This section
2252 describes them all and their syntax.
2257 **Boolean constants**
2258 The two strings '``true``' and '``false``' are both valid constants
2260 **Integer constants**
2261 Standard integers (such as '4') are constants of the
2262 :ref:`integer <t_integer>` type. Negative numbers may be used with
2264 **Floating point constants**
2265 Floating point constants use standard decimal notation (e.g.
2266 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2267 hexadecimal notation (see below). The assembler requires the exact
2268 decimal value of a floating-point constant. For example, the
2269 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2270 decimal in binary. Floating point constants must have a :ref:`floating
2271 point <t_floating>` type.
2272 **Null pointer constants**
2273 The identifier '``null``' is recognized as a null pointer constant
2274 and must be of :ref:`pointer type <t_pointer>`.
2276 The one non-intuitive notation for constants is the hexadecimal form of
2277 floating point constants. For example, the form
2278 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2279 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2280 constants are required (and the only time that they are generated by the
2281 disassembler) is when a floating point constant must be emitted but it
2282 cannot be represented as a decimal floating point number in a reasonable
2283 number of digits. For example, NaN's, infinities, and other special
2284 values are represented in their IEEE hexadecimal format so that assembly
2285 and disassembly do not cause any bits to change in the constants.
2287 When using the hexadecimal form, constants of types half, float, and
2288 double are represented using the 16-digit form shown above (which
2289 matches the IEEE754 representation for double); half and float values
2290 must, however, be exactly representable as IEEE 754 half and single
2291 precision, respectively. Hexadecimal format is always used for long
2292 double, and there are three forms of long double. The 80-bit format used
2293 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2294 128-bit format used by PowerPC (two adjacent doubles) is represented by
2295 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2296 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2297 will only work if they match the long double format on your target.
2298 The IEEE 16-bit format (half precision) is represented by ``0xH``
2299 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2300 (sign bit at the left).
2302 There are no constants of type x86_mmx.
2304 .. _complexconstants:
2309 Complex constants are a (potentially recursive) combination of simple
2310 constants and smaller complex constants.
2312 **Structure constants**
2313 Structure constants are represented with notation similar to
2314 structure type definitions (a comma separated list of elements,
2315 surrounded by braces (``{}``)). For example:
2316 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2317 "``@G = external global i32``". Structure constants must have
2318 :ref:`structure type <t_struct>`, and the number and types of elements
2319 must match those specified by the type.
2321 Array constants are represented with notation similar to array type
2322 definitions (a comma separated list of elements, surrounded by
2323 square brackets (``[]``)). For example:
2324 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2325 :ref:`array type <t_array>`, and the number and types of elements must
2326 match those specified by the type. As a special case, character array
2327 constants may also be represented as a double-quoted string using the ``c``
2328 prefix. For example: "``c"Hello World\0A\00"``".
2329 **Vector constants**
2330 Vector constants are represented with notation similar to vector
2331 type definitions (a comma separated list of elements, surrounded by
2332 less-than/greater-than's (``<>``)). For example:
2333 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2334 must have :ref:`vector type <t_vector>`, and the number and types of
2335 elements must match those specified by the type.
2336 **Zero initialization**
2337 The string '``zeroinitializer``' can be used to zero initialize a
2338 value to zero of *any* type, including scalar and
2339 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2340 having to print large zero initializers (e.g. for large arrays) and
2341 is always exactly equivalent to using explicit zero initializers.
2343 A metadata node is a constant tuple without types. For example:
2344 "``!{!0, !{!2, !0}, !"test"}``". Metadata can reference constant values,
2345 for example: "``!{!0, i32 0, i8* @global, i64 (i64)* @function, !"str"}``".
2346 Unlike other typed constants that are meant to be interpreted as part of
2347 the instruction stream, metadata is a place to attach additional
2348 information such as debug info.
2350 Global Variable and Function Addresses
2351 --------------------------------------
2353 The addresses of :ref:`global variables <globalvars>` and
2354 :ref:`functions <functionstructure>` are always implicitly valid
2355 (link-time) constants. These constants are explicitly referenced when
2356 the :ref:`identifier for the global <identifiers>` is used and always have
2357 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2360 .. code-block:: llvm
2364 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2371 The string '``undef``' can be used anywhere a constant is expected, and
2372 indicates that the user of the value may receive an unspecified
2373 bit-pattern. Undefined values may be of any type (other than '``label``'
2374 or '``void``') and be used anywhere a constant is permitted.
2376 Undefined values are useful because they indicate to the compiler that
2377 the program is well defined no matter what value is used. This gives the
2378 compiler more freedom to optimize. Here are some examples of
2379 (potentially surprising) transformations that are valid (in pseudo IR):
2381 .. code-block:: llvm
2391 This is safe because all of the output bits are affected by the undef
2392 bits. Any output bit can have a zero or one depending on the input bits.
2394 .. code-block:: llvm
2405 These logical operations have bits that are not always affected by the
2406 input. For example, if ``%X`` has a zero bit, then the output of the
2407 '``and``' operation will always be a zero for that bit, no matter what
2408 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2409 optimize or assume that the result of the '``and``' is '``undef``'.
2410 However, it is safe to assume that all bits of the '``undef``' could be
2411 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2412 all the bits of the '``undef``' operand to the '``or``' could be set,
2413 allowing the '``or``' to be folded to -1.
2415 .. code-block:: llvm
2417 %A = select undef, %X, %Y
2418 %B = select undef, 42, %Y
2419 %C = select %X, %Y, undef
2429 This set of examples shows that undefined '``select``' (and conditional
2430 branch) conditions can go *either way*, but they have to come from one
2431 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2432 both known to have a clear low bit, then ``%A`` would have to have a
2433 cleared low bit. However, in the ``%C`` example, the optimizer is
2434 allowed to assume that the '``undef``' operand could be the same as
2435 ``%Y``, allowing the whole '``select``' to be eliminated.
2437 .. code-block:: llvm
2439 %A = xor undef, undef
2456 This example points out that two '``undef``' operands are not
2457 necessarily the same. This can be surprising to people (and also matches
2458 C semantics) where they assume that "``X^X``" is always zero, even if
2459 ``X`` is undefined. This isn't true for a number of reasons, but the
2460 short answer is that an '``undef``' "variable" can arbitrarily change
2461 its value over its "live range". This is true because the variable
2462 doesn't actually *have a live range*. Instead, the value is logically
2463 read from arbitrary registers that happen to be around when needed, so
2464 the value is not necessarily consistent over time. In fact, ``%A`` and
2465 ``%C`` need to have the same semantics or the core LLVM "replace all
2466 uses with" concept would not hold.
2468 .. code-block:: llvm
2476 These examples show the crucial difference between an *undefined value*
2477 and *undefined behavior*. An undefined value (like '``undef``') is
2478 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2479 operation can be constant folded to '``undef``', because the '``undef``'
2480 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2481 However, in the second example, we can make a more aggressive
2482 assumption: because the ``undef`` is allowed to be an arbitrary value,
2483 we are allowed to assume that it could be zero. Since a divide by zero
2484 has *undefined behavior*, we are allowed to assume that the operation
2485 does not execute at all. This allows us to delete the divide and all
2486 code after it. Because the undefined operation "can't happen", the
2487 optimizer can assume that it occurs in dead code.
2489 .. code-block:: llvm
2491 a: store undef -> %X
2492 b: store %X -> undef
2497 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2498 value can be assumed to not have any effect; we can assume that the
2499 value is overwritten with bits that happen to match what was already
2500 there. However, a store *to* an undefined location could clobber
2501 arbitrary memory, therefore, it has undefined behavior.
2508 Poison values are similar to :ref:`undef values <undefvalues>`, however
2509 they also represent the fact that an instruction or constant expression
2510 that cannot evoke side effects has nevertheless detected a condition
2511 that results in undefined behavior.
2513 There is currently no way of representing a poison value in the IR; they
2514 only exist when produced by operations such as :ref:`add <i_add>` with
2517 Poison value behavior is defined in terms of value *dependence*:
2519 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2520 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2521 their dynamic predecessor basic block.
2522 - Function arguments depend on the corresponding actual argument values
2523 in the dynamic callers of their functions.
2524 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2525 instructions that dynamically transfer control back to them.
2526 - :ref:`Invoke <i_invoke>` instructions depend on the
2527 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2528 call instructions that dynamically transfer control back to them.
2529 - Non-volatile loads and stores depend on the most recent stores to all
2530 of the referenced memory addresses, following the order in the IR
2531 (including loads and stores implied by intrinsics such as
2532 :ref:`@llvm.memcpy <int_memcpy>`.)
2533 - An instruction with externally visible side effects depends on the
2534 most recent preceding instruction with externally visible side
2535 effects, following the order in the IR. (This includes :ref:`volatile
2536 operations <volatile>`.)
2537 - An instruction *control-depends* on a :ref:`terminator
2538 instruction <terminators>` if the terminator instruction has
2539 multiple successors and the instruction is always executed when
2540 control transfers to one of the successors, and may not be executed
2541 when control is transferred to another.
2542 - Additionally, an instruction also *control-depends* on a terminator
2543 instruction if the set of instructions it otherwise depends on would
2544 be different if the terminator had transferred control to a different
2546 - Dependence is transitive.
2548 Poison values have the same behavior as :ref:`undef values <undefvalues>`,
2549 with the additional effect that any instruction that has a *dependence*
2550 on a poison value has undefined behavior.
2552 Here are some examples:
2554 .. code-block:: llvm
2557 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2558 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2559 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
2560 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2562 store i32 %poison, i32* @g ; Poison value stored to memory.
2563 %poison2 = load i32* @g ; Poison value loaded back from memory.
2565 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2567 %narrowaddr = bitcast i32* @g to i16*
2568 %wideaddr = bitcast i32* @g to i64*
2569 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2570 %poison4 = load i64* %wideaddr ; Returns a poison value.
2572 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2573 br i1 %cmp, label %true, label %end ; Branch to either destination.
2576 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2577 ; it has undefined behavior.
2581 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2582 ; Both edges into this PHI are
2583 ; control-dependent on %cmp, so this
2584 ; always results in a poison value.
2586 store volatile i32 0, i32* @g ; This would depend on the store in %true
2587 ; if %cmp is true, or the store in %entry
2588 ; otherwise, so this is undefined behavior.
2590 br i1 %cmp, label %second_true, label %second_end
2591 ; The same branch again, but this time the
2592 ; true block doesn't have side effects.
2599 store volatile i32 0, i32* @g ; This time, the instruction always depends
2600 ; on the store in %end. Also, it is
2601 ; control-equivalent to %end, so this is
2602 ; well-defined (ignoring earlier undefined
2603 ; behavior in this example).
2607 Addresses of Basic Blocks
2608 -------------------------
2610 ``blockaddress(@function, %block)``
2612 The '``blockaddress``' constant computes the address of the specified
2613 basic block in the specified function, and always has an ``i8*`` type.
2614 Taking the address of the entry block is illegal.
2616 This value only has defined behavior when used as an operand to the
2617 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2618 against null. Pointer equality tests between labels addresses results in
2619 undefined behavior --- though, again, comparison against null is ok, and
2620 no label is equal to the null pointer. This may be passed around as an
2621 opaque pointer sized value as long as the bits are not inspected. This
2622 allows ``ptrtoint`` and arithmetic to be performed on these values so
2623 long as the original value is reconstituted before the ``indirectbr``
2626 Finally, some targets may provide defined semantics when using the value
2627 as the operand to an inline assembly, but that is target specific.
2631 Constant Expressions
2632 --------------------
2634 Constant expressions are used to allow expressions involving other
2635 constants to be used as constants. Constant expressions may be of any
2636 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2637 that does not have side effects (e.g. load and call are not supported).
2638 The following is the syntax for constant expressions:
2640 ``trunc (CST to TYPE)``
2641 Truncate a constant to another type. The bit size of CST must be
2642 larger than the bit size of TYPE. Both types must be integers.
2643 ``zext (CST to TYPE)``
2644 Zero extend a constant to another type. The bit size of CST must be
2645 smaller than the bit size of TYPE. Both types must be integers.
2646 ``sext (CST to TYPE)``
2647 Sign extend a constant to another type. The bit size of CST must be
2648 smaller than the bit size of TYPE. Both types must be integers.
2649 ``fptrunc (CST to TYPE)``
2650 Truncate a floating point constant to another floating point type.
2651 The size of CST must be larger than the size of TYPE. Both types
2652 must be floating point.
2653 ``fpext (CST to TYPE)``
2654 Floating point extend a constant to another type. The size of CST
2655 must be smaller or equal to the size of TYPE. Both types must be
2657 ``fptoui (CST to TYPE)``
2658 Convert a floating point constant to the corresponding unsigned
2659 integer constant. TYPE must be a scalar or vector integer type. CST
2660 must be of scalar or vector floating point type. Both CST and TYPE
2661 must be scalars, or vectors of the same number of elements. If the
2662 value won't fit in the integer type, the results are undefined.
2663 ``fptosi (CST to TYPE)``
2664 Convert a floating point constant to the corresponding signed
2665 integer constant. TYPE must be a scalar or vector integer type. CST
2666 must be of scalar or vector floating point type. Both CST and TYPE
2667 must be scalars, or vectors of the same number of elements. If the
2668 value won't fit in the integer type, the results are undefined.
2669 ``uitofp (CST to TYPE)``
2670 Convert an unsigned integer constant to the corresponding floating
2671 point constant. TYPE must be a scalar or vector floating point type.
2672 CST must be of scalar or vector integer type. Both CST and TYPE must
2673 be scalars, or vectors of the same number of elements. If the value
2674 won't fit in the floating point type, the results are undefined.
2675 ``sitofp (CST to TYPE)``
2676 Convert a signed integer constant to the corresponding floating
2677 point constant. TYPE must be a scalar or vector floating point type.
2678 CST must be of scalar or vector integer type. Both CST and TYPE must
2679 be scalars, or vectors of the same number of elements. If the value
2680 won't fit in the floating point type, the results are undefined.
2681 ``ptrtoint (CST to TYPE)``
2682 Convert a pointer typed constant to the corresponding integer
2683 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2684 pointer type. The ``CST`` value is zero extended, truncated, or
2685 unchanged to make it fit in ``TYPE``.
2686 ``inttoptr (CST to TYPE)``
2687 Convert an integer constant to a pointer constant. TYPE must be a
2688 pointer type. CST must be of integer type. The CST value is zero
2689 extended, truncated, or unchanged to make it fit in a pointer size.
2690 This one is *really* dangerous!
2691 ``bitcast (CST to TYPE)``
2692 Convert a constant, CST, to another TYPE. The constraints of the
2693 operands are the same as those for the :ref:`bitcast
2694 instruction <i_bitcast>`.
2695 ``addrspacecast (CST to TYPE)``
2696 Convert a constant pointer or constant vector of pointer, CST, to another
2697 TYPE in a different address space. The constraints of the operands are the
2698 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2699 ``getelementptr (CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)``
2700 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2701 constants. As with the :ref:`getelementptr <i_getelementptr>`
2702 instruction, the index list may have zero or more indexes, which are
2703 required to make sense for the type of "CSTPTR".
2704 ``select (COND, VAL1, VAL2)``
2705 Perform the :ref:`select operation <i_select>` on constants.
2706 ``icmp COND (VAL1, VAL2)``
2707 Performs the :ref:`icmp operation <i_icmp>` on constants.
2708 ``fcmp COND (VAL1, VAL2)``
2709 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2710 ``extractelement (VAL, IDX)``
2711 Perform the :ref:`extractelement operation <i_extractelement>` on
2713 ``insertelement (VAL, ELT, IDX)``
2714 Perform the :ref:`insertelement operation <i_insertelement>` on
2716 ``shufflevector (VEC1, VEC2, IDXMASK)``
2717 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2719 ``extractvalue (VAL, IDX0, IDX1, ...)``
2720 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2721 constants. The index list is interpreted in a similar manner as
2722 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2723 least one index value must be specified.
2724 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2725 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2726 The index list is interpreted in a similar manner as indices in a
2727 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2728 value must be specified.
2729 ``OPCODE (LHS, RHS)``
2730 Perform the specified operation of the LHS and RHS constants. OPCODE
2731 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2732 binary <bitwiseops>` operations. The constraints on operands are
2733 the same as those for the corresponding instruction (e.g. no bitwise
2734 operations on floating point values are allowed).
2741 Inline Assembler Expressions
2742 ----------------------------
2744 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2745 Inline Assembly <moduleasm>`) through the use of a special value. This
2746 value represents the inline assembler as a string (containing the
2747 instructions to emit), a list of operand constraints (stored as a
2748 string), a flag that indicates whether or not the inline asm expression
2749 has side effects, and a flag indicating whether the function containing
2750 the asm needs to align its stack conservatively. An example inline
2751 assembler expression is:
2753 .. code-block:: llvm
2755 i32 (i32) asm "bswap $0", "=r,r"
2757 Inline assembler expressions may **only** be used as the callee operand
2758 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2759 Thus, typically we have:
2761 .. code-block:: llvm
2763 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2765 Inline asms with side effects not visible in the constraint list must be
2766 marked as having side effects. This is done through the use of the
2767 '``sideeffect``' keyword, like so:
2769 .. code-block:: llvm
2771 call void asm sideeffect "eieio", ""()
2773 In some cases inline asms will contain code that will not work unless
2774 the stack is aligned in some way, such as calls or SSE instructions on
2775 x86, yet will not contain code that does that alignment within the asm.
2776 The compiler should make conservative assumptions about what the asm
2777 might contain and should generate its usual stack alignment code in the
2778 prologue if the '``alignstack``' keyword is present:
2780 .. code-block:: llvm
2782 call void asm alignstack "eieio", ""()
2784 Inline asms also support using non-standard assembly dialects. The
2785 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2786 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2787 the only supported dialects. An example is:
2789 .. code-block:: llvm
2791 call void asm inteldialect "eieio", ""()
2793 If multiple keywords appear the '``sideeffect``' keyword must come
2794 first, the '``alignstack``' keyword second and the '``inteldialect``'
2800 The call instructions that wrap inline asm nodes may have a
2801 "``!srcloc``" MDNode attached to it that contains a list of constant
2802 integers. If present, the code generator will use the integer as the
2803 location cookie value when report errors through the ``LLVMContext``
2804 error reporting mechanisms. This allows a front-end to correlate backend
2805 errors that occur with inline asm back to the source code that produced
2808 .. code-block:: llvm
2810 call void asm sideeffect "something bad", ""(), !srcloc !42
2812 !42 = !{ i32 1234567 }
2814 It is up to the front-end to make sense of the magic numbers it places
2815 in the IR. If the MDNode contains multiple constants, the code generator
2816 will use the one that corresponds to the line of the asm that the error
2824 LLVM IR allows metadata to be attached to instructions in the program
2825 that can convey extra information about the code to the optimizers and
2826 code generator. One example application of metadata is source-level
2827 debug information. There are two metadata primitives: strings and nodes.
2829 Metadata does not have a type, and is not a value. If referenced from a
2830 ``call`` instruction, it uses the ``metadata`` type.
2832 All metadata are identified in syntax by a exclamation point ('``!``').
2834 Metadata Nodes and Metadata Strings
2835 -----------------------------------
2837 A metadata string is a string surrounded by double quotes. It can
2838 contain any character by escaping non-printable characters with
2839 "``\xx``" where "``xx``" is the two digit hex code. For example:
2842 Metadata nodes are represented with notation similar to structure
2843 constants (a comma separated list of elements, surrounded by braces and
2844 preceded by an exclamation point). Metadata nodes can have any values as
2845 their operand. For example:
2847 .. code-block:: llvm
2849 !{ !"test\00", i32 10}
2851 Metadata nodes that aren't uniqued use the ``distinct`` keyword. For example:
2853 .. code-block:: llvm
2855 !0 = distinct !{!"test\00", i32 10}
2857 ``distinct`` nodes are useful when nodes shouldn't be merged based on their
2858 content. They can also occur when transformations cause uniquing collisions
2859 when metadata operands change.
2861 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2862 metadata nodes, which can be looked up in the module symbol table. For
2865 .. code-block:: llvm
2869 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2870 function is using two metadata arguments:
2872 .. code-block:: llvm
2874 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2876 Metadata can be attached with an instruction. Here metadata ``!21`` is
2877 attached to the ``add`` instruction using the ``!dbg`` identifier:
2879 .. code-block:: llvm
2881 %indvar.next = add i64 %indvar, 1, !dbg !21
2883 More information about specific metadata nodes recognized by the
2884 optimizers and code generator is found below.
2886 Specialized Metadata Nodes
2887 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2889 Specialized metadata nodes are custom data structures in metadata (as opposed
2890 to generic tuples). Their fields are labelled, and can be specified in any
2896 ``MDLocation`` nodes represent source debug locations. The ``scope:`` field is
2899 .. code-block:: llvm
2901 !0 = !MDLocation(line: 2900, column: 42, scope: !1, inlinedAt: !2)
2906 In LLVM IR, memory does not have types, so LLVM's own type system is not
2907 suitable for doing TBAA. Instead, metadata is added to the IR to
2908 describe a type system of a higher level language. This can be used to
2909 implement typical C/C++ TBAA, but it can also be used to implement
2910 custom alias analysis behavior for other languages.
2912 The current metadata format is very simple. TBAA metadata nodes have up
2913 to three fields, e.g.:
2915 .. code-block:: llvm
2917 !0 = !{ !"an example type tree" }
2918 !1 = !{ !"int", !0 }
2919 !2 = !{ !"float", !0 }
2920 !3 = !{ !"const float", !2, i64 1 }
2922 The first field is an identity field. It can be any value, usually a
2923 metadata string, which uniquely identifies the type. The most important
2924 name in the tree is the name of the root node. Two trees with different
2925 root node names are entirely disjoint, even if they have leaves with
2928 The second field identifies the type's parent node in the tree, or is
2929 null or omitted for a root node. A type is considered to alias all of
2930 its descendants and all of its ancestors in the tree. Also, a type is
2931 considered to alias all types in other trees, so that bitcode produced
2932 from multiple front-ends is handled conservatively.
2934 If the third field is present, it's an integer which if equal to 1
2935 indicates that the type is "constant" (meaning
2936 ``pointsToConstantMemory`` should return true; see `other useful
2937 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
2939 '``tbaa.struct``' Metadata
2940 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2942 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
2943 aggregate assignment operations in C and similar languages, however it
2944 is defined to copy a contiguous region of memory, which is more than
2945 strictly necessary for aggregate types which contain holes due to
2946 padding. Also, it doesn't contain any TBAA information about the fields
2949 ``!tbaa.struct`` metadata can describe which memory subregions in a
2950 memcpy are padding and what the TBAA tags of the struct are.
