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], [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 dynamically
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
357 unintrusive as possible. This convention behaves identically 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 use 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 variable 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 aliases 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 be 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 known 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 Aliases 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).
1015 ``dereferenceable_or_null(<n>)``
1016 This indicates that the parameter or return value isn't both
1017 non-null and non-dereferenceable (up to ``<n>`` bytes) at the same
1018 time. All non-null pointers tagged with
1019 ``dereferenceable_or_null(<n>)`` are ``dereferenceable(<n>)``.
1020 For address space 0 ``dereferenceable_or_null(<n>)`` implies that
1021 a pointer is exactly one of ``dereferenceable(<n>)`` or ``null``,
1022 and in other address spaces ``dereferenceable_or_null(<n>)``
1023 implies that a pointer is at least one of ``dereferenceable(<n>)``
1024 or ``null`` (i.e. it may be both ``null`` and
1025 ``dereferenceable(<n>)``). This attribute may only be applied to
1026 pointer typed parameters.
1030 Garbage Collector Strategy Names
1031 --------------------------------
1033 Each function may specify a garbage collector strategy name, which is simply a
1036 .. code-block:: llvm
1038 define void @f() gc "name" { ... }
1040 The supported values of *name* includes those :ref:`built in to LLVM
1041 <builtin-gc-strategies>` and any provided by loaded plugins. Specifying a GC
1042 strategy will cause the compiler to alter its output in order to support the
1043 named garbage collection algorithm. Note that LLVM itself does not contain a
1044 garbage collector, this functionality is restricted to generating machine code
1045 which can interoperate with a collector provided externally.
1052 Prefix data is data associated with a function which the code
1053 generator will emit immediately before the function's entrypoint.
1054 The purpose of this feature is to allow frontends to associate
1055 language-specific runtime metadata with specific functions and make it
1056 available through the function pointer while still allowing the
1057 function pointer to be called.
1059 To access the data for a given function, a program may bitcast the
1060 function pointer to a pointer to the constant's type and dereference
1061 index -1. This implies that the IR symbol points just past the end of
1062 the prefix data. For instance, take the example of a function annotated
1063 with a single ``i32``,
1065 .. code-block:: llvm
1067 define void @f() prefix i32 123 { ... }
1069 The prefix data can be referenced as,
1071 .. code-block:: llvm
1073 %0 = bitcast void* () @f to i32*
1074 %a = getelementptr inbounds i32, i32* %0, i32 -1
1075 %b = load i32, i32* %a
1077 Prefix data is laid out as if it were an initializer for a global variable
1078 of the prefix data's type. The function will be placed such that the
1079 beginning of the prefix data is aligned. This means that if the size
1080 of the prefix data is not a multiple of the alignment size, the
1081 function's entrypoint will not be aligned. If alignment of the
1082 function's entrypoint is desired, padding must be added to the prefix
1085 A function may have prefix data but no body. This has similar semantics
1086 to the ``available_externally`` linkage in that the data may be used by the
1087 optimizers but will not be emitted in the object file.
1094 The ``prologue`` attribute allows arbitrary code (encoded as bytes) to
1095 be inserted prior to the function body. This can be used for enabling
1096 function hot-patching and instrumentation.
1098 To maintain the semantics of ordinary function calls, the prologue data must
1099 have a particular format. Specifically, it must begin with a sequence of
1100 bytes which decode to a sequence of machine instructions, valid for the
1101 module's target, which transfer control to the point immediately succeeding
1102 the prologue data, without performing any other visible action. This allows
1103 the inliner and other passes to reason about the semantics of the function
1104 definition without needing to reason about the prologue data. Obviously this
1105 makes the format of the prologue data highly target dependent.
1107 A trivial example of valid prologue data for the x86 architecture is ``i8 144``,
1108 which encodes the ``nop`` instruction:
1110 .. code-block:: llvm
1112 define void @f() prologue i8 144 { ... }
1114 Generally prologue data can be formed by encoding a relative branch instruction
1115 which skips the metadata, as in this example of valid prologue data for the
1116 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
1118 .. code-block:: llvm
1120 %0 = type <{ i8, i8, i8* }>
1122 define void @f() prologue %0 <{ i8 235, i8 8, i8* @md}> { ... }
1124 A function may have prologue data but no body. This has similar semantics
1125 to the ``available_externally`` linkage in that the data may be used by the
1126 optimizers but will not be emitted in the object file.
1133 Attribute groups are groups of attributes that are referenced by objects within
1134 the IR. They are important for keeping ``.ll`` files readable, because a lot of
1135 functions will use the same set of attributes. In the degenerative case of a
1136 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
1137 group will capture the important command line flags used to build that file.
1139 An attribute group is a module-level object. To use an attribute group, an
1140 object references the attribute group's ID (e.g. ``#37``). An object may refer
1141 to more than one attribute group. In that situation, the attributes from the
1142 different groups are merged.
1144 Here is an example of attribute groups for a function that should always be
1145 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
1147 .. code-block:: llvm
1149 ; Target-independent attributes:
1150 attributes #0 = { alwaysinline alignstack=4 }
1152 ; Target-dependent attributes:
1153 attributes #1 = { "no-sse" }
1155 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
1156 define void @f() #0 #1 { ... }
1163 Function attributes are set to communicate additional information about
1164 a function. Function attributes are considered to be part of the
1165 function, not of the function type, so functions with different function
1166 attributes can have the same function type.
1168 Function attributes are simple keywords that follow the type specified.
1169 If multiple attributes are needed, they are space separated. For
1172 .. code-block:: llvm
1174 define void @f() noinline { ... }
1175 define void @f() alwaysinline { ... }
1176 define void @f() alwaysinline optsize { ... }
1177 define void @f() optsize { ... }
1180 This attribute indicates that, when emitting the prologue and
1181 epilogue, the backend should forcibly align the stack pointer.
1182 Specify the desired alignment, which must be a power of two, in
1185 This attribute indicates that the inliner should attempt to inline
1186 this function into callers whenever possible, ignoring any active
1187 inlining size threshold for this caller.
1189 This indicates that the callee function at a call site should be
1190 recognized as a built-in function, even though the function's declaration
1191 uses the ``nobuiltin`` attribute. This is only valid at call sites for
1192 direct calls to functions that are declared with the ``nobuiltin``
1195 This attribute indicates that this function is rarely called. When
1196 computing edge weights, basic blocks post-dominated by a cold
1197 function call are also considered to be cold; and, thus, given low
1200 This attribute indicates that the source code contained a hint that
1201 inlining this function is desirable (such as the "inline" keyword in
1202 C/C++). It is just a hint; it imposes no requirements on the
1205 This attribute indicates that the function should be added to a
1206 jump-instruction table at code-generation time, and that all address-taken
1207 references to this function should be replaced with a reference to the
1208 appropriate jump-instruction-table function pointer. Note that this creates
1209 a new pointer for the original function, which means that code that depends
1210 on function-pointer identity can break. So, any function annotated with
1211 ``jumptable`` must also be ``unnamed_addr``.
1213 This attribute suggests that optimization passes and code generator
1214 passes make choices that keep the code size of this function as small
1215 as possible and perform optimizations that may sacrifice runtime
1216 performance in order to minimize the size of the generated code.
1218 This attribute disables prologue / epilogue emission for the
1219 function. This can have very system-specific consequences.
1221 This indicates that the callee function at a call site is not recognized as
1222 a built-in function. LLVM will retain the original call and not replace it
1223 with equivalent code based on the semantics of the built-in function, unless
1224 the call site uses the ``builtin`` attribute. This is valid at call sites
1225 and on function declarations and definitions.
1227 This attribute indicates that calls to the function cannot be
1228 duplicated. A call to a ``noduplicate`` function may be moved
1229 within its parent function, but may not be duplicated within
1230 its parent function.
1232 A function containing a ``noduplicate`` call may still
1233 be an inlining candidate, provided that the call is not
1234 duplicated by inlining. That implies that the function has
1235 internal linkage and only has one call site, so the original
1236 call is dead after inlining.
1238 This attributes disables implicit floating point instructions.
1240 This attribute indicates that the inliner should never inline this
1241 function in any situation. This attribute may not be used together
1242 with the ``alwaysinline`` attribute.
1244 This attribute suppresses lazy symbol binding for the function. This
1245 may make calls to the function faster, at the cost of extra program
1246 startup time if the function is not called during program startup.
1248 This attribute indicates that the code generator should not use a
1249 red zone, even if the target-specific ABI normally permits it.