2952 The current metadata format is very simple. ``!tbaa.struct`` metadata
2953 nodes are a list of operands which are in conceptual groups of three.
2954 For each group of three, the first operand gives the byte offset of a
2955 field in bytes, the second gives its size in bytes, and the third gives
2958 .. code-block:: llvm
2960 !4 = !{ i64 0, i64 4, !1, i64 8, i64 4, !2 }
2962 This describes a struct with two fields. The first is at offset 0 bytes
2963 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
2964 and has size 4 bytes and has tbaa tag !2.
2966 Note that the fields need not be contiguous. In this example, there is a
2967 4 byte gap between the two fields. This gap represents padding which
2968 does not carry useful data and need not be preserved.
2970 '``noalias``' and '``alias.scope``' Metadata
2971 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2973 ``noalias`` and ``alias.scope`` metadata provide the ability to specify generic
2974 noalias memory-access sets. This means that some collection of memory access
2975 instructions (loads, stores, memory-accessing calls, etc.) that carry
2976 ``noalias`` metadata can specifically be specified not to alias with some other
2977 collection of memory access instructions that carry ``alias.scope`` metadata.
2978 Each type of metadata specifies a list of scopes where each scope has an id and
2979 a domain. When evaluating an aliasing query, if for some some domain, the set
2980 of scopes with that domain in one instruction's ``alias.scope`` list is a
2981 subset of (or qual to) the set of scopes for that domain in another
2982 instruction's ``noalias`` list, then the two memory accesses are assumed not to
2985 The metadata identifying each domain is itself a list containing one or two
2986 entries. The first entry is the name of the domain. Note that if the name is a
2987 string then it can be combined accross functions and translation units. A
2988 self-reference can be used to create globally unique domain names. A
2989 descriptive string may optionally be provided as a second list entry.
2991 The metadata identifying each scope is also itself a list containing two or
2992 three entries. The first entry is the name of the scope. Note that if the name
2993 is a string then it can be combined accross functions and translation units. A
2994 self-reference can be used to create globally unique scope names. A metadata
2995 reference to the scope's domain is the second entry. A descriptive string may
2996 optionally be provided as a third list entry.
3000 .. code-block:: llvm
3002 ; Two scope domains:
3006 ; Some scopes in these domains:
3012 !5 = !{!4} ; A list containing only scope !4
3016 ; These two instructions don't alias:
3017 %0 = load float* %c, align 4, !alias.scope !5
3018 store float %0, float* %arrayidx.i, align 4, !noalias !5
3020 ; These two instructions also don't alias (for domain !1, the set of scopes
3021 ; in the !alias.scope equals that in the !noalias list):
3022 %2 = load float* %c, align 4, !alias.scope !5
3023 store float %2, float* %arrayidx.i2, align 4, !noalias !6
3025 ; These two instructions don't alias (for domain !0, the set of scopes in
3026 ; the !noalias list is not a superset of, or equal to, the scopes in the
3027 ; !alias.scope list):
3028 %2 = load float* %c, align 4, !alias.scope !6
3029 store float %0, float* %arrayidx.i, align 4, !noalias !7
3031 '``fpmath``' Metadata
3032 ^^^^^^^^^^^^^^^^^^^^^
3034 ``fpmath`` metadata may be attached to any instruction of floating point
3035 type. It can be used to express the maximum acceptable error in the
3036 result of that instruction, in ULPs, thus potentially allowing the
3037 compiler to use a more efficient but less accurate method of computing
3038 it. ULP is defined as follows:
3040 If ``x`` is a real number that lies between two finite consecutive
3041 floating-point numbers ``a`` and ``b``, without being equal to one
3042 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
3043 distance between the two non-equal finite floating-point numbers
3044 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
3046 The metadata node shall consist of a single positive floating point
3047 number representing the maximum relative error, for example:
3049 .. code-block:: llvm
3051 !0 = !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
3053 '``range``' Metadata
3054 ^^^^^^^^^^^^^^^^^^^^
3056 ``range`` metadata may be attached only to ``load``, ``call`` and ``invoke`` of
3057 integer types. It expresses the possible ranges the loaded value or the value
3058 returned by the called function at this call site is in. The ranges are
3059 represented with a flattened list of integers. The loaded value or the value
3060 returned is known to be in the union of the ranges defined by each consecutive
3061 pair. Each pair has the following properties:
3063 - The type must match the type loaded by the instruction.
3064 - The pair ``a,b`` represents the range ``[a,b)``.
3065 - Both ``a`` and ``b`` are constants.
3066 - The range is allowed to wrap.
3067 - The range should not represent the full or empty set. That is,
3070 In addition, the pairs must be in signed order of the lower bound and
3071 they must be non-contiguous.
3075 .. code-block:: llvm
3077 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
3078 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
3079 %c = call i8 @foo(), !range !2 ; Can only be 0, 1, 3, 4 or 5
3080 %d = invoke i8 @bar() to label %cont
3081 unwind label %lpad, !range !3 ; Can only be -2, -1, 3, 4 or 5
3083 !0 = !{ i8 0, i8 2 }
3084 !1 = !{ i8 255, i8 2 }
3085 !2 = !{ i8 0, i8 2, i8 3, i8 6 }
3086 !3 = !{ i8 -2, i8 0, i8 3, i8 6 }
3091 It is sometimes useful to attach information to loop constructs. Currently,
3092 loop metadata is implemented as metadata attached to the branch instruction
3093 in the loop latch block. This type of metadata refer to a metadata node that is
3094 guaranteed to be separate for each loop. The loop identifier metadata is
3095 specified with the name ``llvm.loop``.
3097 The loop identifier metadata is implemented using a metadata that refers to
3098 itself to avoid merging it with any other identifier metadata, e.g.,
3099 during module linkage or function inlining. That is, each loop should refer
3100 to their own identification metadata even if they reside in separate functions.
3101 The following example contains loop identifier metadata for two separate loop
3104 .. code-block:: llvm
3109 The loop identifier metadata can be used to specify additional
3110 per-loop metadata. Any operands after the first operand can be treated
3111 as user-defined metadata. For example the ``llvm.loop.unroll.count``
3112 suggests an unroll factor to the loop unroller:
3114 .. code-block:: llvm
3116 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
3119 !1 = !{!"llvm.loop.unroll.count", i32 4}
3121 '``llvm.loop.vectorize``' and '``llvm.loop.interleave``'
3122 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3124 Metadata prefixed with ``llvm.loop.vectorize`` or ``llvm.loop.interleave`` are
3125 used to control per-loop vectorization and interleaving parameters such as
3126 vectorization width and interleave count. These metadata should be used in
3127 conjunction with ``llvm.loop`` loop identification metadata. The
3128 ``llvm.loop.vectorize`` and ``llvm.loop.interleave`` metadata are only
3129 optimization hints and the optimizer will only interleave and vectorize loops if
3130 it believes it is safe to do so. The ``llvm.mem.parallel_loop_access`` metadata
3131 which contains information about loop-carried memory dependencies can be helpful
3132 in determining the safety of these transformations.
3134 '``llvm.loop.interleave.count``' Metadata
3135 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3137 This metadata suggests an interleave count to the loop interleaver.
3138 The first operand is the string ``llvm.loop.interleave.count`` and the
3139 second operand is an integer specifying the interleave count. For
3142 .. code-block:: llvm
3144 !0 = !{!"llvm.loop.interleave.count", i32 4}
3146 Note that setting ``llvm.loop.interleave.count`` to 1 disables interleaving
3147 multiple iterations of the loop. If ``llvm.loop.interleave.count`` is set to 0
3148 then the interleave count will be determined automatically.
3150 '``llvm.loop.vectorize.enable``' Metadata
3151 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3153 This metadata selectively enables or disables vectorization for the loop. The
3154 first operand is the string ``llvm.loop.vectorize.enable`` and the second operand
3155 is a bit. If the bit operand value is 1 vectorization is enabled. A value of
3156 0 disables vectorization:
3158 .. code-block:: llvm
3160 !0 = !{!"llvm.loop.vectorize.enable", i1 0}
3161 !1 = !{!"llvm.loop.vectorize.enable", i1 1}
3163 '``llvm.loop.vectorize.width``' Metadata
3164 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3166 This metadata sets the target width of the vectorizer. The first
3167 operand is the string ``llvm.loop.vectorize.width`` and the second
3168 operand is an integer specifying the width. For example:
3170 .. code-block:: llvm
3172 !0 = !{!"llvm.loop.vectorize.width", i32 4}
3174 Note that setting ``llvm.loop.vectorize.width`` to 1 disables
3175 vectorization of the loop. If ``llvm.loop.vectorize.width`` is set to
3176 0 or if the loop does not have this metadata the width will be
3177 determined automatically.
3179 '``llvm.loop.unroll``'
3180 ^^^^^^^^^^^^^^^^^^^^^^
3182 Metadata prefixed with ``llvm.loop.unroll`` are loop unrolling
3183 optimization hints such as the unroll factor. ``llvm.loop.unroll``
3184 metadata should be used in conjunction with ``llvm.loop`` loop
3185 identification metadata. The ``llvm.loop.unroll`` metadata are only
3186 optimization hints and the unrolling will only be performed if the
3187 optimizer believes it is safe to do so.
3189 '``llvm.loop.unroll.count``' Metadata
3190 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3192 This metadata suggests an unroll factor to the loop unroller. The
3193 first operand is the string ``llvm.loop.unroll.count`` and the second
3194 operand is a positive integer specifying the unroll factor. For
3197 .. code-block:: llvm
3199 !0 = !{!"llvm.loop.unroll.count", i32 4}
3201 If the trip count of the loop is less than the unroll count the loop
3202 will be partially unrolled.
3204 '``llvm.loop.unroll.disable``' Metadata
3205 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3207 This metadata either disables loop unrolling. The metadata has a single operand
3208 which is the string ``llvm.loop.unroll.disable``. For example:
3210 .. code-block:: llvm
3212 !0 = !{!"llvm.loop.unroll.disable"}
3214 '``llvm.loop.unroll.full``' Metadata
3215 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3217 This metadata either suggests that the loop should be unrolled fully. The
3218 metadata has a single operand which is the string ``llvm.loop.unroll.disable``.
3221 .. code-block:: llvm
3223 !0 = !{!"llvm.loop.unroll.full"}
3228 Metadata types used to annotate memory accesses with information helpful
3229 for optimizations are prefixed with ``llvm.mem``.
3231 '``llvm.mem.parallel_loop_access``' Metadata
3232 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3234 The ``llvm.mem.parallel_loop_access`` metadata refers to a loop identifier,
3235 or metadata containing a list of loop identifiers for nested loops.
3236 The metadata is attached to memory accessing instructions and denotes that
3237 no loop carried memory dependence exist between it and other instructions denoted
3238 with the same loop identifier.
3240 Precisely, given two instructions ``m1`` and ``m2`` that both have the
3241 ``llvm.mem.parallel_loop_access`` metadata, with ``L1`` and ``L2`` being the
3242 set of loops associated with that metadata, respectively, then there is no loop
3243 carried dependence between ``m1`` and ``m2`` for loops in both ``L1`` and
3246 As a special case, if all memory accessing instructions in a loop have
3247 ``llvm.mem.parallel_loop_access`` metadata that refers to that loop, then the
3248 loop has no loop carried memory dependences and is considered to be a parallel
3251 Note that if not all memory access instructions have such metadata referring to
3252 the loop, then the loop is considered not being trivially parallel. Additional
3253 memory dependence analysis is required to make that determination. As a fail
3254 safe mechanism, this causes loops that were originally parallel to be considered
3255 sequential (if optimization passes that are unaware of the parallel semantics
3256 insert new memory instructions into the loop body).
3258 Example of a loop that is considered parallel due to its correct use of
3259 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
3260 metadata types that refer to the same loop identifier metadata.
3262 .. code-block:: llvm
3266 %val0 = load i32* %arrayidx, !llvm.mem.parallel_loop_access !0
3268 store i32 %val0, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3270 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
3276 It is also possible to have nested parallel loops. In that case the
3277 memory accesses refer to a list of loop identifier metadata nodes instead of
3278 the loop identifier metadata node directly:
3280 .. code-block:: llvm
3284 %val1 = load i32* %arrayidx3, !llvm.mem.parallel_loop_access !2
3286 br label %inner.for.body
3290 %val0 = load i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3292 store i32 %val0, i32* %arrayidx2, !llvm.mem.parallel_loop_access !0
3294 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
3298 store i32 %val1, i32* %arrayidx4, !llvm.mem.parallel_loop_access !2
3300 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
3302 outer.for.end: ; preds = %for.body
3304 !0 = !{!1, !2} ; a list of loop identifiers
3305 !1 = !{!1} ; an identifier for the inner loop
3306 !2 = !{!2} ; an identifier for the outer loop
3308 Module Flags Metadata
3309 =====================
3311 Information about the module as a whole is difficult to convey to LLVM's
3312 subsystems. The LLVM IR isn't sufficient to transmit this information.
3313 The ``llvm.module.flags`` named metadata exists in order to facilitate
3314 this. These flags are in the form of key / value pairs --- much like a
3315 dictionary --- making it easy for any subsystem who cares about a flag to
3318 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
3319 Each triplet has the following form:
3321 - The first element is a *behavior* flag, which specifies the behavior
3322 when two (or more) modules are merged together, and it encounters two
3323 (or more) metadata with the same ID. The supported behaviors are
3325 - The second element is a metadata string that is a unique ID for the
3326 metadata. Each module may only have one flag entry for each unique ID (not
3327 including entries with the **Require** behavior).
3328 - The third element is the value of the flag.
3330 When two (or more) modules are merged together, the resulting
3331 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
3332 each unique metadata ID string, there will be exactly one entry in the merged
3333 modules ``llvm.module.flags`` metadata table, and the value for that entry will
3334 be determined by the merge behavior flag, as described below. The only exception
3335 is that entries with the *Require* behavior are always preserved.
3337 The following behaviors are supported:
3348 Emits an error if two values disagree, otherwise the resulting value
3349 is that of the operands.
3353 Emits a warning if two values disagree. The result value will be the
3354 operand for the flag from the first module being linked.
3358 Adds a requirement that another module flag be present and have a
3359 specified value after linking is performed. The value must be a
3360 metadata pair, where the first element of the pair is the ID of the
3361 module flag to be restricted, and the second element of the pair is
3362 the value the module flag should be restricted to. This behavior can
3363 be used to restrict the allowable results (via triggering of an
3364 error) of linking IDs with the **Override** behavior.
3368 Uses the specified value, regardless of the behavior or value of the
3369 other module. If both modules specify **Override**, but the values
3370 differ, an error will be emitted.
3374 Appends the two values, which are required to be metadata nodes.
3378 Appends the two values, which are required to be metadata
3379 nodes. However, duplicate entries in the second list are dropped
3380 during the append operation.
3382 It is an error for a particular unique flag ID to have multiple behaviors,
3383 except in the case of **Require** (which adds restrictions on another metadata
3384 value) or **Override**.
3386 An example of module flags:
3388 .. code-block:: llvm
3390 !0 = !{ i32 1, !"foo", i32 1 }
3391 !1 = !{ i32 4, !"bar", i32 37 }
3392 !2 = !{ i32 2, !"qux", i32 42 }
3393 !3 = !{ i32 3, !"qux",
3398 !llvm.module.flags = !{ !0, !1, !2, !3 }
3400 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
3401 if two or more ``!"foo"`` flags are seen is to emit an error if their
3402 values are not equal.
3404 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
3405 behavior if two or more ``!"bar"`` flags are seen is to use the value
3408 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
3409 behavior if two or more ``!"qux"`` flags are seen is to emit a
3410 warning if their values are not equal.
3412 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
3418 The behavior is to emit an error if the ``llvm.module.flags`` does not
3419 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
3422 Objective-C Garbage Collection Module Flags Metadata
3423 ----------------------------------------------------
3425 On the Mach-O platform, Objective-C stores metadata about garbage
3426 collection in a special section called "image info". The metadata
3427 consists of a version number and a bitmask specifying what types of
3428 garbage collection are supported (if any) by the file. If two or more
3429 modules are linked together their garbage collection metadata needs to
3430 be merged rather than appended together.
3432 The Objective-C garbage collection module flags metadata consists of the
3433 following key-value pairs:
3442 * - ``Objective-C Version``
3443 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
3445 * - ``Objective-C Image Info Version``
3446 - **[Required]** --- The version of the image info section. Currently
3449 * - ``Objective-C Image Info Section``
3450 - **[Required]** --- The section to place the metadata. Valid values are
3451 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
3452 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
3453 Objective-C ABI version 2.
3455 * - ``Objective-C Garbage Collection``
3456 - **[Required]** --- Specifies whether garbage collection is supported or
3457 not. Valid values are 0, for no garbage collection, and 2, for garbage
3458 collection supported.
3460 * - ``Objective-C GC Only``
3461 - **[Optional]** --- Specifies that only garbage collection is supported.
3462 If present, its value must be 6. This flag requires that the
3463 ``Objective-C Garbage Collection`` flag have the value 2.
3465 Some important flag interactions:
3467 - If a module with ``Objective-C Garbage Collection`` set to 0 is
3468 merged with a module with ``Objective-C Garbage Collection`` set to
3469 2, then the resulting module has the
3470 ``Objective-C Garbage Collection`` flag set to 0.
3471 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
3472 merged with a module with ``Objective-C GC Only`` set to 6.
3474 Automatic Linker Flags Module Flags Metadata
3475 --------------------------------------------
3477 Some targets support embedding flags to the linker inside individual object
3478 files. Typically this is used in conjunction with language extensions which
3479 allow source files to explicitly declare the libraries they depend on, and have
3480 these automatically be transmitted to the linker via object files.
3482 These flags are encoded in the IR using metadata in the module flags section,
3483 using the ``Linker Options`` key. The merge behavior for this flag is required
3484 to be ``AppendUnique``, and the value for the key is expected to be a metadata
3485 node which should be a list of other metadata nodes, each of which should be a
3486 list of metadata strings defining linker options.
3488 For example, the following metadata section specifies two separate sets of
3489 linker options, presumably to link against ``libz`` and the ``Cocoa``
3492 !0 = !{ i32 6, !"Linker Options",
3495 !{ !"-framework", !"Cocoa" } } }
3496 !llvm.module.flags = !{ !0 }
3498 The metadata encoding as lists of lists of options, as opposed to a collapsed
3499 list of options, is chosen so that the IR encoding can use multiple option
3500 strings to specify e.g., a single library, while still having that specifier be
3501 preserved as an atomic element that can be recognized by a target specific
3502 assembly writer or object file emitter.
3504 Each individual option is required to be either a valid option for the target's
3505 linker, or an option that is reserved by the target specific assembly writer or
3506 object file emitter. No other aspect of these options is defined by the IR.
3508 C type width Module Flags Metadata
3509 ----------------------------------
3511 The ARM backend emits a section into each generated object file describing the
3512 options that it was compiled with (in a compiler-independent way) to prevent
3513 linking incompatible objects, and to allow automatic library selection. Some
3514 of these options are not visible at the IR level, namely wchar_t width and enum
3517 To pass this information to the backend, these options are encoded in module
3518 flags metadata, using the following key-value pairs:
3528 - * 0 --- sizeof(wchar_t) == 4
3529 * 1 --- sizeof(wchar_t) == 2
3532 - * 0 --- Enums are at least as large as an ``int``.
3533 * 1 --- Enums are stored in the smallest integer type which can
3534 represent all of its values.
3536 For example, the following metadata section specifies that the module was
3537 compiled with a ``wchar_t`` width of 4 bytes, and the underlying type of an
3538 enum is the smallest type which can represent all of its values::
3540 !llvm.module.flags = !{!0, !1}
3541 !0 = !{i32 1, !"short_wchar", i32 1}
3542 !1 = !{i32 1, !"short_enum", i32 0}
3544 .. _intrinsicglobalvariables:
3546 Intrinsic Global Variables
3547 ==========================
3549 LLVM has a number of "magic" global variables that contain data that
3550 affect code generation or other IR semantics. These are documented here.
3551 All globals of this sort should have a section specified as
3552 "``llvm.metadata``". This section and all globals that start with
3553 "``llvm.``" are reserved for use by LLVM.
3557 The '``llvm.used``' Global Variable
3558 -----------------------------------
3560 The ``@llvm.used`` global is an array which has
3561 :ref:`appending linkage <linkage_appending>`. This array contains a list of
3562 pointers to named global variables, functions and aliases which may optionally
3563 have a pointer cast formed of bitcast or getelementptr. For example, a legal
3566 .. code-block:: llvm
3571 @llvm.used = appending global [2 x i8*] [
3573 i8* bitcast (i32* @Y to i8*)
3574 ], section "llvm.metadata"
3576 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
3577 and linker are required to treat the symbol as if there is a reference to the
3578 symbol that it cannot see (which is why they have to be named). For example, if
3579 a variable has internal linkage and no references other than that from the
3580 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
3581 references from inline asms and other things the compiler cannot "see", and
3582 corresponds to "``attribute((used))``" in GNU C.
3584 On some targets, the code generator must emit a directive to the
3585 assembler or object file to prevent the assembler and linker from
3586 molesting the symbol.
3588 .. _gv_llvmcompilerused:
3590 The '``llvm.compiler.used``' Global Variable
3591 --------------------------------------------
3593 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
3594 directive, except that it only prevents the compiler from touching the
3595 symbol. On targets that support it, this allows an intelligent linker to
3596 optimize references to the symbol without being impeded as it would be
3599 This is a rare construct that should only be used in rare circumstances,
3600 and should not be exposed to source languages.
3602 .. _gv_llvmglobalctors:
3604 The '``llvm.global_ctors``' Global Variable
3605 -------------------------------------------
3607 .. code-block:: llvm
3609 %0 = type { i32, void ()*, i8* }
3610 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor, i8* @data }]
3612 The ``@llvm.global_ctors`` array contains a list of constructor
3613 functions, priorities, and an optional associated global or function.
3614 The functions referenced by this array will be called in ascending order
3615 of priority (i.e. lowest first) when the module is loaded. The order of
3616 functions with the same priority is not defined.
3618 If the third field is present, non-null, and points to a global variable
3619 or function, the initializer function will only run if the associated
3620 data from the current module is not discarded.
3622 .. _llvmglobaldtors:
3624 The '``llvm.global_dtors``' Global Variable
3625 -------------------------------------------
3627 .. code-block:: llvm
3629 %0 = type { i32, void ()*, i8* }
3630 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor, i8* @data }]
3632 The ``@llvm.global_dtors`` array contains a list of destructor
3633 functions, priorities, and an optional associated global or function.
3634 The functions referenced by this array will be called in descending
3635 order of priority (i.e. highest first) when the module is unloaded. The
3636 order of functions with the same priority is not defined.
3638 If the third field is present, non-null, and points to a global variable
3639 or function, the destructor function will only run if the associated
3640 data from the current module is not discarded.
3642 Instruction Reference
3643 =====================
3645 The LLVM instruction set consists of several different classifications
3646 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
3647 instructions <binaryops>`, :ref:`bitwise binary
3648 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
3649 :ref:`other instructions <otherops>`.
3653 Terminator Instructions
3654 -----------------------
3656 As mentioned :ref:`previously <functionstructure>`, every basic block in a
3657 program ends with a "Terminator" instruction, which indicates which
3658 block should be executed after the current block is finished. These
3659 terminator instructions typically yield a '``void``' value: they produce
3660 control flow, not values (the one exception being the
3661 ':ref:`invoke <i_invoke>`' instruction).
3663 The terminator instructions are: ':ref:`ret <i_ret>`',
3664 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
3665 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
3666 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
3670 '``ret``' Instruction
3671 ^^^^^^^^^^^^^^^^^^^^^
3678 ret <type> <value> ; Return a value from a non-void function
3679 ret void ; Return from void function
3684 The '``ret``' instruction is used to return control flow (and optionally
3685 a value) from a function back to the caller.
3687 There are two forms of the '``ret``' instruction: one that returns a
3688 value and then causes control flow, and one that just causes control
3694 The '``ret``' instruction optionally accepts a single argument, the
3695 return value. The type of the return value must be a ':ref:`first
3696 class <t_firstclass>`' type.
3698 A function is not :ref:`well formed <wellformed>` if it it has a non-void
3699 return type and contains a '``ret``' instruction with no return value or
3700 a return value with a type that does not match its type, or if it has a
3701 void return type and contains a '``ret``' instruction with a return
3707 When the '``ret``' instruction is executed, control flow returns back to
3708 the calling function's context. If the caller is a
3709 ":ref:`call <i_call>`" instruction, execution continues at the
3710 instruction after the call. If the caller was an
3711 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
3712 beginning of the "normal" destination block. If the instruction returns
3713 a value, that value shall set the call or invoke instruction's return
3719 .. code-block:: llvm
3721 ret i32 5 ; Return an integer value of 5
3722 ret void ; Return from a void function
3723 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
3727 '``br``' Instruction
3728 ^^^^^^^^^^^^^^^^^^^^
3735 br i1 <cond>, label <iftrue>, label <iffalse>
3736 br label <dest> ; Unconditional branch
3741 The '``br``' instruction is used to cause control flow to transfer to a
3742 different basic block in the current function. There are two forms of
3743 this instruction, corresponding to a conditional branch and an
3744 unconditional branch.
3749 The conditional branch form of the '``br``' instruction takes a single
3750 '``i1``' value and two '``label``' values. The unconditional form of the
3751 '``br``' instruction takes a single '``label``' value as a target.
3756 Upon execution of a conditional '``br``' instruction, the '``i1``'
3757 argument is evaluated. If the value is ``true``, control flows to the
3758 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
3759 to the '``iffalse``' ``label`` argument.
3764 .. code-block:: llvm
3767 %cond = icmp eq i32 %a, %b
3768 br i1 %cond, label %IfEqual, label %IfUnequal
3776 '``switch``' Instruction
3777 ^^^^^^^^^^^^^^^^^^^^^^^^
3784 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
3789 The '``switch``' instruction is used to transfer control flow to one of
3790 several different places. It is a generalization of the '``br``'
3791 instruction, allowing a branch to occur to one of many possible
3797 The '``switch``' instruction uses three parameters: an integer
3798 comparison value '``value``', a default '``label``' destination, and an
3799 array of pairs of comparison value constants and '``label``'s. The table
3800 is not allowed to contain duplicate constant entries.
3805 The ``switch`` instruction specifies a table of values and destinations.
3806 When the '``switch``' instruction is executed, this table is searched
3807 for the given value. If the value is found, control flow is transferred
3808 to the corresponding destination; otherwise, control flow is transferred
3809 to the default destination.
3814 Depending on properties of the target machine and the particular
3815 ``switch`` instruction, this instruction may be code generated in
3816 different ways. For example, it could be generated as a series of
3817 chained conditional branches or with a lookup table.
3822 .. code-block:: llvm
3824 ; Emulate a conditional br instruction
3825 %Val = zext i1 %value to i32
3826 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
3828 ; Emulate an unconditional br instruction
3829 switch i32 0, label %dest [ ]
3831 ; Implement a jump table:
3832 switch i32 %val, label %otherwise [ i32 0, label %onzero
3834 i32 2, label %ontwo ]
3838 '``indirectbr``' Instruction
3839 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3846 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
3851 The '``indirectbr``' instruction implements an indirect branch to a
3852 label within the current function, whose address is specified by
3853 "``address``". Address must be derived from a
3854 :ref:`blockaddress <blockaddress>` constant.
3859 The '``address``' argument is the address of the label to jump to. The
3860 rest of the arguments indicate the full set of possible destinations
3861 that the address may point to. Blocks are allowed to occur multiple
3862 times in the destination list, though this isn't particularly useful.
3864 This destination list is required so that dataflow analysis has an
3865 accurate understanding of the CFG.
3870 Control transfers to the block specified in the address argument. All
3871 possible destination blocks must be listed in the label list, otherwise
3872 this instruction has undefined behavior. This implies that jumps to
3873 labels defined in other functions have undefined behavior as well.
3878 This is typically implemented with a jump through a register.
3883 .. code-block:: llvm
3885 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
3889 '``invoke``' Instruction
3890 ^^^^^^^^^^^^^^^^^^^^^^^^
3897 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
3898 to label <normal label> unwind label <exception label>
3903 The '``invoke``' instruction causes control to transfer to a specified
3904 function, with the possibility of control flow transfer to either the
3905 '``normal``' label or the '``exception``' label. If the callee function
3906 returns with the "``ret``" instruction, control flow will return to the
3907 "normal" label. If the callee (or any indirect callees) returns via the
3908 ":ref:`resume <i_resume>`" instruction or other exception handling
3909 mechanism, control is interrupted and continued at the dynamically
3910 nearest "exception" label.
3912 The '``exception``' label is a `landing
3913 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
3914 '``exception``' label is required to have the
3915 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
3916 information about the behavior of the program after unwinding happens,
3917 as its first non-PHI instruction. The restrictions on the
3918 "``landingpad``" instruction's tightly couples it to the "``invoke``"
3919 instruction, so that the important information contained within the
3920 "``landingpad``" instruction can't be lost through normal code motion.
3925 This instruction requires several arguments:
3927 #. The optional "cconv" marker indicates which :ref:`calling
3928 convention <callingconv>` the call should use. If none is
3929 specified, the call defaults to using C calling conventions.
3930 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
3931 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
3933 #. '``ptr to function ty``': shall be the signature of the pointer to
3934 function value being invoked. In most cases, this is a direct
3935 function invocation, but indirect ``invoke``'s are just as possible,
3936 branching off an arbitrary pointer to function value.
3937 #. '``function ptr val``': An LLVM value containing a pointer to a
3938 function to be invoked.
3939 #. '``function args``': argument list whose types match the function
3940 signature argument types and parameter attributes. All arguments must
3941 be of :ref:`first class <t_firstclass>` type. If the function signature
3942 indicates the function accepts a variable number of arguments, the
3943 extra arguments can be specified.
3944 #. '``normal label``': the label reached when the called function
3945 executes a '``ret``' instruction.
3946 #. '``exception label``': the label reached when a callee returns via
3947 the :ref:`resume <i_resume>` instruction or other exception handling
3949 #. The optional :ref:`function attributes <fnattrs>` list. Only
3950 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
3951 attributes are valid here.
3956 This instruction is designed to operate as a standard '``call``'
3957 instruction in most regards. The primary difference is that it
3958 establishes an association with a label, which is used by the runtime
3959 library to unwind the stack.
3961 This instruction is used in languages with destructors to ensure that
3962 proper cleanup is performed in the case of either a ``longjmp`` or a
3963 thrown exception. Additionally, this is important for implementation of
3964 '``catch``' clauses in high-level languages that support them.
3966 For the purposes of the SSA form, the definition of the value returned
3967 by the '``invoke``' instruction is deemed to occur on the edge from the
3968 current block to the "normal" label. If the callee unwinds then no
3969 return value is available.
3974 .. code-block:: llvm
3976 %retval = invoke i32 @Test(i32 15) to label %Continue
3977 unwind label %TestCleanup ; i32:retval set
3978 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
3979 unwind label %TestCleanup ; i32:retval set
3983 '``resume``' Instruction
3984 ^^^^^^^^^^^^^^^^^^^^^^^^
3991 resume <type> <value>
3996 The '``resume``' instruction is a terminator instruction that has no
4002 The '``resume``' instruction requires one argument, which must have the
4003 same type as the result of any '``landingpad``' instruction in the same
4009 The '``resume``' instruction resumes propagation of an existing
4010 (in-flight) exception whose unwinding was interrupted with a
4011 :ref:`landingpad <i_landingpad>` instruction.
4016 .. code-block:: llvm
4018 resume { i8*, i32 } %exn
4022 '``unreachable``' Instruction
4023 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4035 The '``unreachable``' instruction has no defined semantics. This
4036 instruction is used to inform the optimizer that a particular portion of
4037 the code is not reachable. This can be used to indicate that the code
4038 after a no-return function cannot be reached, and other facts.
4043 The '``unreachable``' instruction has no defined semantics.
4050 Binary operators are used to do most of the computation in a program.
4051 They require two operands of the same type, execute an operation on
4052 them, and produce a single value. The operands might represent multiple
4053 data, as is the case with the :ref:`vector <t_vector>` data type. The
4054 result value has the same type as its operands.
4056 There are several different binary operators:
4060 '``add``' Instruction
4061 ^^^^^^^^^^^^^^^^^^^^^
4068 <result> = add <ty> <op1>, <op2> ; yields ty:result
4069 <result> = add nuw <ty> <op1>, <op2> ; yields ty:result
4070 <result> = add nsw <ty> <op1>, <op2> ; yields ty:result
4071 <result> = add nuw nsw <ty> <op1>, <op2> ; yields ty:result
4076 The '``add``' instruction returns the sum of its two operands.
4081 The two arguments to the '``add``' instruction must be
4082 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4083 arguments must have identical types.
4088 The value produced is the integer sum of the two operands.
4090 If the sum has unsigned overflow, the result returned is the
4091 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4094 Because LLVM integers use a two's complement representation, this
4095 instruction is appropriate for both signed and unsigned integers.
4097 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4098 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4099 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
4100 unsigned and/or signed overflow, respectively, occurs.
4105 .. code-block:: llvm
4107 <result> = add i32 4, %var ; yields i32:result = 4 + %var
4111 '``fadd``' Instruction
4112 ^^^^^^^^^^^^^^^^^^^^^^
4119 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4124 The '``fadd``' instruction returns the sum of its two operands.
4129 The two arguments to the '``fadd``' instruction must be :ref:`floating
4130 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4131 Both arguments must have identical types.
4136 The value produced is the floating point sum of the two operands. This
4137 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
4138 which are optimization hints to enable otherwise unsafe floating point
4144 .. code-block:: llvm
4146 <result> = fadd float 4.0, %var ; yields float:result = 4.0 + %var
4148 '``sub``' Instruction
4149 ^^^^^^^^^^^^^^^^^^^^^
4156 <result> = sub <ty> <op1>, <op2> ; yields ty:result
4157 <result> = sub nuw <ty> <op1>, <op2> ; yields ty:result
4158 <result> = sub nsw <ty> <op1>, <op2> ; yields ty:result
4159 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields ty:result
4164 The '``sub``' instruction returns the difference of its two operands.
4166 Note that the '``sub``' instruction is used to represent the '``neg``'
4167 instruction present in most other intermediate representations.
4172 The two arguments to the '``sub``' instruction must be
4173 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4174 arguments must have identical types.
4179 The value produced is the integer difference of the two operands.
4181 If the difference has unsigned overflow, the result returned is the
4182 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4185 Because LLVM integers use a two's complement representation, this
4186 instruction is appropriate for both signed and unsigned integers.
4188 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4189 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4190 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
4191 unsigned and/or signed overflow, respectively, occurs.
4196 .. code-block:: llvm
4198 <result> = sub i32 4, %var ; yields i32:result = 4 - %var
4199 <result> = sub i32 0, %val ; yields i32:result = -%var
4203 '``fsub``' Instruction
4204 ^^^^^^^^^^^^^^^^^^^^^^
4211 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4216 The '``fsub``' instruction returns the difference of its two operands.
4218 Note that the '``fsub``' instruction is used to represent the '``fneg``'
4219 instruction present in most other intermediate representations.
4224 The two arguments to the '``fsub``' instruction must be :ref:`floating
4225 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4226 Both arguments must have identical types.
4231 The value produced is the floating point difference of the two operands.
4232 This instruction can also take any number of :ref:`fast-math
4233 flags <fastmath>`, which are optimization hints to enable otherwise
4234 unsafe floating point optimizations:
4239 .. code-block:: llvm
4241 <result> = fsub float 4.0, %var ; yields float:result = 4.0 - %var
4242 <result> = fsub float -0.0, %val ; yields float:result = -%var
4244 '``mul``' Instruction
4245 ^^^^^^^^^^^^^^^^^^^^^
4252 <result> = mul <ty> <op1>, <op2> ; yields ty:result
4253 <result> = mul nuw <ty> <op1>, <op2> ; yields ty:result
4254 <result> = mul nsw <ty> <op1>, <op2> ; yields ty:result
4255 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields ty:result
4260 The '``mul``' instruction returns the product of its two operands.
4265 The two arguments to the '``mul``' instruction must be
4266 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4267 arguments must have identical types.
4272 The value produced is the integer product of the two operands.
4274 If the result of the multiplication has unsigned overflow, the result
4275 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
4276 bit width of the result.
4278 Because LLVM integers use a two's complement representation, and the
4279 result is the same width as the operands, this instruction returns the
4280 correct result for both signed and unsigned integers. If a full product
4281 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
4282 sign-extended or zero-extended as appropriate to the width of the full
4285 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4286 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4287 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
4288 unsigned and/or signed overflow, respectively, occurs.
4293 .. code-block:: llvm
4295 <result> = mul i32 4, %var ; yields i32:result = 4 * %var
4299 '``fmul``' Instruction
4300 ^^^^^^^^^^^^^^^^^^^^^^
4307 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4312 The '``fmul``' instruction returns the product of its two operands.
4317 The two arguments to the '``fmul``' instruction must be :ref:`floating
4318 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4319 Both arguments must have identical types.
4324 The value produced is the floating point product of the two operands.
4325 This instruction can also take any number of :ref:`fast-math
4326 flags <fastmath>`, which are optimization hints to enable otherwise
4327 unsafe floating point optimizations:
4332 .. code-block:: llvm
4334 <result> = fmul float 4.0, %var ; yields float:result = 4.0 * %var
4336 '``udiv``' Instruction
4337 ^^^^^^^^^^^^^^^^^^^^^^
4344 <result> = udiv <ty> <op1>, <op2> ; yields ty:result
4345 <result> = udiv exact <ty> <op1>, <op2> ; yields ty:result
4350 The '``udiv``' instruction returns the quotient of its two operands.
4355 The two arguments to the '``udiv``' instruction must be
4356 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4357 arguments must have identical types.
4362 The value produced is the unsigned integer quotient of the two operands.
4364 Note that unsigned integer division and signed integer division are
4365 distinct operations; for signed integer division, use '``sdiv``'.
4367 Division by zero leads to undefined behavior.
4369 If the ``exact`` keyword is present, the result value of the ``udiv`` is
4370 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
4371 such, "((a udiv exact b) mul b) == a").
4376 .. code-block:: llvm
4378 <result> = udiv i32 4, %var ; yields i32:result = 4 / %var
4380 '``sdiv``' Instruction
4381 ^^^^^^^^^^^^^^^^^^^^^^
4388 <result> = sdiv <ty> <op1>, <op2> ; yields ty:result
4389 <result> = sdiv exact <ty> <op1>, <op2> ; yields ty:result
4394 The '``sdiv``' instruction returns the quotient of its two operands.
4399 The two arguments to the '``sdiv``' instruction must be
4400 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4401 arguments must have identical types.
4406 The value produced is the signed integer quotient of the two operands
4407 rounded towards zero.
4409 Note that signed integer division and unsigned integer division are
4410 distinct operations; for unsigned integer division, use '``udiv``'.
4412 Division by zero leads to undefined behavior. Overflow also leads to
4413 undefined behavior; this is a rare case, but can occur, for example, by
4414 doing a 32-bit division of -2147483648 by -1.
4416 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
4417 a :ref:`poison value <poisonvalues>` if the result would be rounded.
4422 .. code-block:: llvm
4424 <result> = sdiv i32 4, %var ; yields i32:result = 4 / %var
4428 '``fdiv``' Instruction
4429 ^^^^^^^^^^^^^^^^^^^^^^
4436 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4441 The '``fdiv``' instruction returns the quotient of its two operands.
4446 The two arguments to the '``fdiv``' instruction must be :ref:`floating
4447 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4448 Both arguments must have identical types.
4453 The value produced is the floating point quotient of the two operands.
4454 This instruction can also take any number of :ref:`fast-math
4455 flags <fastmath>`, which are optimization hints to enable otherwise
4456 unsafe floating point optimizations:
4461 .. code-block:: llvm
4463 <result> = fdiv float 4.0, %var ; yields float:result = 4.0 / %var
4465 '``urem``' Instruction
4466 ^^^^^^^^^^^^^^^^^^^^^^
4473 <result> = urem <ty> <op1>, <op2> ; yields ty:result
4478 The '``urem``' instruction returns the remainder from the unsigned
4479 division of its two arguments.
4484 The two arguments to the '``urem``' instruction must be
4485 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4486 arguments must have identical types.
4491 This instruction returns the unsigned integer *remainder* of a division.
4492 This instruction always performs an unsigned division to get the
4495 Note that unsigned integer remainder and signed integer remainder are
4496 distinct operations; for signed integer remainder, use '``srem``'.
4498 Taking the remainder of a division by zero leads to undefined behavior.
4503 .. code-block:: llvm
4505 <result> = urem i32 4, %var ; yields i32:result = 4 % %var
4507 '``srem``' Instruction
4508 ^^^^^^^^^^^^^^^^^^^^^^
4515 <result> = srem <ty> <op1>, <op2> ; yields ty:result
4520 The '``srem``' instruction returns the remainder from the signed
4521 division of its two operands. This instruction can also take
4522 :ref:`vector <t_vector>` versions of the values in which case the elements
4528 The two arguments to the '``srem``' instruction must be
4529 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4530 arguments must have identical types.
4535 This instruction returns the *remainder* of a division (where the result
4536 is either zero or has the same sign as the dividend, ``op1``), not the
4537 *modulo* operator (where the result is either zero or has the same sign
4538 as the divisor, ``op2``) of a value. For more information about the
4539 difference, see `The Math
4540 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
4541 table of how this is implemented in various languages, please see
4543 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
4545 Note that signed integer remainder and unsigned integer remainder are
4546 distinct operations; for unsigned integer remainder, use '``urem``'.
4548 Taking the remainder of a division by zero leads to undefined behavior.
4549 Overflow also leads to undefined behavior; this is a rare case, but can
4550 occur, for example, by taking the remainder of a 32-bit division of
4551 -2147483648 by -1. (The remainder doesn't actually overflow, but this
4552 rule lets srem be implemented using instructions that return both the
4553 result of the division and the remainder.)
4558 .. code-block:: llvm
4560 <result> = srem i32 4, %var ; yields i32:result = 4 % %var
4564 '``frem``' Instruction
4565 ^^^^^^^^^^^^^^^^^^^^^^
4572 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4577 The '``frem``' instruction returns the remainder from the division of
4583 The two arguments to the '``frem``' instruction must be :ref:`floating
4584 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4585 Both arguments must have identical types.
4590 This instruction returns the *remainder* of a division. The remainder
4591 has the same sign as the dividend. This instruction can also take any
4592 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
4593 to enable otherwise unsafe floating point optimizations:
4598 .. code-block:: llvm
4600 <result> = frem float 4.0, %var ; yields float:result = 4.0 % %var
4604 Bitwise Binary Operations
4605 -------------------------
4607 Bitwise binary operators are used to do various forms of bit-twiddling
4608 in a program. They are generally very efficient instructions and can
4609 commonly be strength reduced from other instructions. They require two
4610 operands of the same type, execute an operation on them, and produce a
4611 single value. The resulting value is the same type as its operands.
4613 '``shl``' Instruction
4614 ^^^^^^^^^^^^^^^^^^^^^
4621 <result> = shl <ty> <op1>, <op2> ; yields ty:result
4622 <result> = shl nuw <ty> <op1>, <op2> ; yields ty:result
4623 <result> = shl nsw <ty> <op1>, <op2> ; yields ty:result
4624 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields ty:result
4629 The '``shl``' instruction returns the first operand shifted to the left
4630 a specified number of bits.
4635 Both arguments to the '``shl``' instruction must be the same
4636 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4637 '``op2``' is treated as an unsigned value.
4642 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
4643 where ``n`` is the width of the result. If ``op2`` is (statically or
4644 dynamically) negative or equal to or larger than the number of bits in
4645 ``op1``, the result is undefined. If the arguments are vectors, each
4646 vector element of ``op1`` is shifted by the corresponding shift amount
4649 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
4650 value <poisonvalues>` if it shifts out any non-zero bits. If the
4651 ``nsw`` keyword is present, then the shift produces a :ref:`poison
4652 value <poisonvalues>` if it shifts out any bits that disagree with the
4653 resultant sign bit. As such, NUW/NSW have the same semantics as they
4654 would if the shift were expressed as a mul instruction with the same
4655 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
4660 .. code-block:: llvm
4662 <result> = shl i32 4, %var ; yields i32: 4 << %var
4663 <result> = shl i32 4, 2 ; yields i32: 16
4664 <result> = shl i32 1, 10 ; yields i32: 1024
4665 <result> = shl i32 1, 32 ; undefined
4666 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
4668 '``lshr``' Instruction
4669 ^^^^^^^^^^^^^^^^^^^^^^
4676 <result> = lshr <ty> <op1>, <op2> ; yields ty:result
4677 <result> = lshr exact <ty> <op1>, <op2> ; yields ty:result
4682 The '``lshr``' instruction (logical shift right) returns the first
4683 operand shifted to the right a specified number of bits with zero fill.
4688 Both arguments to the '``lshr``' instruction must be the same
4689 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4690 '``op2``' is treated as an unsigned value.