1251 This function attribute indicates that the function never returns
1252 normally. This produces undefined behavior at runtime if the
1253 function ever does dynamically return.
1255 This function attribute indicates that the function never raises an
1256 exception. If the function does raise an exception, its runtime
1257 behavior is undefined. However, functions marked nounwind may still
1258 trap or generate asynchronous exceptions. Exception handling schemes
1259 that are recognized by LLVM to handle asynchronous exceptions, such
1260 as SEH, will still provide their implementation defined semantics.
1262 This function attribute indicates that the function is not optimized
1263 by any optimization or code generator passes with the
1264 exception of interprocedural optimization passes.
1265 This attribute cannot be used together with the ``alwaysinline``
1266 attribute; this attribute is also incompatible
1267 with the ``minsize`` attribute and the ``optsize`` attribute.
1269 This attribute requires the ``noinline`` attribute to be specified on
1270 the function as well, so the function is never inlined into any caller.
1271 Only functions with the ``alwaysinline`` attribute are valid
1272 candidates for inlining into the body of this function.
1274 This attribute suggests that optimization passes and code generator
1275 passes make choices that keep the code size of this function low,
1276 and otherwise do optimizations specifically to reduce code size as
1277 long as they do not significantly impact runtime performance.
1279 On a function, this attribute indicates that the function computes its
1280 result (or decides to unwind an exception) based strictly on its arguments,
1281 without dereferencing any pointer arguments or otherwise accessing
1282 any mutable state (e.g. memory, control registers, etc) visible to
1283 caller functions. It does not write through any pointer arguments
1284 (including ``byval`` arguments) and never changes any state visible
1285 to callers. This means that it cannot unwind exceptions by calling
1286 the ``C++`` exception throwing methods.
1288 On an argument, this attribute indicates that the function does not
1289 dereference that pointer argument, even though it may read or write the
1290 memory that the pointer points to if accessed through other pointers.
1292 On a function, this attribute indicates that the function does not write
1293 through any pointer arguments (including ``byval`` arguments) or otherwise
1294 modify any state (e.g. memory, control registers, etc) visible to
1295 caller functions. It may dereference pointer arguments and read
1296 state that may be set in the caller. A readonly function always
1297 returns the same value (or unwinds an exception identically) when
1298 called with the same set of arguments and global state. It cannot
1299 unwind an exception by calling the ``C++`` exception throwing
1302 On an argument, this attribute indicates that the function does not write
1303 through this pointer argument, even though it may write to the memory that
1304 the pointer points to.
1306 This attribute indicates that this function can return twice. The C
1307 ``setjmp`` is an example of such a function. The compiler disables
1308 some optimizations (like tail calls) in the caller of these
1310 ``sanitize_address``
1311 This attribute indicates that AddressSanitizer checks
1312 (dynamic address safety analysis) are enabled for this function.
1314 This attribute indicates that MemorySanitizer checks (dynamic detection
1315 of accesses to uninitialized memory) are enabled for this function.
1317 This attribute indicates that ThreadSanitizer checks
1318 (dynamic thread safety analysis) are enabled for this function.
1320 This attribute indicates that the function should emit a stack
1321 smashing protector. It is in the form of a "canary" --- a random value
1322 placed on the stack before the local variables that's checked upon
1323 return from the function to see if it has been overwritten. A
1324 heuristic is used to determine if a function needs stack protectors
1325 or not. The heuristic used will enable protectors for functions with:
1327 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1328 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1329 - Calls to alloca() with variable sizes or constant sizes greater than
1330 ``ssp-buffer-size``.
1332 Variables that are identified as requiring a protector will be arranged
1333 on the stack such that they are adjacent to the stack protector guard.
1335 If a function that has an ``ssp`` attribute is inlined into a
1336 function that doesn't have an ``ssp`` attribute, then the resulting
1337 function will have an ``ssp`` attribute.
1339 This attribute indicates that the function should *always* emit a
1340 stack smashing protector. This overrides the ``ssp`` function
1343 Variables that are identified as requiring a protector will be arranged
1344 on the stack such that they are adjacent to the stack protector guard.
1345 The specific layout rules are:
1347 #. Large arrays and structures containing large arrays
1348 (``>= ssp-buffer-size``) are closest to the stack protector.
1349 #. Small arrays and structures containing small arrays
1350 (``< ssp-buffer-size``) are 2nd closest to the protector.
1351 #. Variables that have had their address taken are 3rd closest to the
1354 If a function that has an ``sspreq`` attribute is inlined into a
1355 function that doesn't have an ``sspreq`` attribute or which has an
1356 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1357 an ``sspreq`` attribute.
1359 This attribute indicates that the function should emit a stack smashing
1360 protector. This attribute causes a strong heuristic to be used when
1361 determining if a function needs stack protectors. The strong heuristic
1362 will enable protectors for functions with:
1364 - Arrays of any size and type
1365 - Aggregates containing an array of any size and type.
1366 - Calls to alloca().
1367 - Local variables that have had their address taken.
1369 Variables that are identified as requiring a protector will be arranged
1370 on the stack such that they are adjacent to the stack protector guard.
1371 The specific layout rules are:
1373 #. Large arrays and structures containing large arrays
1374 (``>= ssp-buffer-size``) are closest to the stack protector.
1375 #. Small arrays and structures containing small arrays
1376 (``< ssp-buffer-size``) are 2nd closest to the protector.
1377 #. Variables that have had their address taken are 3rd closest to the
1380 This overrides the ``ssp`` function attribute.
1382 If a function that has an ``sspstrong`` attribute is inlined into a
1383 function that doesn't have an ``sspstrong`` attribute, then the
1384 resulting function will have an ``sspstrong`` attribute.
1386 This attribute indicates that the function will delegate to some other
1387 function with a tail call. The prototype of a thunk should not be used for
1388 optimization purposes. The caller is expected to cast the thunk prototype to
1389 match the thunk target prototype.
1391 This attribute indicates that the ABI being targeted requires that
1392 an unwind table entry be produce for this function even if we can
1393 show that no exceptions passes by it. This is normally the case for
1394 the ELF x86-64 abi, but it can be disabled for some compilation
1399 Module-Level Inline Assembly
1400 ----------------------------
1402 Modules may contain "module-level inline asm" blocks, which corresponds
1403 to the GCC "file scope inline asm" blocks. These blocks are internally
1404 concatenated by LLVM and treated as a single unit, but may be separated
1405 in the ``.ll`` file if desired. The syntax is very simple:
1407 .. code-block:: llvm
1409 module asm "inline asm code goes here"
1410 module asm "more can go here"
1412 The strings can contain any character by escaping non-printable
1413 characters. The escape sequence used is simply "\\xx" where "xx" is the
1414 two digit hex code for the number.
1416 The inline asm code is simply printed to the machine code .s file when
1417 assembly code is generated.
1419 .. _langref_datalayout:
1424 A module may specify a target specific data layout string that specifies
1425 how data is to be laid out in memory. The syntax for the data layout is
1428 .. code-block:: llvm
1430 target datalayout = "layout specification"
1432 The *layout specification* consists of a list of specifications
1433 separated by the minus sign character ('-'). Each specification starts
1434 with a letter and may include other information after the letter to
1435 define some aspect of the data layout. The specifications accepted are
1439 Specifies that the target lays out data in big-endian form. That is,
1440 the bits with the most significance have the lowest address
1443 Specifies that the target lays out data in little-endian form. That
1444 is, the bits with the least significance have the lowest address
1447 Specifies the natural alignment of the stack in bits. Alignment
1448 promotion of stack variables is limited to the natural stack
1449 alignment to avoid dynamic stack realignment. The stack alignment
1450 must be a multiple of 8-bits. If omitted, the natural stack
1451 alignment defaults to "unspecified", which does not prevent any
1452 alignment promotions.
1453 ``p[n]:<size>:<abi>:<pref>``
1454 This specifies the *size* of a pointer and its ``<abi>`` and
1455 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1456 bits. The address space, ``n`` is optional, and if not specified,
1457 denotes the default address space 0. The value of ``n`` must be
1458 in the range [1,2^23).
1459 ``i<size>:<abi>:<pref>``
1460 This specifies the alignment for an integer type of a given bit
1461 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1462 ``v<size>:<abi>:<pref>``
1463 This specifies the alignment for a vector type of a given bit
1465 ``f<size>:<abi>:<pref>``
1466 This specifies the alignment for a floating point type of a given bit
1467 ``<size>``. Only values of ``<size>`` that are supported by the target
1468 will work. 32 (float) and 64 (double) are supported on all targets; 80
1469 or 128 (different flavors of long double) are also supported on some
1472 This specifies the alignment for an object of aggregate type.
1474 If present, specifies that llvm names are mangled in the output. The
1477 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1478 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1479 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1480 symbols get a ``_`` prefix.