4695 This instruction always performs a logical shift right operation. The
4696 most significant bits of the result will be filled with zero bits after
4697 the shift. If ``op2`` is (statically or dynamically) equal to or larger
4698 than the number of bits in ``op1``, the result is undefined. If the
4699 arguments are vectors, each vector element of ``op1`` is shifted by the
4700 corresponding shift amount in ``op2``.
4702 If the ``exact`` keyword is present, the result value of the ``lshr`` is
4703 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4709 .. code-block:: llvm
4711 <result> = lshr i32 4, 1 ; yields i32:result = 2
4712 <result> = lshr i32 4, 2 ; yields i32:result = 1
4713 <result> = lshr i8 4, 3 ; yields i8:result = 0
4714 <result> = lshr i8 -2, 1 ; yields i8:result = 0x7F
4715 <result> = lshr i32 1, 32 ; undefined
4716 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
4718 '``ashr``' Instruction
4719 ^^^^^^^^^^^^^^^^^^^^^^
4726 <result> = ashr <ty> <op1>, <op2> ; yields ty:result
4727 <result> = ashr exact <ty> <op1>, <op2> ; yields ty:result
4732 The '``ashr``' instruction (arithmetic shift right) returns the first
4733 operand shifted to the right a specified number of bits with sign
4739 Both arguments to the '``ashr``' instruction must be the same
4740 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4741 '``op2``' is treated as an unsigned value.
4746 This instruction always performs an arithmetic shift right operation,
4747 The most significant bits of the result will be filled with the sign bit
4748 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
4749 than the number of bits in ``op1``, the result is undefined. If the
4750 arguments are vectors, each vector element of ``op1`` is shifted by the
4751 corresponding shift amount in ``op2``.
4753 If the ``exact`` keyword is present, the result value of the ``ashr`` is
4754 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4760 .. code-block:: llvm
4762 <result> = ashr i32 4, 1 ; yields i32:result = 2
4763 <result> = ashr i32 4, 2 ; yields i32:result = 1
4764 <result> = ashr i8 4, 3 ; yields i8:result = 0
4765 <result> = ashr i8 -2, 1 ; yields i8:result = -1
4766 <result> = ashr i32 1, 32 ; undefined
4767 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
4769 '``and``' Instruction
4770 ^^^^^^^^^^^^^^^^^^^^^
4777 <result> = and <ty> <op1>, <op2> ; yields ty:result
4782 The '``and``' instruction returns the bitwise logical and of its two
4788 The two arguments to the '``and``' instruction must be
4789 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4790 arguments must have identical types.
4795 The truth table used for the '``and``' instruction is:
4812 .. code-block:: llvm
4814 <result> = and i32 4, %var ; yields i32:result = 4 & %var
4815 <result> = and i32 15, 40 ; yields i32:result = 8
4816 <result> = and i32 4, 8 ; yields i32:result = 0
4818 '``or``' Instruction
4819 ^^^^^^^^^^^^^^^^^^^^
4826 <result> = or <ty> <op1>, <op2> ; yields ty:result
4831 The '``or``' instruction returns the bitwise logical inclusive or of its
4837 The two arguments to the '``or``' instruction must be
4838 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4839 arguments must have identical types.
4844 The truth table used for the '``or``' instruction is:
4863 <result> = or i32 4, %var ; yields i32:result = 4 | %var
4864 <result> = or i32 15, 40 ; yields i32:result = 47
4865 <result> = or i32 4, 8 ; yields i32:result = 12
4867 '``xor``' Instruction
4868 ^^^^^^^^^^^^^^^^^^^^^
4875 <result> = xor <ty> <op1>, <op2> ; yields ty:result
4880 The '``xor``' instruction returns the bitwise logical exclusive or of
4881 its two operands. The ``xor`` is used to implement the "one's
4882 complement" operation, which is the "~" operator in C.
4887 The two arguments to the '``xor``' instruction must be
4888 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4889 arguments must have identical types.
4894 The truth table used for the '``xor``' instruction is:
4911 .. code-block:: llvm
4913 <result> = xor i32 4, %var ; yields i32:result = 4 ^ %var
4914 <result> = xor i32 15, 40 ; yields i32:result = 39
4915 <result> = xor i32 4, 8 ; yields i32:result = 12
4916 <result> = xor i32 %V, -1 ; yields i32:result = ~%V
4921 LLVM supports several instructions to represent vector operations in a
4922 target-independent manner. These instructions cover the element-access
4923 and vector-specific operations needed to process vectors effectively.
4924 While LLVM does directly support these vector operations, many
4925 sophisticated algorithms will want to use target-specific intrinsics to
4926 take full advantage of a specific target.
4928 .. _i_extractelement:
4930 '``extractelement``' Instruction
4931 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4938 <result> = extractelement <n x <ty>> <val>, <ty2> <idx> ; yields <ty>
4943 The '``extractelement``' instruction extracts a single scalar element
4944 from a vector at a specified index.
4949 The first operand of an '``extractelement``' instruction is a value of
4950 :ref:`vector <t_vector>` type. The second operand is an index indicating
4951 the position from which to extract the element. The index may be a
4952 variable of any integer type.
4957 The result is a scalar of the same type as the element type of ``val``.
4958 Its value is the value at position ``idx`` of ``val``. If ``idx``
4959 exceeds the length of ``val``, the results are undefined.
4964 .. code-block:: llvm
4966 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
4968 .. _i_insertelement:
4970 '``insertelement``' Instruction
4971 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4978 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, <ty2> <idx> ; yields <n x <ty>>
4983 The '``insertelement``' instruction inserts a scalar element into a
4984 vector at a specified index.
4989 The first operand of an '``insertelement``' instruction is a value of
4990 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
4991 type must equal the element type of the first operand. The third operand
4992 is an index indicating the position at which to insert the value. The
4993 index may be a variable of any integer type.
4998 The result is a vector of the same type as ``val``. Its element values
4999 are those of ``val`` except at position ``idx``, where it gets the value
5000 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
5006 .. code-block:: llvm
5008 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
5010 .. _i_shufflevector:
5012 '``shufflevector``' Instruction
5013 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5020 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
5025 The '``shufflevector``' instruction constructs a permutation of elements
5026 from two input vectors, returning a vector with the same element type as
5027 the input and length that is the same as the shuffle mask.
5032 The first two operands of a '``shufflevector``' instruction are vectors
5033 with the same type. The third argument is a shuffle mask whose element
5034 type is always 'i32'. The result of the instruction is a vector whose
5035 length is the same as the shuffle mask and whose element type is the
5036 same as the element type of the first two operands.
5038 The shuffle mask operand is required to be a constant vector with either
5039 constant integer or undef values.
5044 The elements of the two input vectors are numbered from left to right
5045 across both of the vectors. The shuffle mask operand specifies, for each
5046 element of the result vector, which element of the two input vectors the
5047 result element gets. The element selector may be undef (meaning "don't
5048 care") and the second operand may be undef if performing a shuffle from
5054 .. code-block:: llvm
5056 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
5057 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
5058 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
5059 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
5060 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
5061 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
5062 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
5063 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
5065 Aggregate Operations
5066 --------------------
5068 LLVM supports several instructions for working with
5069 :ref:`aggregate <t_aggregate>` values.
5073 '``extractvalue``' Instruction
5074 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5081 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
5086 The '``extractvalue``' instruction extracts the value of a member field
5087 from an :ref:`aggregate <t_aggregate>` value.
5092 The first operand of an '``extractvalue``' instruction is a value of
5093 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
5094 constant indices to specify which value to extract in a similar manner
5095 as indices in a '``getelementptr``' instruction.
5097 The major differences to ``getelementptr`` indexing are:
5099 - Since the value being indexed is not a pointer, the first index is
5100 omitted and assumed to be zero.
5101 - At least one index must be specified.
5102 - Not only struct indices but also array indices must be in bounds.
5107 The result is the value at the position in the aggregate specified by
5113 .. code-block:: llvm
5115 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
5119 '``insertvalue``' Instruction
5120 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5127 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
5132 The '``insertvalue``' instruction inserts a value into a member field in
5133 an :ref:`aggregate <t_aggregate>` value.
5138 The first operand of an '``insertvalue``' instruction is a value of
5139 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
5140 a first-class value to insert. The following operands are constant
5141 indices indicating the position at which to insert the value in a
5142 similar manner as indices in a '``extractvalue``' instruction. The value
5143 to insert must have the same type as the value identified by the
5149 The result is an aggregate of the same type as ``val``. Its value is
5150 that of ``val`` except that the value at the position specified by the
5151 indices is that of ``elt``.
5156 .. code-block:: llvm
5158 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
5159 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
5160 %agg3 = insertvalue {i32, {float}} undef, float %val, 1, 0 ; yields {i32 undef, {float %val}}
5164 Memory Access and Addressing Operations
5165 ---------------------------------------
5167 A key design point of an SSA-based representation is how it represents
5168 memory. In LLVM, no memory locations are in SSA form, which makes things
5169 very simple. This section describes how to read, write, and allocate
5174 '``alloca``' Instruction
5175 ^^^^^^^^^^^^^^^^^^^^^^^^
5182 <result> = alloca [inalloca] <type> [, <ty> <NumElements>] [, align <alignment>] ; yields type*:result
5187 The '``alloca``' instruction allocates memory on the stack frame of the
5188 currently executing function, to be automatically released when this
5189 function returns to its caller. The object is always allocated in the
5190 generic address space (address space zero).
5195 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
5196 bytes of memory on the runtime stack, returning a pointer of the
5197 appropriate type to the program. If "NumElements" is specified, it is
5198 the number of elements allocated, otherwise "NumElements" is defaulted
5199 to be one. If a constant alignment is specified, the value result of the
5200 allocation is guaranteed to be aligned to at least that boundary. The
5201 alignment may not be greater than ``1 << 29``. If not specified, or if
5202 zero, the target can choose to align the allocation on any convenient
5203 boundary compatible with the type.
5205 '``type``' may be any sized type.
5210 Memory is allocated; a pointer is returned. The operation is undefined
5211 if there is insufficient stack space for the allocation. '``alloca``'d
5212 memory is automatically released when the function returns. The
5213 '``alloca``' instruction is commonly used to represent automatic
5214 variables that must have an address available. When the function returns
5215 (either with the ``ret`` or ``resume`` instructions), the memory is
5216 reclaimed. Allocating zero bytes is legal, but the result is undefined.
5217 The order in which memory is allocated (ie., which way the stack grows)
5223 .. code-block:: llvm
5225 %ptr = alloca i32 ; yields i32*:ptr
5226 %ptr = alloca i32, i32 4 ; yields i32*:ptr
5227 %ptr = alloca i32, i32 4, align 1024 ; yields i32*:ptr
5228 %ptr = alloca i32, align 1024 ; yields i32*:ptr
5232 '``load``' Instruction
5233 ^^^^^^^^^^^^^^^^^^^^^^
5240 <result> = load [volatile] <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>][, !nonnull !<index>]
5241 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
5242 !<index> = !{ i32 1 }
5247 The '``load``' instruction is used to read from memory.
5252 The argument to the ``load`` instruction specifies the memory address
5253 from which to load. The pointer must point to a :ref:`first
5254 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
5255 then the optimizer is not allowed to modify the number or order of
5256 execution of this ``load`` with other :ref:`volatile
5257 operations <volatile>`.
5259 If the ``load`` is marked as ``atomic``, it takes an extra
5260 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5261 ``release`` and ``acq_rel`` orderings are not valid on ``load``
5262 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5263 when they may see multiple atomic stores. The type of the pointee must
5264 be an integer type whose bit width is a power of two greater than or
5265 equal to eight and less than or equal to a target-specific size limit.
5266 ``align`` must be explicitly specified on atomic loads, and the load has
5267 undefined behavior if the alignment is not set to a value which is at
5268 least the size in bytes of the pointee. ``!nontemporal`` does not have
5269 any defined semantics for atomic loads.
5271 The optional constant ``align`` argument specifies the alignment of the
5272 operation (that is, the alignment of the memory address). A value of 0
5273 or an omitted ``align`` argument means that the operation has the ABI
5274 alignment for the target. It is the responsibility of the code emitter
5275 to ensure that the alignment information is correct. Overestimating the
5276 alignment results in undefined behavior. Underestimating the alignment
5277 may produce less efficient code. An alignment of 1 is always safe. The
5278 maximum possible alignment is ``1 << 29``.
5280 The optional ``!nontemporal`` metadata must reference a single
5281 metadata name ``<index>`` corresponding to a metadata node with one
5282 ``i32`` entry of value 1. The existence of the ``!nontemporal``
5283 metadata on the instruction tells the optimizer and code generator
5284 that this load is not expected to be reused in the cache. The code
5285 generator may select special instructions to save cache bandwidth, such
5286 as the ``MOVNT`` instruction on x86.
5288 The optional ``!invariant.load`` metadata must reference a single
5289 metadata name ``<index>`` corresponding to a metadata node with no
5290 entries. The existence of the ``!invariant.load`` metadata on the
5291 instruction tells the optimizer and code generator that the address
5292 operand to this load points to memory which can be assumed unchanged.
5293 Being invariant does not imply that a location is dereferenceable,
5294 but it does imply that once the location is known dereferenceable
5295 its value is henceforth unchanging.
5297 The optional ``!nonnull`` metadata must reference a single
5298 metadata name ``<index>`` corresponding to a metadata node with no
5299 entries. The existence of the ``!nonnull`` metadata on the
5300 instruction tells the optimizer that the value loaded is known to
5301 never be null. This is analogous to the ''nonnull'' attribute
5302 on parameters and return values. This metadata can only be applied
5303 to loads of a pointer type.
5308 The location of memory pointed to is loaded. If the value being loaded
5309 is of scalar type then the number of bytes read does not exceed the
5310 minimum number of bytes needed to hold all bits of the type. For
5311 example, loading an ``i24`` reads at most three bytes. When loading a
5312 value of a type like ``i20`` with a size that is not an integral number
5313 of bytes, the result is undefined if the value was not originally
5314 written using a store of the same type.
5319 .. code-block:: llvm
5321 %ptr = alloca i32 ; yields i32*:ptr
5322 store i32 3, i32* %ptr ; yields void
5323 %val = load i32* %ptr ; yields i32:val = i32 3
5327 '``store``' Instruction
5328 ^^^^^^^^^^^^^^^^^^^^^^^
5335 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields void
5336 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields void
5341 The '``store``' instruction is used to write to memory.
5346 There are two arguments to the ``store`` instruction: a value to store
5347 and an address at which to store it. The type of the ``<pointer>``
5348 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
5349 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
5350 then the optimizer is not allowed to modify the number or order of
5351 execution of this ``store`` with other :ref:`volatile
5352 operations <volatile>`.
5354 If the ``store`` is marked as ``atomic``, it takes an extra
5355 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5356 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
5357 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5358 when they may see multiple atomic stores. The type of the pointee must
5359 be an integer type whose bit width is a power of two greater than or
5360 equal to eight and less than or equal to a target-specific size limit.
5361 ``align`` must be explicitly specified on atomic stores, and the store
5362 has undefined behavior if the alignment is not set to a value which is
5363 at least the size in bytes of the pointee. ``!nontemporal`` does not
5364 have any defined semantics for atomic stores.
5366 The optional constant ``align`` argument specifies the alignment of the
5367 operation (that is, the alignment of the memory address). A value of 0
5368 or an omitted ``align`` argument means that the operation has the ABI
5369 alignment for the target. It is the responsibility of the code emitter
5370 to ensure that the alignment information is correct. Overestimating the
5371 alignment results in undefined behavior. Underestimating the
5372 alignment may produce less efficient code. An alignment of 1 is always
5373 safe. The maximum possible alignment is ``1 << 29``.
5375 The optional ``!nontemporal`` metadata must reference a single metadata
5376 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
5377 value 1. The existence of the ``!nontemporal`` metadata on the instruction
5378 tells the optimizer and code generator that this load is not expected to
5379 be reused in the cache. The code generator may select special
5380 instructions to save cache bandwidth, such as the MOVNT instruction on
5386 The contents of memory are updated to contain ``<value>`` at the
5387 location specified by the ``<pointer>`` operand. If ``<value>`` is
5388 of scalar type then the number of bytes written does not exceed the
5389 minimum number of bytes needed to hold all bits of the type. For
5390 example, storing an ``i24`` writes at most three bytes. When writing a
5391 value of a type like ``i20`` with a size that is not an integral number
5392 of bytes, it is unspecified what happens to the extra bits that do not
5393 belong to the type, but they will typically be overwritten.
5398 .. code-block:: llvm
5400 %ptr = alloca i32 ; yields i32*:ptr
5401 store i32 3, i32* %ptr ; yields void
5402 %val = load i32* %ptr ; yields i32:val = i32 3
5406 '``fence``' Instruction
5407 ^^^^^^^^^^^^^^^^^^^^^^^
5414 fence [singlethread] <ordering> ; yields void
5419 The '``fence``' instruction is used to introduce happens-before edges
5425 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
5426 defines what *synchronizes-with* edges they add. They can only be given
5427 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
5432 A fence A which has (at least) ``release`` ordering semantics
5433 *synchronizes with* a fence B with (at least) ``acquire`` ordering
5434 semantics if and only if there exist atomic operations X and Y, both
5435 operating on some atomic object M, such that A is sequenced before X, X
5436 modifies M (either directly or through some side effect of a sequence
5437 headed by X), Y is sequenced before B, and Y observes M. This provides a
5438 *happens-before* dependency between A and B. Rather than an explicit
5439 ``fence``, one (but not both) of the atomic operations X or Y might
5440 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
5441 still *synchronize-with* the explicit ``fence`` and establish the
5442 *happens-before* edge.
5444 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
5445 ``acquire`` and ``release`` semantics specified above, participates in
5446 the global program order of other ``seq_cst`` operations and/or fences.
5448 The optional ":ref:`singlethread <singlethread>`" argument specifies
5449 that the fence only synchronizes with other fences in the same thread.
5450 (This is useful for interacting with signal handlers.)
5455 .. code-block:: llvm
5457 fence acquire ; yields void
5458 fence singlethread seq_cst ; yields void
5462 '``cmpxchg``' Instruction
5463 ^^^^^^^^^^^^^^^^^^^^^^^^^
5470 cmpxchg [weak] [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <success ordering> <failure ordering> ; yields { ty, i1 }
5475 The '``cmpxchg``' instruction is used to atomically modify memory. It
5476 loads a value in memory and compares it to a given value. If they are
5477 equal, it tries to store a new value into the memory.
5482 There are three arguments to the '``cmpxchg``' instruction: an address
5483 to operate on, a value to compare to the value currently be at that
5484 address, and a new value to place at that address if the compared values
5485 are equal. The type of '<cmp>' must be an integer type whose bit width
5486 is a power of two greater than or equal to eight and less than or equal
5487 to a target-specific size limit. '<cmp>' and '<new>' must have the same
5488 type, and the type of '<pointer>' must be a pointer to that type. If the
5489 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
5490 to modify the number or order of execution of this ``cmpxchg`` with
5491 other :ref:`volatile operations <volatile>`.
5493 The success and failure :ref:`ordering <ordering>` arguments specify how this
5494 ``cmpxchg`` synchronizes with other atomic operations. Both ordering parameters
5495 must be at least ``monotonic``, the ordering constraint on failure must be no
5496 stronger than that on success, and the failure ordering cannot be either
5497 ``release`` or ``acq_rel``.
5499 The optional "``singlethread``" argument declares that the ``cmpxchg``
5500 is only atomic with respect to code (usually signal handlers) running in
5501 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
5502 respect to all other code in the system.
5504 The pointer passed into cmpxchg must have alignment greater than or
5505 equal to the size in memory of the operand.
5510 The contents of memory at the location specified by the '``<pointer>``' operand
5511 is read and compared to '``<cmp>``'; if the read value is the equal, the
5512 '``<new>``' is written. The original value at the location is returned, together
5513 with a flag indicating success (true) or failure (false).
5515 If the cmpxchg operation is marked as ``weak`` then a spurious failure is
5516 permitted: the operation may not write ``<new>`` even if the comparison
5519 If the cmpxchg operation is strong (the default), the i1 value is 1 if and only
5520 if the value loaded equals ``cmp``.
5522 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose of
5523 identifying release sequences. A failed ``cmpxchg`` is equivalent to an atomic
5524 load with an ordering parameter determined the second ordering parameter.
5529 .. code-block:: llvm
5532 %orig = atomic load i32* %ptr unordered ; yields i32
5536 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
5537 %squared = mul i32 %cmp, %cmp
5538 %val_success = cmpxchg i32* %ptr, i32 %cmp, i32 %squared acq_rel monotonic ; yields { i32, i1 }
5539 %value_loaded = extractvalue { i32, i1 } %val_success, 0
5540 %success = extractvalue { i32, i1 } %val_success, 1
5541 br i1 %success, label %done, label %loop
5548 '``atomicrmw``' Instruction
5549 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
5556 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields ty
5561 The '``atomicrmw``' instruction is used to atomically modify memory.
5566 There are three arguments to the '``atomicrmw``' instruction: an
5567 operation to apply, an address whose value to modify, an argument to the
5568 operation. The operation must be one of the following keywords:
5582 The type of '<value>' must be an integer type whose bit width is a power
5583 of two greater than or equal to eight and less than or equal to a
5584 target-specific size limit. The type of the '``<pointer>``' operand must
5585 be a pointer to that type. If the ``atomicrmw`` is marked as
5586 ``volatile``, then the optimizer is not allowed to modify the number or
5587 order of execution of this ``atomicrmw`` with other :ref:`volatile
5588 operations <volatile>`.
5593 The contents of memory at the location specified by the '``<pointer>``'
5594 operand are atomically read, modified, and written back. The original
5595 value at the location is returned. The modification is specified by the
5598 - xchg: ``*ptr = val``
5599 - add: ``*ptr = *ptr + val``
5600 - sub: ``*ptr = *ptr - val``
5601 - and: ``*ptr = *ptr & val``
5602 - nand: ``*ptr = ~(*ptr & val)``
5603 - or: ``*ptr = *ptr | val``
5604 - xor: ``*ptr = *ptr ^ val``
5605 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
5606 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
5607 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
5609 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
5615 .. code-block:: llvm
5617 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields i32
5619 .. _i_getelementptr:
5621 '``getelementptr``' Instruction
5622 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5629 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
5630 <result> = getelementptr inbounds <pty>* <ptrval>{, <ty> <idx>}*
5631 <result> = getelementptr <ptr vector> ptrval, <vector index type> idx
5636 The '``getelementptr``' instruction is used to get the address of a
5637 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
5638 address calculation only and does not access memory.
5643 The first argument is always a pointer or a vector of pointers, and
5644 forms the basis of the calculation. The remaining arguments are indices
5645 that indicate which of the elements of the aggregate object are indexed.