1481 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1482 functions also get a suffix based on the frame size.
1483 ``n<size1>:<size2>:<size3>...``
1484 This specifies a set of native integer widths for the target CPU in
1485 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1486 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1487 this set are considered to support most general arithmetic operations
1490 On every specification that takes a ``<abi>:<pref>``, specifying the
1491 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1492 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1494 When constructing the data layout for a given target, LLVM starts with a
1495 default set of specifications which are then (possibly) overridden by
1496 the specifications in the ``datalayout`` keyword. The default
1497 specifications are given in this list:
1499 - ``E`` - big endian
1500 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1501 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1502 same as the default address space.
1503 - ``S0`` - natural stack alignment is unspecified
1504 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1505 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1506 - ``i16:16:16`` - i16 is 16-bit aligned
1507 - ``i32:32:32`` - i32 is 32-bit aligned
1508 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1509 alignment of 64-bits
1510 - ``f16:16:16`` - half is 16-bit aligned
1511 - ``f32:32:32`` - float is 32-bit aligned
1512 - ``f64:64:64`` - double is 64-bit aligned
1513 - ``f128:128:128`` - quad is 128-bit aligned
1514 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1515 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1516 - ``a:0:64`` - aggregates are 64-bit aligned
1518 When LLVM is determining the alignment for a given type, it uses the
1521 #. If the type sought is an exact match for one of the specifications,
1522 that specification is used.
1523 #. If no match is found, and the type sought is an integer type, then
1524 the smallest integer type that is larger than the bitwidth of the
1525 sought type is used. If none of the specifications are larger than
1526 the bitwidth then the largest integer type is used. For example,
1527 given the default specifications above, the i7 type will use the
1528 alignment of i8 (next largest) while both i65 and i256 will use the
1529 alignment of i64 (largest specified).
1530 #. If no match is found, and the type sought is a vector type, then the
1531 largest vector type that is smaller than the sought vector type will
1532 be used as a fall back. This happens because <128 x double> can be
1533 implemented in terms of 64 <2 x double>, for example.
1535 The function of the data layout string may not be what you expect.
1536 Notably, this is not a specification from the frontend of what alignment
1537 the code generator should use.
1539 Instead, if specified, the target data layout is required to match what
1540 the ultimate *code generator* expects. This string is used by the
1541 mid-level optimizers to improve code, and this only works if it matches
1542 what the ultimate code generator uses. There is no way to generate IR
1543 that does not embed this target-specific detail into the IR. If you
1544 don't specify the string, the default specifications will be used to
1545 generate a Data Layout and the optimization phases will operate
1546 accordingly and introduce target specificity into the IR with respect to
1547 these default specifications.
1554 A module may specify a target triple string that describes the target
1555 host. The syntax for the target triple is simply:
1557 .. code-block:: llvm
1559 target triple = "x86_64-apple-macosx10.7.0"
1561 The *target triple* string consists of a series of identifiers delimited
1562 by the minus sign character ('-'). The canonical forms are:
1566 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1567 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1569 This information is passed along to the backend so that it generates
1570 code for the proper architecture. It's possible to override this on the
1571 command line with the ``-mtriple`` command line option.
1573 .. _pointeraliasing:
1575 Pointer Aliasing Rules
1576 ----------------------
1578 Any memory access must be done through a pointer value associated with
1579 an address range of the memory access, otherwise the behavior is
1580 undefined. Pointer values are associated with address ranges according
1581 to the following rules:
1583 - A pointer value is associated with the addresses associated with any
1584 value it is *based* on.
1585 - An address of a global variable is associated with the address range
1586 of the variable's storage.
1587 - The result value of an allocation instruction is associated with the
1588 address range of the allocated storage.
1589 - A null pointer in the default address-space is associated with no
1591 - An integer constant other than zero or a pointer value returned from
1592 a function not defined within LLVM may be associated with address
1593 ranges allocated through mechanisms other than those provided by
1594 LLVM. Such ranges shall not overlap with any ranges of addresses
1595 allocated by mechanisms provided by LLVM.
1597 A pointer value is *based* on another pointer value according to the
1600 - A pointer value formed from a ``getelementptr`` operation is *based*
1601 on the first value operand of the ``getelementptr``.
1602 - The result value of a ``bitcast`` is *based* on the operand of the
1604 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1605 values that contribute (directly or indirectly) to the computation of
1606 the pointer's value.
1607 - The "*based* on" relationship is transitive.
1609 Note that this definition of *"based"* is intentionally similar to the
1610 definition of *"based"* in C99, though it is slightly weaker.
1612 LLVM IR does not associate types with memory. The result type of a
1613 ``load`` merely indicates the size and alignment of the memory from
1614 which to load, as well as the interpretation of the value. The first
1615 operand type of a ``store`` similarly only indicates the size and
1616 alignment of the store.
1618 Consequently, type-based alias analysis, aka TBAA, aka
1619 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1620 :ref:`Metadata <metadata>` may be used to encode additional information
1621 which specialized optimization passes may use to implement type-based
1626 Volatile Memory Accesses
1627 ------------------------
1629 Certain memory accesses, such as :ref:`load <i_load>`'s,
1630 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1631 marked ``volatile``. The optimizers must not change the number of
1632 volatile operations or change their order of execution relative to other
1633 volatile operations. The optimizers *may* change the order of volatile
1634 operations relative to non-volatile operations. This is not Java's
1635 "volatile" and has no cross-thread synchronization behavior.
1637 IR-level volatile loads and stores cannot safely be optimized into
1638 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1639 flagged volatile. Likewise, the backend should never split or merge
1640 target-legal volatile load/store instructions.
1642 .. admonition:: Rationale
1644 Platforms may rely on volatile loads and stores of natively supported
1645 data width to be executed as single instruction. For example, in C
1646 this holds for an l-value of volatile primitive type with native
1647 hardware support, but not necessarily for aggregate types. The
1648 frontend upholds these expectations, which are intentionally
1649 unspecified in the IR. The rules above ensure that IR transformation
1650 do not violate the frontend's contract with the language.
1654 Memory Model for Concurrent Operations
1655 --------------------------------------
1657 The LLVM IR does not define any way to start parallel threads of
1658 execution or to register signal handlers. Nonetheless, there are
1659 platform-specific ways to create them, and we define LLVM IR's behavior
1660 in their presence. This model is inspired by the C++0x memory model.
1662 For a more informal introduction to this model, see the :doc:`Atomics`.
1664 We define a *happens-before* partial order as the least partial order
1667 - Is a superset of single-thread program order, and
1668 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1669 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1670 techniques, like pthread locks, thread creation, thread joining,
1671 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1672 Constraints <ordering>`).
1674 Note that program order does not introduce *happens-before* edges
1675 between a thread and signals executing inside that thread.
1677 Every (defined) read operation (load instructions, memcpy, atomic
1678 loads/read-modify-writes, etc.) R reads a series of bytes written by
1679 (defined) write operations (store instructions, atomic
1680 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1681 section, initialized globals are considered to have a write of the
1682 initializer which is atomic and happens before any other read or write
1683 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1684 may see any write to the same byte, except:
1686 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1687 write\ :sub:`2` happens before R\ :sub:`byte`, then
1688 R\ :sub:`byte` does not see write\ :sub:`1`.
1689 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1690 R\ :sub:`byte` does not see write\ :sub:`3`.
1692 Given that definition, R\ :sub:`byte` is defined as follows:
1694 - If R is volatile, the result is target-dependent. (Volatile is
1695 supposed to give guarantees which can support ``sig_atomic_t`` in
1696 C/C++, and may be used for accesses to addresses that do not behave
1697 like normal memory. It does not generally provide cross-thread
1699 - Otherwise, if there is no write to the same byte that happens before
1700 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1701 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1702 R\ :sub:`byte` returns the value written by that write.
1703 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1704 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1705 Memory Ordering Constraints <ordering>` section for additional
1706 constraints on how the choice is made.
1707 - Otherwise R\ :sub:`byte` returns ``undef``.
1709 R returns the value composed of the series of bytes it read. This
1710 implies that some bytes within the value may be ``undef`` **without**
1711 the entire value being ``undef``. Note that this only defines the
1712 semantics of the operation; it doesn't mean that targets will emit more
1713 than one instruction to read the series of bytes.
1715 Note that in cases where none of the atomic intrinsics are used, this
1716 model places only one restriction on IR transformations on top of what
1717 is required for single-threaded execution: introducing a store to a byte
1718 which might not otherwise be stored is not allowed in general.
1719 (Specifically, in the case where another thread might write to and read
1720 from an address, introducing a store can change a load that may see
1721 exactly one write into a load that may see multiple writes.)
1725 Atomic Memory Ordering Constraints
1726 ----------------------------------
1728 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1729 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1730 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1731 ordering parameters that determine which other atomic instructions on
1732 the same address they *synchronize with*. These semantics are borrowed
1733 from Java and C++0x, but are somewhat more colloquial. If these
1734 descriptions aren't precise enough, check those specs (see spec
1735 references in the :doc:`atomics guide <Atomics>`).
1736 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1737 differently since they don't take an address. See that instruction's
1738 documentation for details.
1740 For a simpler introduction to the ordering constraints, see the
1744 The set of values that can be read is governed by the happens-before
1745 partial order. A value cannot be read unless some operation wrote
1746 it. This is intended to provide a guarantee strong enough to model
1747 Java's non-volatile shared variables. This ordering cannot be
1748 specified for read-modify-write operations; it is not strong enough
1749 to make them atomic in any interesting way.
1751 In addition to the guarantees of ``unordered``, there is a single
1752 total order for modifications by ``monotonic`` operations on each
1753 address. All modification orders must be compatible with the
1754 happens-before order. There is no guarantee that the modification
1755 orders can be combined to a global total order for the whole program
1756 (and this often will not be possible). The read in an atomic
1757 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1758 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1759 order immediately before the value it writes. If one atomic read
1760 happens before another atomic read of the same address, the later
1761 read must see the same value or a later value in the address's
1762 modification order. This disallows reordering of ``monotonic`` (or
1763 stronger) operations on the same address. If an address is written
1764 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1765 read that address repeatedly, the other threads must eventually see
1766 the write. This corresponds to the C++0x/C1x
1767 ``memory_order_relaxed``.
1769 In addition to the guarantees of ``monotonic``, a
1770 *synchronizes-with* edge may be formed with a ``release`` operation.
1771 This is intended to model C++'s ``memory_order_acquire``.
1773 In addition to the guarantees of ``monotonic``, if this operation
1774 writes a value which is subsequently read by an ``acquire``
1775 operation, it *synchronizes-with* that operation. (This isn't a
1776 complete description; see the C++0x definition of a release
1777 sequence.) This corresponds to the C++0x/C1x
1778 ``memory_order_release``.
1779 ``acq_rel`` (acquire+release)
1780 Acts as both an ``acquire`` and ``release`` operation on its
1781 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1782 ``seq_cst`` (sequentially consistent)
1783 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1784 operation that only reads, ``release`` for an operation that only
1785 writes), there is a global total order on all
1786 sequentially-consistent operations on all addresses, which is
1787 consistent with the *happens-before* partial order and with the
1788 modification orders of all the affected addresses. Each
1789 sequentially-consistent read sees the last preceding write to the
1790 same address in this global order. This corresponds to the C++0x/C1x
1791 ``memory_order_seq_cst`` and Java volatile.
1795 If an atomic operation is marked ``singlethread``, it only *synchronizes
1796 with* or participates in modification and seq\_cst total orderings with
1797 other operations running in the same thread (for example, in signal
1805 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1806 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1807 :ref:`frem <i_frem>`) have the following flags that can be set to enable
1808 otherwise unsafe floating point operations
1811 No NaNs - Allow optimizations to assume the arguments and result are not
1812 NaN. Such optimizations are required to retain defined behavior over
1813 NaNs, but the value of the result is undefined.
1816 No Infs - Allow optimizations to assume the arguments and result are not
1817 +/-Inf. Such optimizations are required to retain defined behavior over
1818 +/-Inf, but the value of the result is undefined.
1821 No Signed Zeros - Allow optimizations to treat the sign of a zero
1822 argument or result as insignificant.
1825 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1826 argument rather than perform division.
1829 Fast - Allow algebraically equivalent transformations that may
1830 dramatically change results in floating point (e.g. reassociate). This
1831 flag implies all the others.
1835 Use-list Order Directives
1836 -------------------------
1838 Use-list directives encode the in-memory order of each use-list, allowing the
1839 order to be recreated. ``<order-indexes>`` is a comma-separated list of
1840 indexes that are assigned to the referenced value's uses. The referenced
1841 value's use-list is immediately sorted by these indexes.
1843 Use-list directives may appear at function scope or global scope. They are not
1844 instructions, and have no effect on the semantics of the IR. When they're at
1845 function scope, they must appear after the terminator of the final basic block.
1847 If basic blocks have their address taken via ``blockaddress()`` expressions,
1848 ``uselistorder_bb`` can be used to reorder their use-lists from outside their
1855 uselistorder <ty> <value>, { <order-indexes> }
1856 uselistorder_bb @function, %block { <order-indexes> }
1862 define void @foo(i32 %arg1, i32 %arg2) {
1864 ; ... instructions ...
1866 ; ... instructions ...
1868 ; At function scope.
1869 uselistorder i32 %arg1, { 1, 0, 2 }
1870 uselistorder label %bb, { 1, 0 }
1874 uselistorder i32* @global, { 1, 2, 0 }
1875 uselistorder i32 7, { 1, 0 }
1876 uselistorder i32 (i32) @bar, { 1, 0 }
1877 uselistorder_bb @foo, %bb, { 5, 1, 3, 2, 0, 4 }
1884 The LLVM type system is one of the most important features of the
1885 intermediate representation. Being typed enables a number of
1886 optimizations to be performed on the intermediate representation
1887 directly, without having to do extra analyses on the side before the
1888 transformation. A strong type system makes it easier to read the
1889 generated code and enables novel analyses and transformations that are
1890 not feasible to perform on normal three address code representations.
1900 The void type does not represent any value and has no size.
1918 The function type can be thought of as a function signature. It consists of a
1919 return type and a list of formal parameter types. The return type of a function
1920 type is a void type or first class type --- except for :ref:`label <t_label>`
1921 and :ref:`metadata <t_metadata>` types.
1927 <returntype> (<parameter list>)
1929 ...where '``<parameter list>``' is a comma-separated list of type
1930 specifiers. Optionally, the parameter list may include a type ``...``, which
1931 indicates that the function takes a variable number of arguments. Variable
1932 argument functions can access their arguments with the :ref:`variable argument
1933 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
1934 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
1938 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1939 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1940 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1941 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1942 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1943 | ``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. |
1944 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1945 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1946 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1953 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1954 Values of these types are the only ones which can be produced by
1962 These are the types that are valid in registers from CodeGen's perspective.
1971 The integer type is a very simple type that simply specifies an
1972 arbitrary bit width for the integer type desired. Any bit width from 1
1973 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1981 The number of bits the integer will occupy is specified by the ``N``
1987 +----------------+------------------------------------------------+
1988 | ``i1`` | a single-bit integer. |
1989 +----------------+------------------------------------------------+
1990 | ``i32`` | a 32-bit integer. |
1991 +----------------+------------------------------------------------+
1992 | ``i1942652`` | a really big integer of over 1 million bits. |
1993 +----------------+------------------------------------------------+
1997 Floating Point Types
1998 """"""""""""""""""""
2007 - 16-bit floating point value
2010 - 32-bit floating point value
2013 - 64-bit floating point value
2016 - 128-bit floating point value (112-bit mantissa)
2019 - 80-bit floating point value (X87)
2022 - 128-bit floating point value (two 64-bits)
2029 The x86_mmx type represents a value held in an MMX register on an x86
2030 machine. The operations allowed on it are quite limited: parameters and
2031 return values, load and store, and bitcast. User-specified MMX
2032 instructions are represented as intrinsic or asm calls with arguments
2033 and/or results of this type. There are no arrays, vectors or constants
2050 The pointer type is used to specify memory locations. Pointers are
2051 commonly used to reference objects in memory.
2053 Pointer types may have an optional address space attribute defining the
2054 numbered address space where the pointed-to object resides. The default
2055 address space is number zero. The semantics of non-zero address spaces
2056 are target-specific.
2058 Note that LLVM does not permit pointers to void (``void*``) nor does it
2059 permit pointers to labels (``label*``). Use ``i8*`` instead.