5646 The interpretation of each index is dependent on the type being indexed
5647 into. The first index always indexes the pointer value given as the
5648 first argument, the second index indexes a value of the type pointed to
5649 (not necessarily the value directly pointed to, since the first index
5650 can be non-zero), etc. The first type indexed into must be a pointer
5651 value, subsequent types can be arrays, vectors, and structs. Note that
5652 subsequent types being indexed into can never be pointers, since that
5653 would require loading the pointer before continuing calculation.
5655 The type of each index argument depends on the type it is indexing into.
5656 When indexing into a (optionally packed) structure, only ``i32`` integer
5657 **constants** are allowed (when using a vector of indices they must all
5658 be the **same** ``i32`` integer constant). When indexing into an array,
5659 pointer or vector, integers of any width are allowed, and they are not
5660 required to be constant. These integers are treated as signed values
5663 For example, let's consider a C code fragment and how it gets compiled
5679 int *foo(struct ST *s) {
5680 return &s[1].Z.B[5][13];
5683 The LLVM code generated by Clang is:
5685 .. code-block:: llvm
5687 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
5688 %struct.ST = type { i32, double, %struct.RT }
5690 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
5692 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5699 In the example above, the first index is indexing into the
5700 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
5701 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
5702 indexes into the third element of the structure, yielding a
5703 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
5704 structure. The third index indexes into the second element of the
5705 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
5706 dimensions of the array are subscripted into, yielding an '``i32``'
5707 type. The '``getelementptr``' instruction returns a pointer to this
5708 element, thus computing a value of '``i32*``' type.
5710 Note that it is perfectly legal to index partially through a structure,
5711 returning a pointer to an inner element. Because of this, the LLVM code
5712 for the given testcase is equivalent to:
5714 .. code-block:: llvm
5716 define i32* @foo(%struct.ST* %s) {
5717 %t1 = getelementptr %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
5718 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
5719 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
5720 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
5721 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
5725 If the ``inbounds`` keyword is present, the result value of the
5726 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
5727 pointer is not an *in bounds* address of an allocated object, or if any
5728 of the addresses that would be formed by successive addition of the
5729 offsets implied by the indices to the base address with infinitely
5730 precise signed arithmetic are not an *in bounds* address of that
5731 allocated object. The *in bounds* addresses for an allocated object are
5732 all the addresses that point into the object, plus the address one byte
5733 past the end. In cases where the base is a vector of pointers the
5734 ``inbounds`` keyword applies to each of the computations element-wise.
5736 If the ``inbounds`` keyword is not present, the offsets are added to the
5737 base address with silently-wrapping two's complement arithmetic. If the
5738 offsets have a different width from the pointer, they are sign-extended
5739 or truncated to the width of the pointer. The result value of the
5740 ``getelementptr`` may be outside the object pointed to by the base
5741 pointer. The result value may not necessarily be used to access memory
5742 though, even if it happens to point into allocated storage. See the
5743 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
5746 The getelementptr instruction is often confusing. For some more insight
5747 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
5752 .. code-block:: llvm
5754 ; yields [12 x i8]*:aptr
5755 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5757 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
5759 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
5761 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
5763 In cases where the pointer argument is a vector of pointers, each index
5764 must be a vector with the same number of elements. For example:
5766 .. code-block:: llvm
5768 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5770 Conversion Operations
5771 ---------------------
5773 The instructions in this category are the conversion instructions
5774 (casting) which all take a single operand and a type. They perform
5775 various bit conversions on the operand.
5777 '``trunc .. to``' Instruction
5778 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5785 <result> = trunc <ty> <value> to <ty2> ; yields ty2
5790 The '``trunc``' instruction truncates its operand to the type ``ty2``.
5795 The '``trunc``' instruction takes a value to trunc, and a type to trunc
5796 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
5797 of the same number of integers. The bit size of the ``value`` must be
5798 larger than the bit size of the destination type, ``ty2``. Equal sized
5799 types are not allowed.
5804 The '``trunc``' instruction truncates the high order bits in ``value``
5805 and converts the remaining bits to ``ty2``. Since the source size must
5806 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
5807 It will always truncate bits.
5812 .. code-block:: llvm
5814 %X = trunc i32 257 to i8 ; yields i8:1
5815 %Y = trunc i32 123 to i1 ; yields i1:true
5816 %Z = trunc i32 122 to i1 ; yields i1:false
5817 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
5819 '``zext .. to``' Instruction
5820 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5827 <result> = zext <ty> <value> to <ty2> ; yields ty2
5832 The '``zext``' instruction zero extends its operand to type ``ty2``.
5837 The '``zext``' instruction takes a value to cast, and a type to cast it
5838 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5839 the same number of integers. The bit size of the ``value`` must be
5840 smaller than the bit size of the destination type, ``ty2``.
5845 The ``zext`` fills the high order bits of the ``value`` with zero bits
5846 until it reaches the size of the destination type, ``ty2``.
5848 When zero extending from i1, the result will always be either 0 or 1.
5853 .. code-block:: llvm
5855 %X = zext i32 257 to i64 ; yields i64:257
5856 %Y = zext i1 true to i32 ; yields i32:1
5857 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5859 '``sext .. to``' Instruction
5860 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5867 <result> = sext <ty> <value> to <ty2> ; yields ty2
5872 The '``sext``' sign extends ``value`` to the type ``ty2``.
5877 The '``sext``' instruction takes a value to cast, and a type to cast it
5878 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5879 the same number of integers. The bit size of the ``value`` must be
5880 smaller than the bit size of the destination type, ``ty2``.
5885 The '``sext``' instruction performs a sign extension by copying the sign
5886 bit (highest order bit) of the ``value`` until it reaches the bit size
5887 of the type ``ty2``.
5889 When sign extending from i1, the extension always results in -1 or 0.
5894 .. code-block:: llvm
5896 %X = sext i8 -1 to i16 ; yields i16 :65535
5897 %Y = sext i1 true to i32 ; yields i32:-1
5898 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5900 '``fptrunc .. to``' Instruction
5901 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5908 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
5913 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
5918 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
5919 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
5920 The size of ``value`` must be larger than the size of ``ty2``. This
5921 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
5926 The '``fptrunc``' instruction truncates a ``value`` from a larger
5927 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
5928 point <t_floating>` type. If the value cannot fit within the
5929 destination type, ``ty2``, then the results are undefined.
5934 .. code-block:: llvm
5936 %X = fptrunc double 123.0 to float ; yields float:123.0
5937 %Y = fptrunc double 1.0E+300 to float ; yields undefined
5939 '``fpext .. to``' Instruction
5940 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5947 <result> = fpext <ty> <value> to <ty2> ; yields ty2
5952 The '``fpext``' extends a floating point ``value`` to a larger floating
5958 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
5959 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
5960 to. The source type must be smaller than the destination type.
5965 The '``fpext``' instruction extends the ``value`` from a smaller
5966 :ref:`floating point <t_floating>` type to a larger :ref:`floating
5967 point <t_floating>` type. The ``fpext`` cannot be used to make a
5968 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
5969 *no-op cast* for a floating point cast.
5974 .. code-block:: llvm
5976 %X = fpext float 3.125 to double ; yields double:3.125000e+00
5977 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
5979 '``fptoui .. to``' Instruction
5980 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5987 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
5992 The '``fptoui``' converts a floating point ``value`` to its unsigned
5993 integer equivalent of type ``ty2``.
5998 The '``fptoui``' instruction takes a value to cast, which must be a
5999 scalar or vector :ref:`floating point <t_floating>` value, and a type to
6000 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
6001 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
6002 type with the same number of elements as ``ty``
6007 The '``fptoui``' instruction converts its :ref:`floating
6008 point <t_floating>` operand into the nearest (rounding towards zero)
6009 unsigned integer value. If the value cannot fit in ``ty2``, the results
6015 .. code-block:: llvm
6017 %X = fptoui double 123.0 to i32 ; yields i32:123
6018 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
6019 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
6021 '``fptosi .. to``' Instruction
6022 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6029 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
6034 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
6035 ``value`` to type ``ty2``.
6040 The '``fptosi``' instruction takes a value to cast, which must be a
6041 scalar or vector :ref:`floating point <t_floating>` value, and a type to
6042 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
6043 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
6044 type with the same number of elements as ``ty``
6049 The '``fptosi``' instruction converts its :ref:`floating
6050 point <t_floating>` operand into the nearest (rounding towards zero)
6051 signed integer value. If the value cannot fit in ``ty2``, the results
6057 .. code-block:: llvm
6059 %X = fptosi double -123.0 to i32 ; yields i32:-123
6060 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
6061 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
6063 '``uitofp .. to``' Instruction
6064 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6071 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
6076 The '``uitofp``' instruction regards ``value`` as an unsigned integer
6077 and converts that value to the ``ty2`` type.
6082 The '``uitofp``' instruction takes a value to cast, which must be a
6083 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
6084 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
6085 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
6086 type with the same number of elements as ``ty``
6091 The '``uitofp``' instruction interprets its operand as an unsigned
6092 integer quantity and converts it to the corresponding floating point
6093 value. If the value cannot fit in the floating point value, the results
6099 .. code-block:: llvm
6101 %X = uitofp i32 257 to float ; yields float:257.0
6102 %Y = uitofp i8 -1 to double ; yields double:255.0
6104 '``sitofp .. to``' Instruction
6105 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6112 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
6117 The '``sitofp``' instruction regards ``value`` as a signed integer and
6118 converts that value to the ``ty2`` type.
6123 The '``sitofp``' instruction takes a value to cast, which must be a
6124 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
6125 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
6126 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
6127 type with the same number of elements as ``ty``
6132 The '``sitofp``' instruction interprets its operand as a signed integer
6133 quantity and converts it to the corresponding floating point value. If
6134 the value cannot fit in the floating point value, the results are
6140 .. code-block:: llvm
6142 %X = sitofp i32 257 to float ; yields float:257.0
6143 %Y = sitofp i8 -1 to double ; yields double:-1.0
6147 '``ptrtoint .. to``' Instruction
6148 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6155 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
6160 The '``ptrtoint``' instruction converts the pointer or a vector of
6161 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
6166 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
6167 a a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
6168 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
6169 a vector of integers type.
6174 The '``ptrtoint``' instruction converts ``value`` to integer type
6175 ``ty2`` by interpreting the pointer value as an integer and either
6176 truncating or zero extending that value to the size of the integer type.
6177 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
6178 ``value`` is larger than ``ty2`` then a truncation is done. If they are
6179 the same size, then nothing is done (*no-op cast*) other than a type
6185 .. code-block:: llvm
6187 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
6188 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
6189 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
6193 '``inttoptr .. to``' Instruction
6194 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6201 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
6206 The '``inttoptr``' instruction converts an integer ``value`` to a
6207 pointer type, ``ty2``.
6212 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
6213 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
6219 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
6220 applying either a zero extension or a truncation depending on the size
6221 of the integer ``value``. If ``value`` is larger than the size of a
6222 pointer then a truncation is done. If ``value`` is smaller than the size
6223 of a pointer then a zero extension is done. If they are the same size,
6224 nothing is done (*no-op cast*).
6229 .. code-block:: llvm
6231 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
6232 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
6233 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
6234 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
6238 '``bitcast .. to``' Instruction
6239 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6246 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
6251 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
6257 The '``bitcast``' instruction takes a value to cast, which must be a
6258 non-aggregate first class value, and a type to cast it to, which must
6259 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
6260 bit sizes of ``value`` and the destination type, ``ty2``, must be
6261 identical. If the source type is a pointer, the destination type must
6262 also be a pointer of the same size. This instruction supports bitwise
6263 conversion of vectors to integers and to vectors of other types (as
6264 long as they have the same size).
6269 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
6270 is always a *no-op cast* because no bits change with this
6271 conversion. The conversion is done as if the ``value`` had been stored
6272 to memory and read back as type ``ty2``. Pointer (or vector of
6273 pointers) types may only be converted to other pointer (or vector of
6274 pointers) types with the same address space through this instruction.
6275 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
6276 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
6281 .. code-block:: llvm
6283 %X = bitcast i8 255 to i8 ; yields i8 :-1
6284 %Y = bitcast i32* %x to sint* ; yields sint*:%x
6285 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
6286 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
6288 .. _i_addrspacecast:
6290 '``addrspacecast .. to``' Instruction
6291 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6298 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
6303 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
6304 address space ``n`` to type ``pty2`` in address space ``m``.
6309 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
6310 to cast and a pointer type to cast it to, which must have a different
6316 The '``addrspacecast``' instruction converts the pointer value
6317 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
6318 value modification, depending on the target and the address space
6319 pair. Pointer conversions within the same address space must be
6320 performed with the ``bitcast`` instruction. Note that if the address space
6321 conversion is legal then both result and operand refer to the same memory
6327 .. code-block:: llvm
6329 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
6330 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
6331 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
6338 The instructions in this category are the "miscellaneous" instructions,
6339 which defy better classification.
6343 '``icmp``' Instruction
6344 ^^^^^^^^^^^^^^^^^^^^^^
6351 <result> = icmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6356 The '``icmp``' instruction returns a boolean value or a vector of
6357 boolean values based on comparison of its two integer, integer vector,
6358 pointer, or pointer vector operands.
6363 The '``icmp``' instruction takes three operands. The first operand is
6364 the condition code indicating the kind of comparison to perform. It is
6365 not a value, just a keyword. The possible condition code are:
6368 #. ``ne``: not equal
6369 #. ``ugt``: unsigned greater than
6370 #. ``uge``: unsigned greater or equal
6371 #. ``ult``: unsigned less than
6372 #. ``ule``: unsigned less or equal
6373 #. ``sgt``: signed greater than
6374 #. ``sge``: signed greater or equal
6375 #. ``slt``: signed less than
6376 #. ``sle``: signed less or equal
6378 The remaining two arguments must be :ref:`integer <t_integer>` or
6379 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
6380 must also be identical types.
6385 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
6386 code given as ``cond``. The comparison performed always yields either an
6387 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
6389 #. ``eq``: yields ``true`` if the operands are equal, ``false``
6390 otherwise. No sign interpretation is necessary or performed.
6391 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
6392 otherwise. No sign interpretation is necessary or performed.
6393 #. ``ugt``: interprets the operands as unsigned values and yields
6394 ``true`` if ``op1`` is greater than ``op2``.
6395 #. ``uge``: interprets the operands as unsigned values and yields
6396 ``true`` if ``op1`` is greater than or equal to ``op2``.
6397 #. ``ult``: interprets the operands as unsigned values and yields
6398 ``true`` if ``op1`` is less than ``op2``.
6399 #. ``ule``: interprets the operands as unsigned values and yields
6400 ``true`` if ``op1`` is less than or equal to ``op2``.
6401 #. ``sgt``: interprets the operands as signed values and yields ``true``
6402 if ``op1`` is greater than ``op2``.
6403 #. ``sge``: interprets the operands as signed values and yields ``true``
6404 if ``op1`` is greater than or equal to ``op2``.
6405 #. ``slt``: interprets the operands as signed values and yields ``true``
6406 if ``op1`` is less than ``op2``.
6407 #. ``sle``: interprets the operands as signed values and yields ``true``
6408 if ``op1`` is less than or equal to ``op2``.
6410 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
6411 are compared as if they were integers.
6413 If the operands are integer vectors, then they are compared element by
6414 element. The result is an ``i1`` vector with the same number of elements
6415 as the values being compared. Otherwise, the result is an ``i1``.
6420 .. code-block:: llvm
6422 <result> = icmp eq i32 4, 5 ; yields: result=false
6423 <result> = icmp ne float* %X, %X ; yields: result=false
6424 <result> = icmp ult i16 4, 5 ; yields: result=true
6425 <result> = icmp sgt i16 4, 5 ; yields: result=false
6426 <result> = icmp ule i16 -4, 5 ; yields: result=false
6427 <result> = icmp sge i16 4, 5 ; yields: result=false
6429 Note that the code generator does not yet support vector types with the
6430 ``icmp`` instruction.
6434 '``fcmp``' Instruction
6435 ^^^^^^^^^^^^^^^^^^^^^^
6442 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6447 The '``fcmp``' instruction returns a boolean value or vector of boolean
6448 values based on comparison of its operands.
6450 If the operands are floating point scalars, then the result type is a
6451 boolean (:ref:`i1 <t_integer>`).
6453 If the operands are floating point vectors, then the result type is a
6454 vector of boolean with the same number of elements as the operands being
6460 The '``fcmp``' instruction takes three operands. The first operand is
6461 the condition code indicating the kind of comparison to perform. It is
6462 not a value, just a keyword. The possible condition code are:
6464 #. ``false``: no comparison, always returns false
6465 #. ``oeq``: ordered and equal
6466 #. ``ogt``: ordered and greater than
6467 #. ``oge``: ordered and greater than or equal
6468 #. ``olt``: ordered and less than
6469 #. ``ole``: ordered and less than or equal
6470 #. ``one``: ordered and not equal
6471 #. ``ord``: ordered (no nans)
6472 #. ``ueq``: unordered or equal
6473 #. ``ugt``: unordered or greater than
6474 #. ``uge``: unordered or greater than or equal
6475 #. ``ult``: unordered or less than
6476 #. ``ule``: unordered or less than or equal
6477 #. ``une``: unordered or not equal
6478 #. ``uno``: unordered (either nans)
6479 #. ``true``: no comparison, always returns true
6481 *Ordered* means that neither operand is a QNAN while *unordered* means
6482 that either operand may be a QNAN.
6484 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
6485 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
6486 type. They must have identical types.
6491 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
6492 condition code given as ``cond``. If the operands are vectors, then the
6493 vectors are compared element by element. Each comparison performed
6494 always yields an :ref:`i1 <t_integer>` result, as follows:
6496 #. ``false``: always yields ``false``, regardless of operands.
6497 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
6498 is equal to ``op2``.
6499 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
6500 is greater than ``op2``.
6501 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
6502 is greater than or equal to ``op2``.
6503 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
6504 is less than ``op2``.
6505 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
6506 is less than or equal to ``op2``.
6507 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
6508 is not equal to ``op2``.
6509 #. ``ord``: yields ``true`` if both operands are not a QNAN.
6510 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
6512 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
6513 greater than ``op2``.
6514 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
6515 greater than or equal to ``op2``.
6516 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
6518 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
6519 less than or equal to ``op2``.
6520 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
6521 not equal to ``op2``.
6522 #. ``uno``: yields ``true`` if either operand is a QNAN.
6523 #. ``true``: always yields ``true``, regardless of operands.
6528 .. code-block:: llvm
6530 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
6531 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
6532 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
6533 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
6535 Note that the code generator does not yet support vector types with the
6536 ``fcmp`` instruction.
6540 '``phi``' Instruction
6541 ^^^^^^^^^^^^^^^^^^^^^
6548 <result> = phi <ty> [ <val0>, <label0>], ...
6553 The '``phi``' instruction is used to implement the φ node in the SSA
6554 graph representing the function.
6559 The type of the incoming values is specified with the first type field.
6560 After this, the '``phi``' instruction takes a list of pairs as
6561 arguments, with one pair for each predecessor basic block of the current
6562 block. Only values of :ref:`first class <t_firstclass>` type may be used as
6563 the value arguments to the PHI node. Only labels may be used as the
6566 There must be no non-phi instructions between the start of a basic block
6567 and the PHI instructions: i.e. PHI instructions must be first in a basic
6570 For the purposes of the SSA form, the use of each incoming value is
6571 deemed to occur on the edge from the corresponding predecessor block to
6572 the current block (but after any definition of an '``invoke``'
6573 instruction's return value on the same edge).
6578 At runtime, the '``phi``' instruction logically takes on the value
6579 specified by the pair corresponding to the predecessor basic block that
6580 executed just prior to the current block.
6585 .. code-block:: llvm
6587 Loop: ; Infinite loop that counts from 0 on up...
6588 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
6589 %nextindvar = add i32 %indvar, 1
6594 '``select``' Instruction
6595 ^^^^^^^^^^^^^^^^^^^^^^^^
6602 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
6604 selty is either i1 or {<N x i1>}
6609 The '``select``' instruction is used to choose one value based on a
6610 condition, without IR-level branching.
6615 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
6616 values indicating the condition, and two values of the same :ref:`first
6617 class <t_firstclass>` type. If the val1/val2 are vectors and the
6618 condition is a scalar, then entire vectors are selected, not individual
6624 If the condition is an i1 and it evaluates to 1, the instruction returns
6625 the first value argument; otherwise, it returns the second value
6628 If the condition is a vector of i1, then the value arguments must be
6629 vectors of the same size, and the selection is done element by element.
6634 .. code-block:: llvm
6636 %X = select i1 true, i8 17, i8 42 ; yields i8:17
6640 '``call``' Instruction
6641 ^^^^^^^^^^^^^^^^^^^^^^
6648 <result> = [tail | musttail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
6653 The '``call``' instruction represents a simple function call.
6658 This instruction requires several arguments:
6660 #. The optional ``tail`` and ``musttail`` markers indicate that the optimizers
6661 should perform tail call optimization. The ``tail`` marker is a hint that
6662 `can be ignored <CodeGenerator.html#sibcallopt>`_. The ``musttail`` marker
6663 means that the call must be tail call optimized in order for the program to
6664 be correct. The ``musttail`` marker provides these guarantees:
6666 #. The call will not cause unbounded stack growth if it is part of a
6667 recursive cycle in the call graph.
6668 #. Arguments with the :ref:`inalloca <attr_inalloca>` attribute are
6671 Both markers imply that the callee does not access allocas or varargs from
6672 the caller. Calls marked ``musttail`` must obey the following additional
6675 - The call must immediately precede a :ref:`ret <i_ret>` instruction,
6676 or a pointer bitcast followed by a ret instruction.
6677 - The ret instruction must return the (possibly bitcasted) value
6678 produced by the call or void.
6679 - The caller and callee prototypes must match. Pointer types of
6680 parameters or return types may differ in pointee type, but not
6682 - The calling conventions of the caller and callee must match.
6683 - All ABI-impacting function attributes, such as sret, byval, inreg,
6684 returned, and inalloca, must match.
6685 - The callee must be varargs iff the caller is varargs. Bitcasting a
6686 non-varargs function to the appropriate varargs type is legal so
6687 long as the non-varargs prefixes obey the other rules.
6689 Tail call optimization for calls marked ``tail`` is guaranteed to occur if
6690 the following conditions are met:
6692 - Caller and callee both have the calling convention ``fastcc``.
6693 - The call is in tail position (ret immediately follows call and ret
6694 uses value of call or is void).
6695 - Option ``-tailcallopt`` is enabled, or
6696 ``llvm::GuaranteedTailCallOpt`` is ``true``.