2069 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2070 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
2071 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2072 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
2073 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2074 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
2075 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2084 A vector type is a simple derived type that represents a vector of
2085 elements. Vector types are used when multiple primitive data are
2086 operated in parallel using a single instruction (SIMD). A vector type
2087 requires a size (number of elements) and an underlying primitive data
2088 type. Vector types are considered :ref:`first class <t_firstclass>`.
2094 < <# elements> x <elementtype> >
2096 The number of elements is a constant integer value larger than 0;
2097 elementtype may be any integer, floating point or pointer type. Vectors
2098 of size zero are not allowed.
2102 +-------------------+--------------------------------------------------+
2103 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
2104 +-------------------+--------------------------------------------------+
2105 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
2106 +-------------------+--------------------------------------------------+
2107 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
2108 +-------------------+--------------------------------------------------+
2109 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
2110 +-------------------+--------------------------------------------------+
2119 The label type represents code labels.
2134 The metadata type represents embedded metadata. No derived types may be
2135 created from metadata except for :ref:`function <t_function>` arguments.
2148 Aggregate Types are a subset of derived types that can contain multiple
2149 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
2150 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
2160 The array type is a very simple derived type that arranges elements
2161 sequentially in memory. The array type requires a size (number of
2162 elements) and an underlying data type.
2168 [<# elements> x <elementtype>]
2170 The number of elements is a constant integer value; ``elementtype`` may
2171 be any type with a size.
2175 +------------------+--------------------------------------+
2176 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
2177 +------------------+--------------------------------------+
2178 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
2179 +------------------+--------------------------------------+
2180 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
2181 +------------------+--------------------------------------+
2183 Here are some examples of multidimensional arrays:
2185 +-----------------------------+----------------------------------------------------------+
2186 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
2187 +-----------------------------+----------------------------------------------------------+
2188 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
2189 +-----------------------------+----------------------------------------------------------+
2190 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
2191 +-----------------------------+----------------------------------------------------------+
2193 There is no restriction on indexing beyond the end of the array implied
2194 by a static type (though there are restrictions on indexing beyond the
2195 bounds of an allocated object in some cases). This means that
2196 single-dimension 'variable sized array' addressing can be implemented in
2197 LLVM with a zero length array type. An implementation of 'pascal style
2198 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
2208 The structure type is used to represent a collection of data members
2209 together in memory. The elements of a structure may be any type that has
2212 Structures in memory are accessed using '``load``' and '``store``' by
2213 getting a pointer to a field with the '``getelementptr``' instruction.
2214 Structures in registers are accessed using the '``extractvalue``' and
2215 '``insertvalue``' instructions.
2217 Structures may optionally be "packed" structures, which indicate that
2218 the alignment of the struct is one byte, and that there is no padding
2219 between the elements. In non-packed structs, padding between field types
2220 is inserted as defined by the DataLayout string in the module, which is
2221 required to match what the underlying code generator expects.
2223 Structures can either be "literal" or "identified". A literal structure
2224 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
2225 identified types are always defined at the top level with a name.
2226 Literal types are uniqued by their contents and can never be recursive
2227 or opaque since there is no way to write one. Identified types can be
2228 recursive, can be opaqued, and are never uniqued.
2234 %T1 = type { <type list> } ; Identified normal struct type
2235 %T2 = type <{ <type list> }> ; Identified packed struct type
2239 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2240 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
2241 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2242 | ``{ 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``. |
2243 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2244 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
2245 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2249 Opaque Structure Types
2250 """"""""""""""""""""""
2254 Opaque structure types are used to represent named structure types that
2255 do not have a body specified. This corresponds (for example) to the C
2256 notion of a forward declared structure.
2267 +--------------+-------------------+
2268 | ``opaque`` | An opaque type. |
2269 +--------------+-------------------+
2276 LLVM has several different basic types of constants. This section
2277 describes them all and their syntax.
2282 **Boolean constants**
2283 The two strings '``true``' and '``false``' are both valid constants
2285 **Integer constants**
2286 Standard integers (such as '4') are constants of the
2287 :ref:`integer <t_integer>` type. Negative numbers may be used with
2289 **Floating point constants**
2290 Floating point constants use standard decimal notation (e.g.
2291 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2292 hexadecimal notation (see below). The assembler requires the exact
2293 decimal value of a floating-point constant. For example, the
2294 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2295 decimal in binary. Floating point constants must have a :ref:`floating
2296 point <t_floating>` type.
2297 **Null pointer constants**
2298 The identifier '``null``' is recognized as a null pointer constant
2299 and must be of :ref:`pointer type <t_pointer>`.
2301 The one non-intuitive notation for constants is the hexadecimal form of
2302 floating point constants. For example, the form
2303 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2304 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2305 constants are required (and the only time that they are generated by the
2306 disassembler) is when a floating point constant must be emitted but it
2307 cannot be represented as a decimal floating point number in a reasonable
2308 number of digits. For example, NaN's, infinities, and other special
2309 values are represented in their IEEE hexadecimal format so that assembly
2310 and disassembly do not cause any bits to change in the constants.
2312 When using the hexadecimal form, constants of types half, float, and
2313 double are represented using the 16-digit form shown above (which
2314 matches the IEEE754 representation for double); half and float values
2315 must, however, be exactly representable as IEEE 754 half and single
2316 precision, respectively. Hexadecimal format is always used for long
2317 double, and there are three forms of long double. The 80-bit format used
2318 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2319 128-bit format used by PowerPC (two adjacent doubles) is represented by
2320 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2321 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2322 will only work if they match the long double format on your target.
2323 The IEEE 16-bit format (half precision) is represented by ``0xH``
2324 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2325 (sign bit at the left).
2327 There are no constants of type x86_mmx.
2329 .. _complexconstants:
2334 Complex constants are a (potentially recursive) combination of simple
2335 constants and smaller complex constants.
2337 **Structure constants**
2338 Structure constants are represented with notation similar to
2339 structure type definitions (a comma separated list of elements,
2340 surrounded by braces (``{}``)). For example:
2341 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2342 "``@G = external global i32``". Structure constants must have
2343 :ref:`structure type <t_struct>`, and the number and types of elements
2344 must match those specified by the type.
2346 Array constants are represented with notation similar to array type
2347 definitions (a comma separated list of elements, surrounded by
2348 square brackets (``[]``)). For example:
2349 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2350 :ref:`array type <t_array>`, and the number and types of elements must
2351 match those specified by the type. As a special case, character array
2352 constants may also be represented as a double-quoted string using the ``c``
2353 prefix. For example: "``c"Hello World\0A\00"``".
2354 **Vector constants**
2355 Vector constants are represented with notation similar to vector
2356 type definitions (a comma separated list of elements, surrounded by
2357 less-than/greater-than's (``<>``)). For example:
2358 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2359 must have :ref:`vector type <t_vector>`, and the number and types of
2360 elements must match those specified by the type.
2361 **Zero initialization**
2362 The string '``zeroinitializer``' can be used to zero initialize a
2363 value to zero of *any* type, including scalar and
2364 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2365 having to print large zero initializers (e.g. for large arrays) and
2366 is always exactly equivalent to using explicit zero initializers.
2368 A metadata node is a constant tuple without types. For example:
2369 "``!{!0, !{!2, !0}, !"test"}``". Metadata can reference constant values,
2370 for example: "``!{!0, i32 0, i8* @global, i64 (i64)* @function, !"str"}``".
2371 Unlike other typed constants that are meant to be interpreted as part of
2372 the instruction stream, metadata is a place to attach additional
2373 information such as debug info.
2375 Global Variable and Function Addresses
2376 --------------------------------------
2378 The addresses of :ref:`global variables <globalvars>` and
2379 :ref:`functions <functionstructure>` are always implicitly valid
2380 (link-time) constants. These constants are explicitly referenced when
2381 the :ref:`identifier for the global <identifiers>` is used and always have
2382 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2385 .. code-block:: llvm
2389 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2396 The string '``undef``' can be used anywhere a constant is expected, and
2397 indicates that the user of the value may receive an unspecified
2398 bit-pattern. Undefined values may be of any type (other than '``label``'
2399 or '``void``') and be used anywhere a constant is permitted.
2401 Undefined values are useful because they indicate to the compiler that
2402 the program is well defined no matter what value is used. This gives the
2403 compiler more freedom to optimize. Here are some examples of
2404 (potentially surprising) transformations that are valid (in pseudo IR):
2406 .. code-block:: llvm
2416 This is safe because all of the output bits are affected by the undef
2417 bits. Any output bit can have a zero or one depending on the input bits.