6697 - `Platform-specific constraints are
6698 met. <CodeGenerator.html#tailcallopt>`_
6700 #. The optional "cconv" marker indicates which :ref:`calling
6701 convention <callingconv>` the call should use. If none is
6702 specified, the call defaults to using C calling conventions. The
6703 calling convention of the call must match the calling convention of
6704 the target function, or else the behavior is undefined.
6705 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
6706 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
6708 #. '``ty``': the type of the call instruction itself which is also the
6709 type of the return value. Functions that return no value are marked
6711 #. '``fnty``': shall be the signature of the pointer to function value
6712 being invoked. The argument types must match the types implied by
6713 this signature. This type can be omitted if the function is not
6714 varargs and if the function type does not return a pointer to a
6716 #. '``fnptrval``': An LLVM value containing a pointer to a function to
6717 be invoked. In most cases, this is a direct function invocation, but
6718 indirect ``call``'s are just as possible, calling an arbitrary pointer
6720 #. '``function args``': argument list whose types match the function
6721 signature argument types and parameter attributes. All arguments must
6722 be of :ref:`first class <t_firstclass>` type. If the function signature
6723 indicates the function accepts a variable number of arguments, the
6724 extra arguments can be specified.
6725 #. The optional :ref:`function attributes <fnattrs>` list. Only
6726 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
6727 attributes are valid here.
6732 The '``call``' instruction is used to cause control flow to transfer to
6733 a specified function, with its incoming arguments bound to the specified
6734 values. Upon a '``ret``' instruction in the called function, control
6735 flow continues with the instruction after the function call, and the
6736 return value of the function is bound to the result argument.
6741 .. code-block:: llvm
6743 %retval = call i32 @test(i32 %argc)
6744 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
6745 %X = tail call i32 @foo() ; yields i32
6746 %Y = tail call fastcc i32 @foo() ; yields i32
6747 call void %foo(i8 97 signext)
6749 %struct.A = type { i32, i8 }
6750 %r = call %struct.A @foo() ; yields { i32, i8 }
6751 %gr = extractvalue %struct.A %r, 0 ; yields i32
6752 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
6753 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
6754 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
6756 llvm treats calls to some functions with names and arguments that match
6757 the standard C99 library as being the C99 library functions, and may
6758 perform optimizations or generate code for them under that assumption.
6759 This is something we'd like to change in the future to provide better
6760 support for freestanding environments and non-C-based languages.
6764 '``va_arg``' Instruction
6765 ^^^^^^^^^^^^^^^^^^^^^^^^
6772 <resultval> = va_arg <va_list*> <arglist>, <argty>
6777 The '``va_arg``' instruction is used to access arguments passed through
6778 the "variable argument" area of a function call. It is used to implement
6779 the ``va_arg`` macro in C.
6784 This instruction takes a ``va_list*`` value and the type of the
6785 argument. It returns a value of the specified argument type and
6786 increments the ``va_list`` to point to the next argument. The actual
6787 type of ``va_list`` is target specific.
6792 The '``va_arg``' instruction loads an argument of the specified type
6793 from the specified ``va_list`` and causes the ``va_list`` to point to
6794 the next argument. For more information, see the variable argument
6795 handling :ref:`Intrinsic Functions <int_varargs>`.
6797 It is legal for this instruction to be called in a function which does
6798 not take a variable number of arguments, for example, the ``vfprintf``
6801 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
6802 function <intrinsics>` because it takes a type as an argument.
6807 See the :ref:`variable argument processing <int_varargs>` section.
6809 Note that the code generator does not yet fully support va\_arg on many
6810 targets. Also, it does not currently support va\_arg with aggregate
6811 types on any target.
6815 '``landingpad``' Instruction
6816 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6823 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
6824 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
6826 <clause> := catch <type> <value>
6827 <clause> := filter <array constant type> <array constant>
6832 The '``landingpad``' instruction is used by `LLVM's exception handling
6833 system <ExceptionHandling.html#overview>`_ to specify that a basic block
6834 is a landing pad --- one where the exception lands, and corresponds to the
6835 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
6836 defines values supplied by the personality function (``pers_fn``) upon
6837 re-entry to the function. The ``resultval`` has the type ``resultty``.
6842 This instruction takes a ``pers_fn`` value. This is the personality
6843 function associated with the unwinding mechanism. The optional
6844 ``cleanup`` flag indicates that the landing pad block is a cleanup.
6846 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
6847 contains the global variable representing the "type" that may be caught
6848 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
6849 clause takes an array constant as its argument. Use
6850 "``[0 x i8**] undef``" for a filter which cannot throw. The
6851 '``landingpad``' instruction must contain *at least* one ``clause`` or
6852 the ``cleanup`` flag.
6857 The '``landingpad``' instruction defines the values which are set by the
6858 personality function (``pers_fn``) upon re-entry to the function, and
6859 therefore the "result type" of the ``landingpad`` instruction. As with
6860 calling conventions, how the personality function results are
6861 represented in LLVM IR is target specific.
6863 The clauses are applied in order from top to bottom. If two
6864 ``landingpad`` instructions are merged together through inlining, the
6865 clauses from the calling function are appended to the list of clauses.
6866 When the call stack is being unwound due to an exception being thrown,
6867 the exception is compared against each ``clause`` in turn. If it doesn't
6868 match any of the clauses, and the ``cleanup`` flag is not set, then
6869 unwinding continues further up the call stack.
6871 The ``landingpad`` instruction has several restrictions:
6873 - A landing pad block is a basic block which is the unwind destination
6874 of an '``invoke``' instruction.
6875 - A landing pad block must have a '``landingpad``' instruction as its
6876 first non-PHI instruction.
6877 - There can be only one '``landingpad``' instruction within the landing
6879 - A basic block that is not a landing pad block may not include a
6880 '``landingpad``' instruction.
6881 - All '``landingpad``' instructions in a function must have the same
6882 personality function.
6887 .. code-block:: llvm
6889 ;; A landing pad which can catch an integer.
6890 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6892 ;; A landing pad that is a cleanup.
6893 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6895 ;; A landing pad which can catch an integer and can only throw a double.
6896 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6898 filter [1 x i8**] [@_ZTId]
6905 LLVM supports the notion of an "intrinsic function". These functions
6906 have well known names and semantics and are required to follow certain
6907 restrictions. Overall, these intrinsics represent an extension mechanism
6908 for the LLVM language that does not require changing all of the
6909 transformations in LLVM when adding to the language (or the bitcode
6910 reader/writer, the parser, etc...).
6912 Intrinsic function names must all start with an "``llvm.``" prefix. This
6913 prefix is reserved in LLVM for intrinsic names; thus, function names may
6914 not begin with this prefix. Intrinsic functions must always be external
6915 functions: you cannot define the body of intrinsic functions. Intrinsic
6916 functions may only be used in call or invoke instructions: it is illegal
6917 to take the address of an intrinsic function. Additionally, because
6918 intrinsic functions are part of the LLVM language, it is required if any
6919 are added that they be documented here.
6921 Some intrinsic functions can be overloaded, i.e., the intrinsic
6922 represents a family of functions that perform the same operation but on
6923 different data types. Because LLVM can represent over 8 million
6924 different integer types, overloading is used commonly to allow an
6925 intrinsic function to operate on any integer type. One or more of the
6926 argument types or the result type can be overloaded to accept any
6927 integer type. Argument types may also be defined as exactly matching a
6928 previous argument's type or the result type. This allows an intrinsic
6929 function which accepts multiple arguments, but needs all of them to be
6930 of the same type, to only be overloaded with respect to a single
6931 argument or the result.
6933 Overloaded intrinsics will have the names of its overloaded argument
6934 types encoded into its function name, each preceded by a period. Only
6935 those types which are overloaded result in a name suffix. Arguments
6936 whose type is matched against another type do not. For example, the
6937 ``llvm.ctpop`` function can take an integer of any width and returns an
6938 integer of exactly the same integer width. This leads to a family of
6939 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
6940 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
6941 overloaded, and only one type suffix is required. Because the argument's
6942 type is matched against the return type, it does not require its own
6945 To learn how to add an intrinsic function, please see the `Extending
6946 LLVM Guide <ExtendingLLVM.html>`_.
6950 Variable Argument Handling Intrinsics
6951 -------------------------------------
6953 Variable argument support is defined in LLVM with the
6954 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
6955 functions. These functions are related to the similarly named macros
6956 defined in the ``<stdarg.h>`` header file.
6958 All of these functions operate on arguments that use a target-specific
6959 value type "``va_list``". The LLVM assembly language reference manual
6960 does not define what this type is, so all transformations should be
6961 prepared to handle these functions regardless of the type used.
6963 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
6964 variable argument handling intrinsic functions are used.
6966 .. code-block:: llvm
6968 ; This struct is different for every platform. For most platforms,
6969 ; it is merely an i8*.
6970 %struct.va_list = type { i8* }
6972 ; For Unix x86_64 platforms, va_list is the following struct:
6973 ; %struct.va_list = type { i32, i32, i8*, i8* }
6975 define i32 @test(i32 %X, ...) {
6976 ; Initialize variable argument processing
6977 %ap = alloca %struct.va_list
6978 %ap2 = bitcast %struct.va_list* %ap to i8*
6979 call void @llvm.va_start(i8* %ap2)
6981 ; Read a single integer argument
6982 %tmp = va_arg i8* %ap2, i32
6984 ; Demonstrate usage of llvm.va_copy and llvm.va_end
6986 %aq2 = bitcast i8** %aq to i8*
6987 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
6988 call void @llvm.va_end(i8* %aq2)
6990 ; Stop processing of arguments.
6991 call void @llvm.va_end(i8* %ap2)
6995 declare void @llvm.va_start(i8*)
6996 declare void @llvm.va_copy(i8*, i8*)
6997 declare void @llvm.va_end(i8*)
7001 '``llvm.va_start``' Intrinsic
7002 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7009 declare void @llvm.va_start(i8* <arglist>)
7014 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
7015 subsequent use by ``va_arg``.
7020 The argument is a pointer to a ``va_list`` element to initialize.
7025 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
7026 available in C. In a target-dependent way, it initializes the
7027 ``va_list`` element to which the argument points, so that the next call
7028 to ``va_arg`` will produce the first variable argument passed to the
7029 function. Unlike the C ``va_start`` macro, this intrinsic does not need
7030 to know the last argument of the function as the compiler can figure
7033 '``llvm.va_end``' Intrinsic
7034 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7041 declare void @llvm.va_end(i8* <arglist>)
7046 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
7047 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
7052 The argument is a pointer to a ``va_list`` to destroy.
7057 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
7058 available in C. In a target-dependent way, it destroys the ``va_list``
7059 element to which the argument points. Calls to
7060 :ref:`llvm.va_start <int_va_start>` and
7061 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
7066 '``llvm.va_copy``' Intrinsic
7067 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7074 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
7079 The '``llvm.va_copy``' intrinsic copies the current argument position
7080 from the source argument list to the destination argument list.
7085 The first argument is a pointer to a ``va_list`` element to initialize.
7086 The second argument is a pointer to a ``va_list`` element to copy from.
7091 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
7092 available in C. In a target-dependent way, it copies the source
7093 ``va_list`` element into the destination ``va_list`` element. This
7094 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
7095 arbitrarily complex and require, for example, memory allocation.
7097 Accurate Garbage Collection Intrinsics
7098 --------------------------------------
7100 LLVM support for `Accurate Garbage Collection <GarbageCollection.html>`_
7101 (GC) requires the implementation and generation of these intrinsics.
7102 These intrinsics allow identification of :ref:`GC roots on the
7103 stack <int_gcroot>`, as well as garbage collector implementations that
7104 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
7105 Front-ends for type-safe garbage collected languages should generate
7106 these intrinsics to make use of the LLVM garbage collectors. For more
7107 details, see `Accurate Garbage Collection with
7108 LLVM <GarbageCollection.html>`_.
7110 The garbage collection intrinsics only operate on objects in the generic
7111 address space (address space zero).
7115 '``llvm.gcroot``' Intrinsic
7116 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7123 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
7128 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
7129 the code generator, and allows some metadata to be associated with it.
7134 The first argument specifies the address of a stack object that contains
7135 the root pointer. The second pointer (which must be either a constant or
7136 a global value address) contains the meta-data to be associated with the
7142 At runtime, a call to this intrinsic stores a null pointer into the
7143 "ptrloc" location. At compile-time, the code generator generates
7144 information to allow the runtime to find the pointer at GC safe points.
7145 The '``llvm.gcroot``' intrinsic may only be used in a function which
7146 :ref:`specifies a GC algorithm <gc>`.
7150 '``llvm.gcread``' Intrinsic
7151 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7158 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
7163 The '``llvm.gcread``' intrinsic identifies reads of references from heap
7164 locations, allowing garbage collector implementations that require read
7170 The second argument is the address to read from, which should be an
7171 address allocated from the garbage collector. The first object is a
7172 pointer to the start of the referenced object, if needed by the language
7173 runtime (otherwise null).
7178 The '``llvm.gcread``' intrinsic has the same semantics as a load
7179 instruction, but may be replaced with substantially more complex code by
7180 the garbage collector runtime, as needed. The '``llvm.gcread``'
7181 intrinsic may only be used in a function which :ref:`specifies a GC
7186 '``llvm.gcwrite``' Intrinsic
7187 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7194 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
7199 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
7200 locations, allowing garbage collector implementations that require write
7201 barriers (such as generational or reference counting collectors).
7206 The first argument is the reference to store, the second is the start of
7207 the object to store it to, and the third is the address of the field of
7208 Obj to store to. If the runtime does not require a pointer to the
7209 object, Obj may be null.
7214 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
7215 instruction, but may be replaced with substantially more complex code by
7216 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
7217 intrinsic may only be used in a function which :ref:`specifies a GC
7220 Code Generator Intrinsics
7221 -------------------------
7223 These intrinsics are provided by LLVM to expose special features that
7224 may only be implemented with code generator support.
7226 '``llvm.returnaddress``' Intrinsic
7227 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7234 declare i8 *@llvm.returnaddress(i32 <level>)
7239 The '``llvm.returnaddress``' intrinsic attempts to compute a
7240 target-specific value indicating the return address of the current
7241 function or one of its callers.
7246 The argument to this intrinsic indicates which function to return the
7247 address for. Zero indicates the calling function, one indicates its
7248 caller, etc. The argument is **required** to be a constant integer
7254 The '``llvm.returnaddress``' intrinsic either returns a pointer
7255 indicating the return address of the specified call frame, or zero if it
7256 cannot be identified. The value returned by this intrinsic is likely to
7257 be incorrect or 0 for arguments other than zero, so it should only be
7258 used for debugging purposes.
7260 Note that calling this intrinsic does not prevent function inlining or
7261 other aggressive transformations, so the value returned may not be that
7262 of the obvious source-language caller.
7264 '``llvm.frameaddress``' Intrinsic
7265 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7272 declare i8* @llvm.frameaddress(i32 <level>)
7277 The '``llvm.frameaddress``' intrinsic attempts to return the
7278 target-specific frame pointer value for the specified stack frame.
7283 The argument to this intrinsic indicates which function to return the
7284 frame pointer for. Zero indicates the calling function, one indicates
7285 its caller, etc. The argument is **required** to be a constant integer
7291 The '``llvm.frameaddress``' intrinsic either returns a pointer
7292 indicating the frame address of the specified call frame, or zero if it
7293 cannot be identified. The value returned by this intrinsic is likely to
7294 be incorrect or 0 for arguments other than zero, so it should only be
7295 used for debugging purposes.
7297 Note that calling this intrinsic does not prevent function inlining or
7298 other aggressive transformations, so the value returned may not be that
7299 of the obvious source-language caller.
7301 '``llvm.frameallocate``' and '``llvm.framerecover``' Intrinsics
7302 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7309 declare i8* @llvm.frameallocate(i32 %size)
7310 declare i8* @llvm.framerecover(i8* %func, i8* %fp)
7315 The '``llvm.frameallocate``' intrinsic allocates stack memory at some fixed
7316 offset from the frame pointer, and the '``llvm.framerecover``'
7317 intrinsic applies that offset to a live frame pointer to recover the address of
7318 the allocation. The offset is computed during frame layout of the caller of
7319 ``llvm.frameallocate``.
7324 The ``size`` argument to '``llvm.frameallocate``' must be a constant integer
7325 indicating the amount of stack memory to allocate. As with allocas, allocating
7326 zero bytes is legal, but the result is undefined.
7328 The ``func`` argument to '``llvm.framerecover``' must be a constant
7329 bitcasted pointer to a function defined in the current module. The code
7330 generator cannot determine the frame allocation offset of functions defined in
7333 The ``fp`` argument to '``llvm.framerecover``' must be a frame
7334 pointer of a call frame that is currently live. The return value of
7335 '``llvm.frameaddress``' is one way to produce such a value, but most platforms
7336 also expose the frame pointer through stack unwinding mechanisms.
7341 These intrinsics allow a group of functions to access one stack memory
7342 allocation in an ancestor stack frame. The memory returned from
7343 '``llvm.frameallocate``' may be allocated prior to stack realignment, so the
7344 memory is only aligned to the ABI-required stack alignment. Each function may
7345 only call '``llvm.frameallocate``' one or zero times from the function entry
7346 block. The frame allocation intrinsic inhibits inlining, as any frame
7347 allocations in the inlined function frame are likely to be at a different
7348 offset from the one used by '``llvm.framerecover``' called with the
7351 .. _int_read_register:
7352 .. _int_write_register:
7354 '``llvm.read_register``' and '``llvm.write_register``' Intrinsics
7355 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7362 declare i32 @llvm.read_register.i32(metadata)
7363 declare i64 @llvm.read_register.i64(metadata)
7364 declare void @llvm.write_register.i32(metadata, i32 @value)
7365 declare void @llvm.write_register.i64(metadata, i64 @value)
7371 The '``llvm.read_register``' and '``llvm.write_register``' intrinsics
7372 provides access to the named register. The register must be valid on
7373 the architecture being compiled to. The type needs to be compatible
7374 with the register being read.
7379 The '``llvm.read_register``' intrinsic returns the current value of the
7380 register, where possible. The '``llvm.write_register``' intrinsic sets
7381 the current value of the register, where possible.
7383 This is useful to implement named register global variables that need
7384 to always be mapped to a specific register, as is common practice on
7385 bare-metal programs including OS kernels.
7387 The compiler doesn't check for register availability or use of the used
7388 register in surrounding code, including inline assembly. Because of that,
7389 allocatable registers are not supported.
7391 Warning: So far it only works with the stack pointer on selected
7392 architectures (ARM, AArch64, PowerPC and x86_64). Significant amount of
7393 work is needed to support other registers and even more so, allocatable
7398 '``llvm.stacksave``' Intrinsic
7399 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7406 declare i8* @llvm.stacksave()
7411 The '``llvm.stacksave``' intrinsic is used to remember the current state
7412 of the function stack, for use with
7413 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
7414 implementing language features like scoped automatic variable sized
7420 This intrinsic returns a opaque pointer value that can be passed to
7421 :ref:`llvm.stackrestore <int_stackrestore>`. When an
7422 ``llvm.stackrestore`` intrinsic is executed with a value saved from
7423 ``llvm.stacksave``, it effectively restores the state of the stack to
7424 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
7425 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
7426 were allocated after the ``llvm.stacksave`` was executed.
7428 .. _int_stackrestore:
7430 '``llvm.stackrestore``' Intrinsic
7431 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7438 declare void @llvm.stackrestore(i8* %ptr)
7443 The '``llvm.stackrestore``' intrinsic is used to restore the state of
7444 the function stack to the state it was in when the corresponding
7445 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
7446 useful for implementing language features like scoped automatic variable
7447 sized arrays in C99.
7452 See the description for :ref:`llvm.stacksave <int_stacksave>`.
7454 '``llvm.prefetch``' Intrinsic
7455 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7462 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
7467 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
7468 insert a prefetch instruction if supported; otherwise, it is a noop.
7469 Prefetches have no effect on the behavior of the program but can change
7470 its performance characteristics.
7475 ``address`` is the address to be prefetched, ``rw`` is the specifier
7476 determining if the fetch should be for a read (0) or write (1), and
7477 ``locality`` is a temporal locality specifier ranging from (0) - no
7478 locality, to (3) - extremely local keep in cache. The ``cache type``
7479 specifies whether the prefetch is performed on the data (1) or
7480 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
7481 arguments must be constant integers.
7486 This intrinsic does not modify the behavior of the program. In
7487 particular, prefetches cannot trap and do not produce a value. On
7488 targets that support this intrinsic, the prefetch can provide hints to
7489 the processor cache for better performance.
7491 '``llvm.pcmarker``' Intrinsic
7492 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7499 declare void @llvm.pcmarker(i32 <id>)
7504 The '``llvm.pcmarker``' intrinsic is a method to export a Program
7505 Counter (PC) in a region of code to simulators and other tools. The
7506 method is target specific, but it is expected that the marker will use
7507 exported symbols to transmit the PC of the marker. The marker makes no
7508 guarantees that it will remain with any specific instruction after
7509 optimizations. It is possible that the presence of a marker will inhibit
7510 optimizations. The intended use is to be inserted after optimizations to
7511 allow correlations of simulation runs.
7516 ``id`` is a numerical id identifying the marker.
7521 This intrinsic does not modify the behavior of the program. Backends
7522 that do not support this intrinsic may ignore it.
7524 '``llvm.readcyclecounter``' Intrinsic
7525 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7532 declare i64 @llvm.readcyclecounter()
7537 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
7538 counter register (or similar low latency, high accuracy clocks) on those
7539 targets that support it. On X86, it should map to RDTSC. On Alpha, it
7540 should map to RPCC. As the backing counters overflow quickly (on the
7541 order of 9 seconds on alpha), this should only be used for small
7547 When directly supported, reading the cycle counter should not modify any
7548 memory. Implementations are allowed to either return a application
7549 specific value or a system wide value. On backends without support, this
7550 is lowered to a constant 0.
7552 Note that runtime support may be conditional on the privilege-level code is
7553 running at and the host platform.
7555 '``llvm.clear_cache``' Intrinsic
7556 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7563 declare void @llvm.clear_cache(i8*, i8*)
7568 The '``llvm.clear_cache``' intrinsic ensures visibility of modifications
7569 in the specified range to the execution unit of the processor. On
7570 targets with non-unified instruction and data cache, the implementation
7571 flushes the instruction cache.
7576 On platforms with coherent instruction and data caches (e.g. x86), this
7577 intrinsic is a nop. On platforms with non-coherent instruction and data
7578 cache (e.g. ARM, MIPS), the intrinsic is lowered either to appropriate
7579 instructions or a system call, if cache flushing requires special
7582 The default behavior is to emit a call to ``__clear_cache`` from the run
7585 This instrinsic does *not* empty the instruction pipeline. Modifications
7586 of the current function are outside the scope of the intrinsic.