2419 .. code-block:: llvm
2430 These logical operations have bits that are not always affected by the
2431 input. For example, if ``%X`` has a zero bit, then the output of the
2432 '``and``' operation will always be a zero for that bit, no matter what
2433 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2434 optimize or assume that the result of the '``and``' is '``undef``'.
2435 However, it is safe to assume that all bits of the '``undef``' could be
2436 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2437 all the bits of the '``undef``' operand to the '``or``' could be set,
2438 allowing the '``or``' to be folded to -1.
2440 .. code-block:: llvm
2442 %A = select undef, %X, %Y
2443 %B = select undef, 42, %Y
2444 %C = select %X, %Y, undef
2454 This set of examples shows that undefined '``select``' (and conditional
2455 branch) conditions can go *either way*, but they have to come from one
2456 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2457 both known to have a clear low bit, then ``%A`` would have to have a
2458 cleared low bit. However, in the ``%C`` example, the optimizer is
2459 allowed to assume that the '``undef``' operand could be the same as
2460 ``%Y``, allowing the whole '``select``' to be eliminated.
2462 .. code-block:: llvm
2464 %A = xor undef, undef
2481 This example points out that two '``undef``' operands are not
2482 necessarily the same. This can be surprising to people (and also matches
2483 C semantics) where they assume that "``X^X``" is always zero, even if
2484 ``X`` is undefined. This isn't true for a number of reasons, but the
2485 short answer is that an '``undef``' "variable" can arbitrarily change
2486 its value over its "live range". This is true because the variable
2487 doesn't actually *have a live range*. Instead, the value is logically
2488 read from arbitrary registers that happen to be around when needed, so
2489 the value is not necessarily consistent over time. In fact, ``%A`` and
2490 ``%C`` need to have the same semantics or the core LLVM "replace all
2491 uses with" concept would not hold.
2493 .. code-block:: llvm
2501 These examples show the crucial difference between an *undefined value*
2502 and *undefined behavior*. An undefined value (like '``undef``') is
2503 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2504 operation can be constant folded to '``undef``', because the '``undef``'
2505 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2506 However, in the second example, we can make a more aggressive
2507 assumption: because the ``undef`` is allowed to be an arbitrary value,
2508 we are allowed to assume that it could be zero. Since a divide by zero
2509 has *undefined behavior*, we are allowed to assume that the operation
2510 does not execute at all. This allows us to delete the divide and all
2511 code after it. Because the undefined operation "can't happen", the
2512 optimizer can assume that it occurs in dead code.
2514 .. code-block:: llvm
2516 a: store undef -> %X
2517 b: store %X -> undef
2522 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2523 value can be assumed to not have any effect; we can assume that the
2524 value is overwritten with bits that happen to match what was already
2525 there. However, a store *to* an undefined location could clobber
2526 arbitrary memory, therefore, it has undefined behavior.
2533 Poison values are similar to :ref:`undef values <undefvalues>`, however
2534 they also represent the fact that an instruction or constant expression
2535 that cannot evoke side effects has nevertheless detected a condition
2536 that results in undefined behavior.
2538 There is currently no way of representing a poison value in the IR; they
2539 only exist when produced by operations such as :ref:`add <i_add>` with
2542 Poison value behavior is defined in terms of value *dependence*:
2544 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2545 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2546 their dynamic predecessor basic block.
2547 - Function arguments depend on the corresponding actual argument values
2548 in the dynamic callers of their functions.
2549 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2550 instructions that dynamically transfer control back to them.
2551 - :ref:`Invoke <i_invoke>` instructions depend on the
2552 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2553 call instructions that dynamically transfer control back to them.
2554 - Non-volatile loads and stores depend on the most recent stores to all
2555 of the referenced memory addresses, following the order in the IR
2556 (including loads and stores implied by intrinsics such as
2557 :ref:`@llvm.memcpy <int_memcpy>`.)
2558 - An instruction with externally visible side effects depends on the
2559 most recent preceding instruction with externally visible side
2560 effects, following the order in the IR. (This includes :ref:`volatile
2561 operations <volatile>`.)
2562 - An instruction *control-depends* on a :ref:`terminator
2563 instruction <terminators>` if the terminator instruction has
2564 multiple successors and the instruction is always executed when
2565 control transfers to one of the successors, and may not be executed
2566 when control is transferred to another.
2567 - Additionally, an instruction also *control-depends* on a terminator
2568 instruction if the set of instructions it otherwise depends on would
2569 be different if the terminator had transferred control to a different
2571 - Dependence is transitive.
2573 Poison values have the same behavior as :ref:`undef values <undefvalues>`,
2574 with the additional effect that any instruction that has a *dependence*
2575 on a poison value has undefined behavior.
2577 Here are some examples:
2579 .. code-block:: llvm
2582 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2583 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2584 %poison_yet_again = getelementptr i32, i32* @h, i32 %still_poison
2585 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2587 store i32 %poison, i32* @g ; Poison value stored to memory.
2588 %poison2 = load i32, i32* @g ; Poison value loaded back from memory.
2590 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2592 %narrowaddr = bitcast i32* @g to i16*
2593 %wideaddr = bitcast i32* @g to i64*
2594 %poison3 = load i16, i16* %narrowaddr ; Returns a poison value.
2595 %poison4 = load i64, i64* %wideaddr ; Returns a poison value.
2597 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2598 br i1 %cmp, label %true, label %end ; Branch to either destination.
2601 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2602 ; it has undefined behavior.
2606 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2607 ; Both edges into this PHI are
2608 ; control-dependent on %cmp, so this
2609 ; always results in a poison value.
2611 store volatile i32 0, i32* @g ; This would depend on the store in %true
2612 ; if %cmp is true, or the store in %entry
2613 ; otherwise, so this is undefined behavior.
2615 br i1 %cmp, label %second_true, label %second_end
2616 ; The same branch again, but this time the
2617 ; true block doesn't have side effects.
2624 store volatile i32 0, i32* @g ; This time, the instruction always depends
2625 ; on the store in %end. Also, it is
2626 ; control-equivalent to %end, so this is
2627 ; well-defined (ignoring earlier undefined
2628 ; behavior in this example).
2632 Addresses of Basic Blocks
2633 -------------------------
2635 ``blockaddress(@function, %block)``
2637 The '``blockaddress``' constant computes the address of the specified
2638 basic block in the specified function, and always has an ``i8*`` type.
2639 Taking the address of the entry block is illegal.
2641 This value only has defined behavior when used as an operand to the
2642 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2643 against null. Pointer equality tests between labels addresses results in
2644 undefined behavior --- though, again, comparison against null is ok, and
2645 no label is equal to the null pointer. This may be passed around as an
2646 opaque pointer sized value as long as the bits are not inspected. This
2647 allows ``ptrtoint`` and arithmetic to be performed on these values so
2648 long as the original value is reconstituted before the ``indirectbr``
2651 Finally, some targets may provide defined semantics when using the value
2652 as the operand to an inline assembly, but that is target specific.
2656 Constant Expressions
2657 --------------------
2659 Constant expressions are used to allow expressions involving other
2660 constants to be used as constants. Constant expressions may be of any
2661 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2662 that does not have side effects (e.g. load and call are not supported).
2663 The following is the syntax for constant expressions:
2665 ``trunc (CST to TYPE)``
2666 Truncate a constant to another type. The bit size of CST must be
2667 larger than the bit size of TYPE. Both types must be integers.
2668 ``zext (CST to TYPE)``
2669 Zero extend a constant to another type. The bit size of CST must be
2670 smaller than the bit size of TYPE. Both types must be integers.
2671 ``sext (CST to TYPE)``
2672 Sign extend a constant to another type. The bit size of CST must be
2673 smaller than the bit size of TYPE. Both types must be integers.
2674 ``fptrunc (CST to TYPE)``
2675 Truncate a floating point constant to another floating point type.
2676 The size of CST must be larger than the size of TYPE. Both types
2677 must be floating point.
2678 ``fpext (CST to TYPE)``
2679 Floating point extend a constant to another type. The size of CST
2680 must be smaller or equal to the size of TYPE. Both types must be
2682 ``fptoui (CST to TYPE)``
2683 Convert a floating point constant to the corresponding unsigned
2684 integer constant. TYPE must be a scalar or vector integer type. CST
2685 must be of scalar or vector floating point type. Both CST and TYPE
2686 must be scalars, or vectors of the same number of elements. If the
2687 value won't fit in the integer type, the results are undefined.