7588 '``llvm.instrprof_increment``' Intrinsic
7589 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7596 declare void @llvm.instrprof_increment(i8* <name>, i64 <hash>,
7597 i32 <num-counters>, i32 <index>)
7602 The '``llvm.instrprof_increment``' intrinsic can be emitted by a
7603 frontend for use with instrumentation based profiling. These will be
7604 lowered by the ``-instrprof`` pass to generate execution counts of a
7610 The first argument is a pointer to a global variable containing the
7611 name of the entity being instrumented. This should generally be the
7612 (mangled) function name for a set of counters.
7614 The second argument is a hash value that can be used by the consumer
7615 of the profile data to detect changes to the instrumented source, and
7616 the third is the number of counters associated with ``name``. It is an
7617 error if ``hash`` or ``num-counters`` differ between two instances of
7618 ``instrprof_increment`` that refer to the same name.
7620 The last argument refers to which of the counters for ``name`` should
7621 be incremented. It should be a value between 0 and ``num-counters``.
7626 This intrinsic represents an increment of a profiling counter. It will
7627 cause the ``-instrprof`` pass to generate the appropriate data
7628 structures and the code to increment the appropriate value, in a
7629 format that can be written out by a compiler runtime and consumed via
7630 the ``llvm-profdata`` tool.
7632 Standard C Library Intrinsics
7633 -----------------------------
7635 LLVM provides intrinsics for a few important standard C library
7636 functions. These intrinsics allow source-language front-ends to pass
7637 information about the alignment of the pointer arguments to the code
7638 generator, providing opportunity for more efficient code generation.
7642 '``llvm.memcpy``' Intrinsic
7643 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7648 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
7649 integer bit width and for different address spaces. Not all targets
7650 support all bit widths however.
7654 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7655 i32 <len>, i32 <align>, i1 <isvolatile>)
7656 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7657 i64 <len>, i32 <align>, i1 <isvolatile>)
7662 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7663 source location to the destination location.
7665 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
7666 intrinsics do not return a value, takes extra alignment/isvolatile
7667 arguments and the pointers can be in specified address spaces.
7672 The first argument is a pointer to the destination, the second is a
7673 pointer to the source. The third argument is an integer argument
7674 specifying the number of bytes to copy, the fourth argument is the
7675 alignment of the source and destination locations, and the fifth is a
7676 boolean indicating a volatile access.
7678 If the call to this intrinsic has an alignment value that is not 0 or 1,
7679 then the caller guarantees that both the source and destination pointers
7680 are aligned to that boundary.
7682 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
7683 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7684 very cleanly specified and it is unwise to depend on it.
7689 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7690 source location to the destination location, which are not allowed to
7691 overlap. It copies "len" bytes of memory over. If the argument is known
7692 to be aligned to some boundary, this can be specified as the fourth
7693 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
7695 '``llvm.memmove``' Intrinsic
7696 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7701 This is an overloaded intrinsic. You can use llvm.memmove on any integer
7702 bit width and for different address space. Not all targets support all
7707 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7708 i32 <len>, i32 <align>, i1 <isvolatile>)
7709 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7710 i64 <len>, i32 <align>, i1 <isvolatile>)
7715 The '``llvm.memmove.*``' intrinsics move a block of memory from the
7716 source location to the destination location. It is similar to the
7717 '``llvm.memcpy``' intrinsic but allows the two memory locations to
7720 Note that, unlike the standard libc function, the ``llvm.memmove.*``
7721 intrinsics do not return a value, takes extra alignment/isvolatile
7722 arguments and the pointers can be in specified address spaces.
7727 The first argument is a pointer to the destination, the second is a
7728 pointer to the source. The third argument is an integer argument
7729 specifying the number of bytes to copy, the fourth argument is the
7730 alignment of the source and destination locations, and the fifth is a
7731 boolean indicating a volatile access.
7733 If the call to this intrinsic has an alignment value that is not 0 or 1,
7734 then the caller guarantees that the source and destination pointers are
7735 aligned to that boundary.
7737 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
7738 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
7739 not very cleanly specified and it is unwise to depend on it.
7744 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
7745 source location to the destination location, which may overlap. It
7746 copies "len" bytes of memory over. If the argument is known to be
7747 aligned to some boundary, this can be specified as the fourth argument,
7748 otherwise it should be set to 0 or 1 (both meaning no alignment).
7750 '``llvm.memset.*``' Intrinsics
7751 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7756 This is an overloaded intrinsic. You can use llvm.memset on any integer
7757 bit width and for different address spaces. However, not all targets
7758 support all bit widths.
7762 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
7763 i32 <len>, i32 <align>, i1 <isvolatile>)
7764 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
7765 i64 <len>, i32 <align>, i1 <isvolatile>)
7770 The '``llvm.memset.*``' intrinsics fill a block of memory with a
7771 particular byte value.
7773 Note that, unlike the standard libc function, the ``llvm.memset``
7774 intrinsic does not return a value and takes extra alignment/volatile
7775 arguments. Also, the destination can be in an arbitrary address space.
7780 The first argument is a pointer to the destination to fill, the second
7781 is the byte value with which to fill it, the third argument is an
7782 integer argument specifying the number of bytes to fill, and the fourth
7783 argument is the known alignment of the destination location.
7785 If the call to this intrinsic has an alignment value that is not 0 or 1,
7786 then the caller guarantees that the destination pointer is aligned to
7789 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
7790 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7791 very cleanly specified and it is unwise to depend on it.
7796 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
7797 at the destination location. If the argument is known to be aligned to
7798 some boundary, this can be specified as the fourth argument, otherwise
7799 it should be set to 0 or 1 (both meaning no alignment).
7801 '``llvm.sqrt.*``' Intrinsic
7802 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7807 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
7808 floating point or vector of floating point type. Not all targets support
7813 declare float @llvm.sqrt.f32(float %Val)
7814 declare double @llvm.sqrt.f64(double %Val)
7815 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
7816 declare fp128 @llvm.sqrt.f128(fp128 %Val)
7817 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
7822 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
7823 returning the same value as the libm '``sqrt``' functions would. Unlike
7824 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
7825 negative numbers other than -0.0 (which allows for better optimization,
7826 because there is no need to worry about errno being set).
7827 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
7832 The argument and return value are floating point numbers of the same
7838 This function returns the sqrt of the specified operand if it is a
7839 nonnegative floating point number.
7841 '``llvm.powi.*``' Intrinsic
7842 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7847 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
7848 floating point or vector of floating point type. Not all targets support
7853 declare float @llvm.powi.f32(float %Val, i32 %power)
7854 declare double @llvm.powi.f64(double %Val, i32 %power)
7855 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7856 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7857 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
7862 The '``llvm.powi.*``' intrinsics return the first operand raised to the
7863 specified (positive or negative) power. The order of evaluation of
7864 multiplications is not defined. When a vector of floating point type is
7865 used, the second argument remains a scalar integer value.
7870 The second argument is an integer power, and the first is a value to
7871 raise to that power.
7876 This function returns the first value raised to the second power with an
7877 unspecified sequence of rounding operations.
7879 '``llvm.sin.*``' Intrinsic
7880 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7885 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
7886 floating point or vector of floating point type. Not all targets support
7891 declare float @llvm.sin.f32(float %Val)
7892 declare double @llvm.sin.f64(double %Val)
7893 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7894 declare fp128 @llvm.sin.f128(fp128 %Val)
7895 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7900 The '``llvm.sin.*``' intrinsics return the sine of the operand.
7905 The argument and return value are floating point numbers of the same
7911 This function returns the sine of the specified operand, returning the
7912 same values as the libm ``sin`` functions would, and handles error
7913 conditions in the same way.
7915 '``llvm.cos.*``' Intrinsic
7916 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7921 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
7922 floating point or vector of floating point type. Not all targets support
7927 declare float @llvm.cos.f32(float %Val)
7928 declare double @llvm.cos.f64(double %Val)
7929 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7930 declare fp128 @llvm.cos.f128(fp128 %Val)
7931 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7936 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
7941 The argument and return value are floating point numbers of the same
7947 This function returns the cosine of the specified operand, returning the
7948 same values as the libm ``cos`` functions would, and handles error
7949 conditions in the same way.
7951 '``llvm.pow.*``' Intrinsic
7952 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7957 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
7958 floating point or vector of floating point type. Not all targets support
7963 declare float @llvm.pow.f32(float %Val, float %Power)
7964 declare double @llvm.pow.f64(double %Val, double %Power)
7965 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7966 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7967 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7972 The '``llvm.pow.*``' intrinsics return the first operand raised to the
7973 specified (positive or negative) power.
7978 The second argument is a floating point power, and the first is a value
7979 to raise to that power.
7984 This function returns the first value raised to the second power,
7985 returning the same values as the libm ``pow`` functions would, and
7986 handles error conditions in the same way.
7988 '``llvm.exp.*``' Intrinsic
7989 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7994 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
7995 floating point or vector of floating point type. Not all targets support
8000 declare float @llvm.exp.f32(float %Val)
8001 declare double @llvm.exp.f64(double %Val)
8002 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
8003 declare fp128 @llvm.exp.f128(fp128 %Val)
8004 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
8009 The '``llvm.exp.*``' intrinsics perform the exp function.
8014 The argument and return value are floating point numbers of the same
8020 This function returns the same values as the libm ``exp`` functions
8021 would, and handles error conditions in the same way.
8023 '``llvm.exp2.*``' Intrinsic
8024 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8029 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
8030 floating point or vector of floating point type. Not all targets support
8035 declare float @llvm.exp2.f32(float %Val)
8036 declare double @llvm.exp2.f64(double %Val)
8037 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
8038 declare fp128 @llvm.exp2.f128(fp128 %Val)
8039 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
8044 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
8049 The argument and return value are floating point numbers of the same
8055 This function returns the same values as the libm ``exp2`` functions
8056 would, and handles error conditions in the same way.
8058 '``llvm.log.*``' Intrinsic
8059 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8064 This is an overloaded intrinsic. You can use ``llvm.log`` on any
8065 floating point or vector of floating point type. Not all targets support
8070 declare float @llvm.log.f32(float %Val)
8071 declare double @llvm.log.f64(double %Val)
8072 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
8073 declare fp128 @llvm.log.f128(fp128 %Val)
8074 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
8079 The '``llvm.log.*``' intrinsics perform the log function.
8084 The argument and return value are floating point numbers of the same
8090 This function returns the same values as the libm ``log`` functions
8091 would, and handles error conditions in the same way.
8093 '``llvm.log10.*``' Intrinsic
8094 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8099 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
8100 floating point or vector of floating point type. Not all targets support
8105 declare float @llvm.log10.f32(float %Val)
8106 declare double @llvm.log10.f64(double %Val)
8107 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
8108 declare fp128 @llvm.log10.f128(fp128 %Val)
8109 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
8114 The '``llvm.log10.*``' intrinsics perform the log10 function.
8119 The argument and return value are floating point numbers of the same
8125 This function returns the same values as the libm ``log10`` functions
8126 would, and handles error conditions in the same way.
8128 '``llvm.log2.*``' Intrinsic
8129 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8134 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
8135 floating point or vector of floating point type. Not all targets support
8140 declare float @llvm.log2.f32(float %Val)
8141 declare double @llvm.log2.f64(double %Val)
8142 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
8143 declare fp128 @llvm.log2.f128(fp128 %Val)
8144 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
8149 The '``llvm.log2.*``' intrinsics perform the log2 function.
8154 The argument and return value are floating point numbers of the same
8160 This function returns the same values as the libm ``log2`` functions
8161 would, and handles error conditions in the same way.
8163 '``llvm.fma.*``' Intrinsic
8164 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8169 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
8170 floating point or vector of floating point type. Not all targets support
8175 declare float @llvm.fma.f32(float %a, float %b, float %c)
8176 declare double @llvm.fma.f64(double %a, double %b, double %c)
8177 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
8178 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
8179 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
8184 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
8190 The argument and return value are floating point numbers of the same
8196 This function returns the same values as the libm ``fma`` functions
8197 would, and does not set errno.
8199 '``llvm.fabs.*``' Intrinsic
8200 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8205 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
8206 floating point or vector of floating point type. Not all targets support
8211 declare float @llvm.fabs.f32(float %Val)
8212 declare double @llvm.fabs.f64(double %Val)
8213 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
8214 declare fp128 @llvm.fabs.f128(fp128 %Val)
8215 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
8220 The '``llvm.fabs.*``' intrinsics return the absolute value of the
8226 The argument and return value are floating point numbers of the same
8232 This function returns the same values as the libm ``fabs`` functions
8233 would, and handles error conditions in the same way.
8235 '``llvm.minnum.*``' Intrinsic
8236 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8241 This is an overloaded intrinsic. You can use ``llvm.minnum`` on any
8242 floating point or vector of floating point type. Not all targets support
8247 declare float @llvm.minnum.f32(float %Val0, float %Val1)
8248 declare double @llvm.minnum.f64(double %Val0, double %Val1)
8249 declare x86_fp80 @llvm.minnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
8250 declare fp128 @llvm.minnum.f128(fp128 %Val0, fp128 %Val1)
8251 declare ppc_fp128 @llvm.minnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
8256 The '``llvm.minnum.*``' intrinsics return the minimum of the two
8263 The arguments and return value are floating point numbers of the same
8269 Follows the IEEE-754 semantics for minNum, which also match for libm's
8272 If either operand is a NaN, returns the other non-NaN operand. Returns
8273 NaN only if both operands are NaN. If the operands compare equal,
8274 returns a value that compares equal to both operands. This means that
8275 fmin(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
8277 '``llvm.maxnum.*``' Intrinsic
8278 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8283 This is an overloaded intrinsic. You can use ``llvm.maxnum`` on any
8284 floating point or vector of floating point type. Not all targets support
8289 declare float @llvm.maxnum.f32(float %Val0, float %Val1l)
8290 declare double @llvm.maxnum.f64(double %Val0, double %Val1)
8291 declare x86_fp80 @llvm.maxnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
8292 declare fp128 @llvm.maxnum.f128(fp128 %Val0, fp128 %Val1)
8293 declare ppc_fp128 @llvm.maxnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
8298 The '``llvm.maxnum.*``' intrinsics return the maximum of the two
8305 The arguments and return value are floating point numbers of the same
8310 Follows the IEEE-754 semantics for maxNum, which also match for libm's
8313 If either operand is a NaN, returns the other non-NaN operand. Returns
8314 NaN only if both operands are NaN. If the operands compare equal,
8315 returns a value that compares equal to both operands. This means that
8316 fmax(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
8318 '``llvm.copysign.*``' Intrinsic
8319 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8324 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
8325 floating point or vector of floating point type. Not all targets support
8330 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
8331 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
8332 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
8333 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
8334 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
8339 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
8340 first operand and the sign of the second operand.
8345 The arguments and return value are floating point numbers of the same
8351 This function returns the same values as the libm ``copysign``
8352 functions would, and handles error conditions in the same way.
8354 '``llvm.floor.*``' Intrinsic
8355 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8360 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
8361 floating point or vector of floating point type. Not all targets support
8366 declare float @llvm.floor.f32(float %Val)
8367 declare double @llvm.floor.f64(double %Val)
8368 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
8369 declare fp128 @llvm.floor.f128(fp128 %Val)
8370 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
8375 The '``llvm.floor.*``' intrinsics return the floor of the operand.
8380 The argument and return value are floating point numbers of the same
8386 This function returns the same values as the libm ``floor`` functions
8387 would, and handles error conditions in the same way.
8389 '``llvm.ceil.*``' Intrinsic
8390 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8395 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
8396 floating point or vector of floating point type. Not all targets support
8401 declare float @llvm.ceil.f32(float %Val)
8402 declare double @llvm.ceil.f64(double %Val)
8403 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
8404 declare fp128 @llvm.ceil.f128(fp128 %Val)
8405 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
8410 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
8415 The argument and return value are floating point numbers of the same
8421 This function returns the same values as the libm ``ceil`` functions
8422 would, and handles error conditions in the same way.
8424 '``llvm.trunc.*``' Intrinsic
8425 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8430 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
8431 floating point or vector of floating point type. Not all targets support
8436 declare float @llvm.trunc.f32(float %Val)
8437 declare double @llvm.trunc.f64(double %Val)
8438 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
8439 declare fp128 @llvm.trunc.f128(fp128 %Val)
8440 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
8445 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
8446 nearest integer not larger in magnitude than the operand.
8451 The argument and return value are floating point numbers of the same
8457 This function returns the same values as the libm ``trunc`` functions
8458 would, and handles error conditions in the same way.
8460 '``llvm.rint.*``' Intrinsic
8461 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8466 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
8467 floating point or vector of floating point type. Not all targets support
8472 declare float @llvm.rint.f32(float %Val)
8473 declare double @llvm.rint.f64(double %Val)
8474 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
8475 declare fp128 @llvm.rint.f128(fp128 %Val)
8476 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
8481 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
8482 nearest integer. It may raise an inexact floating-point exception if the
8483 operand isn't an integer.
8488 The argument and return value are floating point numbers of the same
8494 This function returns the same values as the libm ``rint`` functions
8495 would, and handles error conditions in the same way.
8497 '``llvm.nearbyint.*``' Intrinsic
8498 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8503 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
8504 floating point or vector of floating point type. Not all targets support
8509 declare float @llvm.nearbyint.f32(float %Val)
8510 declare double @llvm.nearbyint.f64(double %Val)
8511 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
8512 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
8513 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
8518 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
8524 The argument and return value are floating point numbers of the same
8530 This function returns the same values as the libm ``nearbyint``
8531 functions would, and handles error conditions in the same way.
8533 '``llvm.round.*``' Intrinsic
8534 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8539 This is an overloaded intrinsic. You can use ``llvm.round`` on any
8540 floating point or vector of floating point type. Not all targets support
8545 declare float @llvm.round.f32(float %Val)
8546 declare double @llvm.round.f64(double %Val)
8547 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
8548 declare fp128 @llvm.round.f128(fp128 %Val)
8549 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
8554 The '``llvm.round.*``' intrinsics returns the operand rounded to the
8560 The argument and return value are floating point numbers of the same
8566 This function returns the same values as the libm ``round``
8567 functions would, and handles error conditions in the same way.
8569 Bit Manipulation Intrinsics
8570 ---------------------------
8572 LLVM provides intrinsics for a few important bit manipulation
8573 operations. These allow efficient code generation for some algorithms.
8575 '``llvm.bswap.*``' Intrinsics
8576 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8581 This is an overloaded intrinsic function. You can use bswap on any
8582 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
8586 declare i16 @llvm.bswap.i16(i16 <id>)
8587 declare i32 @llvm.bswap.i32(i32 <id>)
8588 declare i64 @llvm.bswap.i64(i64 <id>)
8593 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
8594 values with an even number of bytes (positive multiple of 16 bits).
8595 These are useful for performing operations on data that is not in the
8596 target's native byte order.
8601 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
8602 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
8603 intrinsic returns an i32 value that has the four bytes of the input i32
8604 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
8605 returned i32 will have its bytes in 3, 2, 1, 0 order. The
8606 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
8607 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
8610 '``llvm.ctpop.*``' Intrinsic
8611 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8616 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
8617 bit width, or on any vector with integer elements. Not all targets
8618 support all bit widths or vector types, however.
8622 declare i8 @llvm.ctpop.i8(i8 <src>)
8623 declare i16 @llvm.ctpop.i16(i16 <src>)
8624 declare i32 @llvm.ctpop.i32(i32 <src>)
8625 declare i64 @llvm.ctpop.i64(i64 <src>)
8626 declare i256 @llvm.ctpop.i256(i256 <src>)
8627 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
8632 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
8638 The only argument is the value to be counted. The argument may be of any
8639 integer type, or a vector with integer elements. The return type must
8640 match the argument type.
8645 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
8646 each element of a vector.
8648 '``llvm.ctlz.*``' Intrinsic
8649 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8654 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
8655 integer bit width, or any vector whose elements are integers. Not all
8656 targets support all bit widths or vector types, however.
8660 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
8661 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
8662 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
8663 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
8664 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
8665 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
8670 The '``llvm.ctlz``' family of intrinsic functions counts the number of
8671 leading zeros in a variable.
8676 The first argument is the value to be counted. This argument may be of
8677 any integer type, or a vector with integer element type. The return
8678 type must match the first argument type.
8680 The second argument must be a constant and is a flag to indicate whether
8681 the intrinsic should ensure that a zero as the first argument produces a
8682 defined result. Historically some architectures did not provide a
8683 defined result for zero values as efficiently, and many algorithms are
8684 now predicated on avoiding zero-value inputs.
8689 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
8690 zeros in a variable, or within each element of the vector. If
8691 ``src == 0`` then the result is the size in bits of the type of ``src``
8692 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
8693 ``llvm.ctlz(i32 2) = 30``.
8695 '``llvm.cttz.*``' Intrinsic
8696 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8701 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
8702 integer bit width, or any vector of integer elements. Not all targets
8703 support all bit widths or vector types, however.
8707 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
8708 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
8709 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
8710 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
8711 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
8712 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
8717 The '``llvm.cttz``' family of intrinsic functions counts the number of
8723 The first argument is the value to be counted. This argument may be of
8724 any integer type, or a vector with integer element type. The return
8725 type must match the first argument type.
8727 The second argument must be a constant and is a flag to indicate whether
8728 the intrinsic should ensure that a zero as the first argument produces a
8729 defined result. Historically some architectures did not provide a
8730 defined result for zero values as efficiently, and many algorithms are
8731 now predicated on avoiding zero-value inputs.
8736 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
8737 zeros in a variable, or within each element of a vector. If ``src == 0``
8738 then the result is the size in bits of the type of ``src`` if
8739 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
8740 ``llvm.cttz(2) = 1``.
8742 Arithmetic with Overflow Intrinsics
8743 -----------------------------------
8745 LLVM provides intrinsics for some arithmetic with overflow operations.
8747 '``llvm.sadd.with.overflow.*``' Intrinsics
8748 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8753 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
8754 on any integer bit width.
8758 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
8759 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8760 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
8765 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8766 a signed addition of the two arguments, and indicate whether an overflow
8767 occurred during the signed summation.
8772 The arguments (%a and %b) and the first element of the result structure
8773 may be of integer types of any bit width, but they must have the same
8774 bit width. The second element of the result structure must be of type
8775 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8781 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8782 a signed addition of the two variables. They return a structure --- the
8783 first element of which is the signed summation, and the second element
8784 of which is a bit specifying if the signed summation resulted in an
8790 .. code-block:: llvm
8792 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8793 %sum = extractvalue {i32, i1} %res, 0
8794 %obit = extractvalue {i32, i1} %res, 1
8795 br i1 %obit, label %overflow, label %normal
8797 '``llvm.uadd.with.overflow.*``' Intrinsics
8798 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8803 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
8804 on any integer bit width.