2688 ``fptosi (CST to TYPE)``
2689 Convert a floating point constant to the corresponding signed
2690 integer constant. TYPE must be a scalar or vector integer type. CST
2691 must be of scalar or vector floating point type. Both CST and TYPE
2692 must be scalars, or vectors of the same number of elements. If the
2693 value won't fit in the integer type, the results are undefined.
2694 ``uitofp (CST to TYPE)``
2695 Convert an unsigned integer constant to the corresponding floating
2696 point constant. TYPE must be a scalar or vector floating point type.
2697 CST must be of scalar or vector integer type. Both CST and TYPE must
2698 be scalars, or vectors of the same number of elements. If the value
2699 won't fit in the floating point type, the results are undefined.
2700 ``sitofp (CST to TYPE)``
2701 Convert a signed integer constant to the corresponding floating
2702 point constant. TYPE must be a scalar or vector floating point type.
2703 CST must be of scalar or vector integer type. Both CST and TYPE must
2704 be scalars, or vectors of the same number of elements. If the value
2705 won't fit in the floating point type, the results are undefined.
2706 ``ptrtoint (CST to TYPE)``
2707 Convert a pointer typed constant to the corresponding integer
2708 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2709 pointer type. The ``CST`` value is zero extended, truncated, or
2710 unchanged to make it fit in ``TYPE``.
2711 ``inttoptr (CST to TYPE)``
2712 Convert an integer constant to a pointer constant. TYPE must be a
2713 pointer type. CST must be of integer type. The CST value is zero
2714 extended, truncated, or unchanged to make it fit in a pointer size.
2715 This one is *really* dangerous!
2716 ``bitcast (CST to TYPE)``
2717 Convert a constant, CST, to another TYPE. The constraints of the
2718 operands are the same as those for the :ref:`bitcast
2719 instruction <i_bitcast>`.
2720 ``addrspacecast (CST to TYPE)``
2721 Convert a constant pointer or constant vector of pointer, CST, to another
2722 TYPE in a different address space. The constraints of the operands are the
2723 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2724 ``getelementptr (TY, CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (TY, CSTPTR, IDX0, IDX1, ...)``
2725 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2726 constants. As with the :ref:`getelementptr <i_getelementptr>`
2727 instruction, the index list may have zero or more indexes, which are
2728 required to make sense for the type of "pointer to TY".
2729 ``select (COND, VAL1, VAL2)``
2730 Perform the :ref:`select operation <i_select>` on constants.
2731 ``icmp COND (VAL1, VAL2)``
2732 Performs the :ref:`icmp operation <i_icmp>` on constants.
2733 ``fcmp COND (VAL1, VAL2)``
2734 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2735 ``extractelement (VAL, IDX)``
2736 Perform the :ref:`extractelement operation <i_extractelement>` on
2738 ``insertelement (VAL, ELT, IDX)``
2739 Perform the :ref:`insertelement operation <i_insertelement>` on
2741 ``shufflevector (VEC1, VEC2, IDXMASK)``
2742 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2744 ``extractvalue (VAL, IDX0, IDX1, ...)``
2745 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2746 constants. The index list is interpreted in a similar manner as
2747 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2748 least one index value must be specified.
2749 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2750 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2751 The index list is interpreted in a similar manner as indices in a
2752 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2753 value must be specified.
2754 ``OPCODE (LHS, RHS)``
2755 Perform the specified operation of the LHS and RHS constants. OPCODE
2756 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2757 binary <bitwiseops>` operations. The constraints on operands are
2758 the same as those for the corresponding instruction (e.g. no bitwise
2759 operations on floating point values are allowed).
2766 Inline Assembler Expressions
2767 ----------------------------
2769 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2770 Inline Assembly <moduleasm>`) through the use of a special value. This
2771 value represents the inline assembler as a string (containing the
2772 instructions to emit), a list of operand constraints (stored as a
2773 string), a flag that indicates whether or not the inline asm expression
2774 has side effects, and a flag indicating whether the function containing
2775 the asm needs to align its stack conservatively. An example inline
2776 assembler expression is:
2778 .. code-block:: llvm
2780 i32 (i32) asm "bswap $0", "=r,r"
2782 Inline assembler expressions may **only** be used as the callee operand
2783 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2784 Thus, typically we have:
2786 .. code-block:: llvm
2788 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2790 Inline asms with side effects not visible in the constraint list must be
2791 marked as having side effects. This is done through the use of the
2792 '``sideeffect``' keyword, like so:
2794 .. code-block:: llvm
2796 call void asm sideeffect "eieio", ""()
2798 In some cases inline asms will contain code that will not work unless
2799 the stack is aligned in some way, such as calls or SSE instructions on
2800 x86, yet will not contain code that does that alignment within the asm.
2801 The compiler should make conservative assumptions about what the asm
2802 might contain and should generate its usual stack alignment code in the
2803 prologue if the '``alignstack``' keyword is present:
2805 .. code-block:: llvm
2807 call void asm alignstack "eieio", ""()
2809 Inline asms also support using non-standard assembly dialects. The
2810 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2811 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2812 the only supported dialects. An example is:
2814 .. code-block:: llvm
2816 call void asm inteldialect "eieio", ""()
2818 If multiple keywords appear the '``sideeffect``' keyword must come
2819 first, the '``alignstack``' keyword second and the '``inteldialect``'
2825 The call instructions that wrap inline asm nodes may have a
2826 "``!srcloc``" MDNode attached to it that contains a list of constant
2827 integers. If present, the code generator will use the integer as the
2828 location cookie value when report errors through the ``LLVMContext``
2829 error reporting mechanisms. This allows a front-end to correlate backend
2830 errors that occur with inline asm back to the source code that produced
2833 .. code-block:: llvm
2835 call void asm sideeffect "something bad", ""(), !srcloc !42
2837 !42 = !{ i32 1234567 }
2839 It is up to the front-end to make sense of the magic numbers it places
2840 in the IR. If the MDNode contains multiple constants, the code generator
2841 will use the one that corresponds to the line of the asm that the error
2849 LLVM IR allows metadata to be attached to instructions in the program
2850 that can convey extra information about the code to the optimizers and
2851 code generator. One example application of metadata is source-level
2852 debug information. There are two metadata primitives: strings and nodes.
2854 Metadata does not have a type, and is not a value. If referenced from a
2855 ``call`` instruction, it uses the ``metadata`` type.
2857 All metadata are identified in syntax by a exclamation point ('``!``').
2859 .. _metadata-string:
2861 Metadata Nodes and Metadata Strings
2862 -----------------------------------
2864 A metadata string is a string surrounded by double quotes. It can
2865 contain any character by escaping non-printable characters with
2866 "``\xx``" where "``xx``" is the two digit hex code. For example:
2869 Metadata nodes are represented with notation similar to structure
2870 constants (a comma separated list of elements, surrounded by braces and
2871 preceded by an exclamation point). Metadata nodes can have any values as
2872 their operand. For example:
2874 .. code-block:: llvm
2876 !{ !"test\00", i32 10}
2878 Metadata nodes that aren't uniqued use the ``distinct`` keyword. For example:
2880 .. code-block:: llvm
2882 !0 = distinct !{!"test\00", i32 10}
2884 ``distinct`` nodes are useful when nodes shouldn't be merged based on their
2885 content. They can also occur when transformations cause uniquing collisions
2886 when metadata operands change.
2888 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2889 metadata nodes, which can be looked up in the module symbol table. For
2892 .. code-block:: llvm
2896 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2897 function is using two metadata arguments:
2899 .. code-block:: llvm
2901 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2903 Metadata can be attached with an instruction. Here metadata ``!21`` is
2904 attached to the ``add`` instruction using the ``!dbg`` identifier:
2906 .. code-block:: llvm
2908 %indvar.next = add i64 %indvar, 1, !dbg !21
2910 More information about specific metadata nodes recognized by the
2911 optimizers and code generator is found below.
2913 .. _specialized-metadata:
2915 Specialized Metadata Nodes
2916 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2918 Specialized metadata nodes are custom data structures in metadata (as opposed
2919 to generic tuples). Their fields are labelled, and can be specified in any
2922 These aren't inherently debug info centric, but currently all the specialized
2923 metadata nodes are related to debug info.
2930 ``MDCompileUnit`` nodes represent a compile unit. The ``enums:``,
2931 ``retainedTypes:``, ``subprograms:``, ``globals:`` and ``imports:`` fields are
2932 tuples containing the debug info to be emitted along with the compile unit,
2933 regardless of code optimizations (some nodes are only emitted if there are
2934 references to them from instructions).