8808 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
8809 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8810 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
8815 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8816 an unsigned addition of the two arguments, and indicate whether a carry
8817 occurred during the unsigned summation.
8822 The arguments (%a and %b) and the first element of the result structure
8823 may be of integer types of any bit width, but they must have the same
8824 bit width. The second element of the result structure must be of type
8825 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8831 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8832 an unsigned addition of the two arguments. They return a structure --- the
8833 first element of which is the sum, and the second element of which is a
8834 bit specifying if the unsigned summation resulted in a carry.
8839 .. code-block:: llvm
8841 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8842 %sum = extractvalue {i32, i1} %res, 0
8843 %obit = extractvalue {i32, i1} %res, 1
8844 br i1 %obit, label %carry, label %normal
8846 '``llvm.ssub.with.overflow.*``' Intrinsics
8847 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8852 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
8853 on any integer bit width.
8857 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
8858 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8859 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
8864 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8865 a signed subtraction of the two arguments, and indicate whether an
8866 overflow occurred during the signed subtraction.
8871 The arguments (%a and %b) and the first element of the result structure
8872 may be of integer types of any bit width, but they must have the same
8873 bit width. The second element of the result structure must be of type
8874 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8880 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8881 a signed subtraction of the two arguments. They return a structure --- the
8882 first element of which is the subtraction, and the second element of
8883 which is a bit specifying if the signed subtraction resulted in an
8889 .. code-block:: llvm
8891 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8892 %sum = extractvalue {i32, i1} %res, 0
8893 %obit = extractvalue {i32, i1} %res, 1
8894 br i1 %obit, label %overflow, label %normal
8896 '``llvm.usub.with.overflow.*``' Intrinsics
8897 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8902 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
8903 on any integer bit width.
8907 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
8908 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8909 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
8914 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8915 an unsigned subtraction of the two arguments, and indicate whether an
8916 overflow occurred during the unsigned subtraction.
8921 The arguments (%a and %b) and the first element of the result structure
8922 may be of integer types of any bit width, but they must have the same
8923 bit width. The second element of the result structure must be of type
8924 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8930 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8931 an unsigned subtraction of the two arguments. They return a structure ---
8932 the first element of which is the subtraction, and the second element of
8933 which is a bit specifying if the unsigned subtraction resulted in an
8939 .. code-block:: llvm
8941 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8942 %sum = extractvalue {i32, i1} %res, 0
8943 %obit = extractvalue {i32, i1} %res, 1
8944 br i1 %obit, label %overflow, label %normal
8946 '``llvm.smul.with.overflow.*``' Intrinsics
8947 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8952 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
8953 on any integer bit width.
8957 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
8958 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8959 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
8964 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8965 a signed multiplication of the two arguments, and indicate whether an
8966 overflow occurred during the signed multiplication.
8971 The arguments (%a and %b) and the first element of the result structure
8972 may be of integer types of any bit width, but they must have the same
8973 bit width. The second element of the result structure must be of type
8974 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8980 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8981 a signed multiplication of the two arguments. They return a structure ---
8982 the first element of which is the multiplication, and the second element
8983 of which is a bit specifying if the signed multiplication resulted in an
8989 .. code-block:: llvm
8991 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8992 %sum = extractvalue {i32, i1} %res, 0
8993 %obit = extractvalue {i32, i1} %res, 1
8994 br i1 %obit, label %overflow, label %normal
8996 '``llvm.umul.with.overflow.*``' Intrinsics
8997 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9002 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
9003 on any integer bit width.
9007 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
9008 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
9009 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
9014 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
9015 a unsigned multiplication of the two arguments, and indicate whether an
9016 overflow occurred during the unsigned multiplication.
9021 The arguments (%a and %b) and the first element of the result structure
9022 may be of integer types of any bit width, but they must have the same
9023 bit width. The second element of the result structure must be of type
9024 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
9030 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
9031 an unsigned multiplication of the two arguments. They return a structure ---
9032 the first element of which is the multiplication, and the second
9033 element of which is a bit specifying if the unsigned multiplication
9034 resulted in an overflow.
9039 .. code-block:: llvm
9041 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
9042 %sum = extractvalue {i32, i1} %res, 0
9043 %obit = extractvalue {i32, i1} %res, 1
9044 br i1 %obit, label %overflow, label %normal
9046 Specialised Arithmetic Intrinsics
9047 ---------------------------------
9049 '``llvm.fmuladd.*``' Intrinsic
9050 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9057 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
9058 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
9063 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
9064 expressions that can be fused if the code generator determines that (a) the
9065 target instruction set has support for a fused operation, and (b) that the
9066 fused operation is more efficient than the equivalent, separate pair of mul
9067 and add instructions.
9072 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
9073 multiplicands, a and b, and an addend c.
9082 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
9084 is equivalent to the expression a \* b + c, except that rounding will
9085 not be performed between the multiplication and addition steps if the
9086 code generator fuses the operations. Fusion is not guaranteed, even if
9087 the target platform supports it. If a fused multiply-add is required the
9088 corresponding llvm.fma.\* intrinsic function should be used
9089 instead. This never sets errno, just as '``llvm.fma.*``'.
9094 .. code-block:: llvm
9096 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields float:r2 = (a * b) + c
9098 Half Precision Floating Point Intrinsics
9099 ----------------------------------------
9101 For most target platforms, half precision floating point is a
9102 storage-only format. This means that it is a dense encoding (in memory)
9103 but does not support computation in the format.
9105 This means that code must first load the half-precision floating point
9106 value as an i16, then convert it to float with
9107 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
9108 then be performed on the float value (including extending to double
9109 etc). To store the value back to memory, it is first converted to float
9110 if needed, then converted to i16 with
9111 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
9114 .. _int_convert_to_fp16:
9116 '``llvm.convert.to.fp16``' Intrinsic
9117 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9124 declare i16 @llvm.convert.to.fp16.f32(float %a)
9125 declare i16 @llvm.convert.to.fp16.f64(double %a)
9130 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
9131 conventional floating point type to half precision floating point format.
9136 The intrinsic function contains single argument - the value to be
9142 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
9143 conventional floating point format to half precision floating point format. The
9144 return value is an ``i16`` which contains the converted number.
9149 .. code-block:: llvm
9151 %res = call i16 @llvm.convert.to.fp16.f32(float %a)
9152 store i16 %res, i16* @x, align 2
9154 .. _int_convert_from_fp16:
9156 '``llvm.convert.from.fp16``' Intrinsic
9157 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9164 declare float @llvm.convert.from.fp16.f32(i16 %a)
9165 declare double @llvm.convert.from.fp16.f64(i16 %a)
9170 The '``llvm.convert.from.fp16``' intrinsic function performs a
9171 conversion from half precision floating point format to single precision
9172 floating point format.
9177 The intrinsic function contains single argument - the value to be
9183 The '``llvm.convert.from.fp16``' intrinsic function performs a
9184 conversion from half single precision floating point format to single
9185 precision floating point format. The input half-float value is
9186 represented by an ``i16`` value.
9191 .. code-block:: llvm
9193 %a = load i16* @x, align 2
9194 %res = call float @llvm.convert.from.fp16(i16 %a)
9199 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
9200 prefix), are described in the `LLVM Source Level
9201 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
9204 Exception Handling Intrinsics
9205 -----------------------------
9207 The LLVM exception handling intrinsics (which all start with
9208 ``llvm.eh.`` prefix), are described in the `LLVM Exception
9209 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
9213 Trampoline Intrinsics
9214 ---------------------
9216 These intrinsics make it possible to excise one parameter, marked with
9217 the :ref:`nest <nest>` attribute, from a function. The result is a
9218 callable function pointer lacking the nest parameter - the caller does
9219 not need to provide a value for it. Instead, the value to use is stored
9220 in advance in a "trampoline", a block of memory usually allocated on the
9221 stack, which also contains code to splice the nest value into the
9222 argument list. This is used to implement the GCC nested function address
9225 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
9226 then the resulting function pointer has signature ``i32 (i32, i32)*``.
9227 It can be created as follows:
9229 .. code-block:: llvm
9231 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
9232 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
9233 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
9234 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
9235 %fp = bitcast i8* %p to i32 (i32, i32)*
9237 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
9238 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
9242 '``llvm.init.trampoline``' Intrinsic
9243 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9250 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
9255 This fills the memory pointed to by ``tramp`` with executable code,
9256 turning it into a trampoline.
9261 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
9262 pointers. The ``tramp`` argument must point to a sufficiently large and
9263 sufficiently aligned block of memory; this memory is written to by the
9264 intrinsic. Note that the size and the alignment are target-specific -
9265 LLVM currently provides no portable way of determining them, so a
9266 front-end that generates this intrinsic needs to have some
9267 target-specific knowledge. The ``func`` argument must hold a function
9268 bitcast to an ``i8*``.
9273 The block of memory pointed to by ``tramp`` is filled with target
9274 dependent code, turning it into a function. Then ``tramp`` needs to be
9275 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
9276 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
9277 function's signature is the same as that of ``func`` with any arguments
9278 marked with the ``nest`` attribute removed. At most one such ``nest``
9279 argument is allowed, and it must be of pointer type. Calling the new
9280 function is equivalent to calling ``func`` with the same argument list,
9281 but with ``nval`` used for the missing ``nest`` argument. If, after
9282 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
9283 modified, then the effect of any later call to the returned function
9284 pointer is undefined.
9288 '``llvm.adjust.trampoline``' Intrinsic
9289 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9296 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
9301 This performs any required machine-specific adjustment to the address of
9302 a trampoline (passed as ``tramp``).
9307 ``tramp`` must point to a block of memory which already has trampoline
9308 code filled in by a previous call to
9309 :ref:`llvm.init.trampoline <int_it>`.
9314 On some architectures the address of the code to be executed needs to be
9315 different than the address where the trampoline is actually stored. This
9316 intrinsic returns the executable address corresponding to ``tramp``
9317 after performing the required machine specific adjustments. The pointer
9318 returned can then be :ref:`bitcast and executed <int_trampoline>`.
9320 Masked Vector Load and Store Intrinsics
9321 ---------------------------------------
9323 LLVM provides intrinsics for predicated vector load and store operations. The predicate is specified by a mask operand, which holds one bit per vector element, switching the associated vector lane on or off. The memory addresses corresponding to the "off" lanes are not accessed. When all bits of the mask are on, the intrinsic is identical to a regular vector load or store. When all bits are off, no memory is accessed.
9327 '``llvm.masked.load.*``' Intrinsics
9328 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9332 This is an overloaded intrinsic. The loaded data is a vector of any integer or floating point data type.
9336 declare <16 x float> @llvm.masked.load.v16f32 (<16 x float>* <ptr>, i32 <alignment>, <16 x i1> <mask>, <16 x float> <passthru>)
9337 declare <2 x double> @llvm.masked.load.v2f64 (<2 x double>* <ptr>, i32 <alignment>, <2 x i1> <mask>, <2 x double> <passthru>)
9342 Reads a vector from memory according to the provided mask. The mask holds a bit for each vector lane, and is used to prevent memory accesses to the masked-off lanes. The masked-off lanes in the result vector are taken from the corresponding lanes in the passthru operand.
9348 The first operand is the base pointer for the load. The second operand is the alignment of the source location. It must be a constant integer value. The third operand, mask, is a vector of boolean 'i1' values with the same number of elements as the return type. The fourth is a pass-through value that is used to fill the masked-off lanes of the result. The return type, underlying type of the base pointer and the type of passthru operand are the same vector types.
9354 The '``llvm.masked.load``' intrinsic is designed for conditional reading of selected vector elements in a single IR operation. It is useful for targets that support vector masked loads and allows vectorizing predicated basic blocks on these targets. Other targets may support this intrinsic differently, for example by lowering it into a sequence of branches that guard scalar load operations.
9355 The result of this operation is equivalent to a regular vector load instruction followed by a 'select' between the loaded and the passthru values, predicated on the same mask. However, using this intrinsic prevents exceptions on memory access to masked-off lanes.
9360 %res = call <16 x float> @llvm.masked.load.v16f32 (<16 x float>* %ptr, i32 4, <16 x i1>%mask, <16 x float> %passthru)
9362 ;; The result of the two following instructions is identical aside from potential memory access exception
9363 %loadlal = load <16 x float>* %ptr, align 4
9364 %res = select <16 x i1> %mask, <16 x float> %loadlal, <16 x float> %passthru
9368 '``llvm.masked.store.*``' Intrinsics
9369 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9373 This is an overloaded intrinsic. The data stored in memory is a vector of any integer or floating point data type.
9377 declare void @llvm.masked.store.v8i32 (<8 x i32> <value>, <8 x i32> * <ptr>, i32 <alignment>, <8 x i1> <mask>)
9378 declare void @llvm.masked.store.v16f32(<16 x i32> <value>, <16 x i32>* <ptr>, i32 <alignment>, <16 x i1> <mask>)
9383 Writes a vector to memory according to the provided mask. The mask holds a bit for each vector lane, and is used to prevent memory accesses to the masked-off lanes.
9388 The first operand is the vector value to be written to memory. The second operand is the base pointer for the store, it has the same underlying type as the value operand. The third operand is the alignment of the destination location. The fourth operand, mask, is a vector of boolean values. The types of the mask and the value operand must have the same number of vector elements.
9394 The '``llvm.masked.store``' intrinsics is designed for conditional writing of selected vector elements in a single IR operation. It is useful for targets that support vector masked store and allows vectorizing predicated basic blocks on these targets. Other targets may support this intrinsic differently, for example by lowering it into a sequence of branches that guard scalar store operations.
9395 The result of this operation is equivalent to a load-modify-store sequence. However, using this intrinsic prevents exceptions and data races on memory access to masked-off lanes.
9399 call void @llvm.masked.store.v16f32(<16 x float> %value, <16 x float>* %ptr, i32 4, <16 x i1> %mask)
9401 ;; The result of the following instructions is identical aside from potential data races and memory access exceptions
9402 %oldval = load <16 x float>* %ptr, align 4
9403 %res = select <16 x i1> %mask, <16 x float> %value, <16 x float> %oldval
9404 store <16 x float> %res, <16 x float>* %ptr, align 4
9410 This class of intrinsics provides information about the lifetime of
9411 memory objects and ranges where variables are immutable.
9415 '``llvm.lifetime.start``' Intrinsic
9416 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9423 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
9428 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
9434 The first argument is a constant integer representing the size of the
9435 object, or -1 if it is variable sized. The second argument is a pointer
9441 This intrinsic indicates that before this point in the code, the value
9442 of the memory pointed to by ``ptr`` is dead. This means that it is known
9443 to never be used and has an undefined value. A load from the pointer
9444 that precedes this intrinsic can be replaced with ``'undef'``.
9448 '``llvm.lifetime.end``' Intrinsic
9449 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9456 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
9461 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
9467 The first argument is a constant integer representing the size of the
9468 object, or -1 if it is variable sized. The second argument is a pointer
9474 This intrinsic indicates that after this point in the code, the value of
9475 the memory pointed to by ``ptr`` is dead. This means that it is known to
9476 never be used and has an undefined value. Any stores into the memory
9477 object following this intrinsic may be removed as dead.
9479 '``llvm.invariant.start``' Intrinsic
9480 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9487 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
9492 The '``llvm.invariant.start``' intrinsic specifies that the contents of
9493 a memory object will not change.
9498 The first argument is a constant integer representing the size of the
9499 object, or -1 if it is variable sized. The second argument is a pointer
9505 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
9506 the return value, the referenced memory location is constant and
9509 '``llvm.invariant.end``' Intrinsic
9510 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9517 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
9522 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
9523 memory object are mutable.
9528 The first argument is the matching ``llvm.invariant.start`` intrinsic.
9529 The second argument is a constant integer representing the size of the
9530 object, or -1 if it is variable sized and the third argument is a
9531 pointer to the object.
9536 This intrinsic indicates that the memory is mutable again.
9541 This class of intrinsics is designed to be generic and has no specific
9544 '``llvm.var.annotation``' Intrinsic
9545 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9552 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9557 The '``llvm.var.annotation``' intrinsic.
9562 The first argument is a pointer to a value, the second is a pointer to a
9563 global string, the third is a pointer to a global string which is the
9564 source file name, and the last argument is the line number.
9569 This intrinsic allows annotation of local variables with arbitrary
9570 strings. This can be useful for special purpose optimizations that want
9571 to look for these annotations. These have no other defined use; they are
9572 ignored by code generation and optimization.
9574 '``llvm.ptr.annotation.*``' Intrinsic
9575 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9580 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
9581 pointer to an integer of any width. *NOTE* you must specify an address space for
9582 the pointer. The identifier for the default address space is the integer
9587 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9588 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
9589 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
9590 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
9591 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
9596 The '``llvm.ptr.annotation``' intrinsic.
9601 The first argument is a pointer to an integer value of arbitrary bitwidth
9602 (result of some expression), the second is a pointer to a global string, the
9603 third is a pointer to a global string which is the source file name, and the
9604 last argument is the line number. It returns the value of the first argument.
9609 This intrinsic allows annotation of a pointer to an integer with arbitrary
9610 strings. This can be useful for special purpose optimizations that want to look
9611 for these annotations. These have no other defined use; they are ignored by code
9612 generation and optimization.
9614 '``llvm.annotation.*``' Intrinsic
9615 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9620 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
9621 any integer bit width.
9625 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
9626 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
9627 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
9628 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
9629 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
9634 The '``llvm.annotation``' intrinsic.
9639 The first argument is an integer value (result of some expression), the
9640 second is a pointer to a global string, the third is a pointer to a
9641 global string which is the source file name, and the last argument is
9642 the line number. It returns the value of the first argument.
9647 This intrinsic allows annotations to be put on arbitrary expressions
9648 with arbitrary strings. This can be useful for special purpose
9649 optimizations that want to look for these annotations. These have no
9650 other defined use; they are ignored by code generation and optimization.
9652 '``llvm.trap``' Intrinsic
9653 ^^^^^^^^^^^^^^^^^^^^^^^^^
9660 declare void @llvm.trap() noreturn nounwind
9665 The '``llvm.trap``' intrinsic.
9675 This intrinsic is lowered to the target dependent trap instruction. If
9676 the target does not have a trap instruction, this intrinsic will be
9677 lowered to a call of the ``abort()`` function.
9679 '``llvm.debugtrap``' Intrinsic
9680 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9687 declare void @llvm.debugtrap() nounwind
9692 The '``llvm.debugtrap``' intrinsic.
9702 This intrinsic is lowered to code which is intended to cause an
9703 execution trap with the intention of requesting the attention of a
9706 '``llvm.stackprotector``' Intrinsic
9707 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9714 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
9719 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
9720 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
9721 is placed on the stack before local variables.
9726 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
9727 The first argument is the value loaded from the stack guard
9728 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
9729 enough space to hold the value of the guard.
9734 This intrinsic causes the prologue/epilogue inserter to force the position of
9735 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
9736 to ensure that if a local variable on the stack is overwritten, it will destroy
9737 the value of the guard. When the function exits, the guard on the stack is
9738 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
9739 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
9740 calling the ``__stack_chk_fail()`` function.
9742 '``llvm.stackprotectorcheck``' Intrinsic
9743 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9750 declare void @llvm.stackprotectorcheck(i8** <guard>)
9755 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
9756 created stack protector and if they are not equal calls the
9757 ``__stack_chk_fail()`` function.
9762 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
9763 the variable ``@__stack_chk_guard``.
9768 This intrinsic is provided to perform the stack protector check by comparing
9769 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
9770 values do not match call the ``__stack_chk_fail()`` function.
9772 The reason to provide this as an IR level intrinsic instead of implementing it
9773 via other IR operations is that in order to perform this operation at the IR
9774 level without an intrinsic, one would need to create additional basic blocks to
9775 handle the success/failure cases. This makes it difficult to stop the stack
9776 protector check from disrupting sibling tail calls in Codegen. With this
9777 intrinsic, we are able to generate the stack protector basic blocks late in
9778 codegen after the tail call decision has occurred.
9780 '``llvm.objectsize``' Intrinsic
9781 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9788 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
9789 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
9794 The ``llvm.objectsize`` intrinsic is designed to provide information to
9795 the optimizers to determine at compile time whether a) an operation
9796 (like memcpy) will overflow a buffer that corresponds to an object, or
9797 b) that a runtime check for overflow isn't necessary. An object in this
9798 context means an allocation of a specific class, structure, array, or
9804 The ``llvm.objectsize`` intrinsic takes two arguments. The first
9805 argument is a pointer to or into the ``object``. The second argument is
9806 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
9807 or -1 (if false) when the object size is unknown. The second argument
9808 only accepts constants.
9813 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
9814 the size of the object concerned. If the size cannot be determined at
9815 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
9816 on the ``min`` argument).
9818 '``llvm.expect``' Intrinsic
9819 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9824 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
9829 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
9830 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
9831 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
9836 The ``llvm.expect`` intrinsic provides information about expected (the
9837 most probable) value of ``val``, which can be used by optimizers.
9842 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
9843 a value. The second argument is an expected value, this needs to be a
9844 constant value, variables are not allowed.
9849 This intrinsic is lowered to the ``val``.
9851 '``llvm.assume``' Intrinsic
9852 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9859 declare void @llvm.assume(i1 %cond)
9864 The ``llvm.assume`` allows the optimizer to assume that the provided
9865 condition is true. This information can then be used in simplifying other parts
9871 The condition which the optimizer may assume is always true.
9876 The intrinsic allows the optimizer to assume that the provided condition is
9877 always true whenever the control flow reaches the intrinsic call. No code is
9878 generated for this intrinsic, and instructions that contribute only to the
9879 provided condition are not used for code generation. If the condition is
9880 violated during execution, the behavior is undefined.
9882 Note that the optimizer might limit the transformations performed on values
9883 used by the ``llvm.assume`` intrinsic in order to preserve the instructions
9884 only used to form the intrinsic's input argument. This might prove undesirable
9885 if the extra information provided by the ``llvm.assume`` intrinsic does not cause
9886 sufficient overall improvement in code quality. For this reason,
9887 ``llvm.assume`` should not be used to document basic mathematical invariants
9888 that the optimizer can otherwise deduce or facts that are of little use to the
9891 '``llvm.donothing``' Intrinsic
9892 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9899 declare void @llvm.donothing() nounwind readnone
9904 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's one of only
9905 two intrinsics (besides ``llvm.experimental.patchpoint``) that can be called
9906 with an invoke instruction.
9916 This intrinsic does nothing, and it's removed by optimizers and ignored
9919 Stack Map Intrinsics
9920 --------------------
9922 LLVM provides experimental intrinsics to support runtime patching
9923 mechanisms commonly desired in dynamic language JITs. These intrinsics
9924 are described in :doc:`StackMaps`.