2936 .. code-block:: llvm
2938 !0 = !MDCompileUnit(language: DW_LANG_C99, file: !1, producer: "clang",
2939 isOptimized: true, flags: "-O2", runtimeVersion: 2,
2940 splitDebugFilename: "abc.debug", emissionKind: 1,
2941 enums: !2, retainedTypes: !3, subprograms: !4,
2942 globals: !5, imports: !6)
2944 Compile unit descriptors provide the root scope for objects declared in a
2945 specific compilation unit. File descriptors are defined using this scope.
2946 These descriptors are collected by a named metadata ``!llvm.dbg.cu``. They
2947 keep track of subprograms, global variables, type information, and imported
2948 entities (declarations and namespaces).
2955 ``MDFile`` nodes represent files. The ``filename:`` can include slashes.
2957 .. code-block:: llvm
2959 !0 = !MDFile(filename: "path/to/file", directory: "/path/to/dir")
2961 Files are sometimes used in ``scope:`` fields, and are the only valid target
2962 for ``file:`` fields.
2969 ``MDBasicType`` nodes represent primitive types, such as ``int``, ``bool`` and
2970 ``float``. ``tag:`` defaults to ``DW_TAG_base_type``.
2972 .. code-block:: llvm
2974 !0 = !MDBasicType(name: "unsigned char", size: 8, align: 8,
2975 encoding: DW_ATE_unsigned_char)
2976 !1 = !MDBasicType(tag: DW_TAG_unspecified_type, name: "decltype(nullptr)")
2978 The ``encoding:`` describes the details of the type. Usually it's one of the
2981 .. code-block:: llvm
2987 DW_ATE_signed_char = 6
2989 DW_ATE_unsigned_char = 8
2991 .. _MDSubroutineType:
2996 ``MDSubroutineType`` nodes represent subroutine types. Their ``types:`` field
2997 refers to a tuple; the first operand is the return type, while the rest are the
2998 types of the formal arguments in order. If the first operand is ``null``, that
2999 represents a function with no return value (such as ``void foo() {}`` in C++).
3001 .. code-block:: llvm
3003 !0 = !BasicType(name: "int", size: 32, align: 32, DW_ATE_signed)
3004 !1 = !BasicType(name: "char", size: 8, align: 8, DW_ATE_signed_char)
3005 !2 = !MDSubroutineType(types: !{null, !0, !1}) ; void (int, char)
3012 ``MDDerivedType`` nodes represent types derived from other types, such as
3015 .. code-block:: llvm
3017 !0 = !MDBasicType(name: "unsigned char", size: 8, align: 8,
3018 encoding: DW_ATE_unsigned_char)
3019 !1 = !MDDerivedType(tag: DW_TAG_pointer_type, baseType: !0, size: 32,
3022 The following ``tag:`` values are valid:
3024 .. code-block:: llvm
3026 DW_TAG_formal_parameter = 5
3028 DW_TAG_pointer_type = 15
3029 DW_TAG_reference_type = 16
3031 DW_TAG_ptr_to_member_type = 31
3032 DW_TAG_const_type = 38
3033 DW_TAG_volatile_type = 53
3034 DW_TAG_restrict_type = 55
3036 ``DW_TAG_member`` is used to define a member of a :ref:`composite type
3037 <MDCompositeType>` or :ref:`subprogram <MDSubprogram>`. The type of the member
3038 is the ``baseType:``. The ``offset:`` is the member's bit offset.
3039 ``DW_TAG_formal_parameter`` is used to define a member which is a formal
3040 argument of a subprogram.
3042 ``DW_TAG_typedef`` is used to provide a name for the ``baseType:``.
3044 ``DW_TAG_pointer_type``, ``DW_TAG_reference_type``, ``DW_TAG_const_type``,
3045 ``DW_TAG_volatile_type`` and ``DW_TAG_restrict_type`` are used to qualify the
3048 Note that the ``void *`` type is expressed as a type derived from NULL.
3050 .. _MDCompositeType:
3055 ``MDCompositeType`` nodes represent types composed of other types, like
3056 structures and unions. ``elements:`` points to a tuple of the composed types.
3058 If the source language supports ODR, the ``identifier:`` field gives the unique
3059 identifier used for type merging between modules. When specified, other types
3060 can refer to composite types indirectly via a :ref:`metadata string
3061 <metadata-string>` that matches their identifier.
3063 .. code-block:: llvm
3065 !0 = !MDEnumerator(name: "SixKind", value: 7)
3066 !1 = !MDEnumerator(name: "SevenKind", value: 7)
3067 !2 = !MDEnumerator(name: "NegEightKind", value: -8)
3068 !3 = !MDCompositeType(tag: DW_TAG_enumeration_type, name: "Enum", file: !12,
3069 line: 2, size: 32, align: 32, identifier: "_M4Enum",
3070 elements: !{!0, !1, !2})
3072 The following ``tag:`` values are valid:
3074 .. code-block:: llvm
3076 DW_TAG_array_type = 1
3077 DW_TAG_class_type = 2
3078 DW_TAG_enumeration_type = 4
3079 DW_TAG_structure_type = 19
3080 DW_TAG_union_type = 23
3081 DW_TAG_subroutine_type = 21
3082 DW_TAG_inheritance = 28
3085 For ``DW_TAG_array_type``, the ``elements:`` should be :ref:`subrange
3086 descriptors <MDSubrange>`, each representing the range of subscripts at that
3087 level of indexing. The ``DIFlagVector`` flag to ``flags:`` indicates that an
3088 array type is a native packed vector.
3090 For ``DW_TAG_enumeration_type``, the ``elements:`` should be :ref:`enumerator
3091 descriptors <MDEnumerator>`, each representing the definition of an enumeration
3092 value for the set. All enumeration type descriptors are collected in the
3093 ``enums:`` field of the :ref:`compile unit <MDCompileUnit>`.
3095 For ``DW_TAG_structure_type``, ``DW_TAG_class_type``, and
3096 ``DW_TAG_union_type``, the ``elements:`` should be :ref:`derived types
3097 <MDDerivedType>` with ``tag: DW_TAG_member`` or ``tag: DW_TAG_inheritance``.
3104 ``MDSubrange`` nodes are the elements for ``DW_TAG_array_type`` variants of
3105 :ref:`MDCompositeType`. ``count: -1`` indicates an empty array.
3107 .. code-block:: llvm
3109 !0 = !MDSubrange(count: 5, lowerBound: 0) ; array counting from 0
3110 !1 = !MDSubrange(count: 5, lowerBound: 1) ; array counting from 1
3111 !2 = !MDSubrange(count: -1) ; empty array.
3118 ``MDEnumerator`` nodes are the elements for ``DW_TAG_enumeration_type``
3119 variants of :ref:`MDCompositeType`.
3121 .. code-block:: llvm
3123 !0 = !MDEnumerator(name: "SixKind", value: 7)
3124 !1 = !MDEnumerator(name: "SevenKind", value: 7)
3125 !2 = !MDEnumerator(name: "NegEightKind", value: -8)
3127 MDTemplateTypeParameter
3128 """""""""""""""""""""""
3130 ``MDTemplateTypeParameter`` nodes represent type parameters to generic source
3131 language constructs. They are used (optionally) in :ref:`MDCompositeType` and
3132 :ref:`MDSubprogram` ``templateParams:`` fields.
3134 .. code-block:: llvm
3136 !0 = !MDTemplateTypeParameter(name: "Ty", type: !1)
3138 MDTemplateValueParameter
3139 """"""""""""""""""""""""
3141 ``MDTemplateValueParameter`` nodes represent value parameters to generic source
3142 language constructs. ``tag:`` defaults to ``DW_TAG_template_value_parameter``,
3143 but if specified can also be set to ``DW_TAG_GNU_template_template_param`` or
3144 ``DW_TAG_GNU_template_param_pack``. They are used (optionally) in
3145 :ref:`MDCompositeType` and :ref:`MDSubprogram` ``templateParams:`` fields.
3147 .. code-block:: llvm
3149 !0 = !MDTemplateValueParameter(name: "Ty", type: !1, value: i32 7)
3154 ``MDNamespace`` nodes represent namespaces in the source language.
3156 .. code-block:: llvm
3158 !0 = !MDNamespace(name: "myawesomeproject", scope: !1, file: !2, line: 7)
3163 ``MDGlobalVariable`` nodes represent global variables in the source language.
3165 .. code-block:: llvm
3167 !0 = !MDGlobalVariable(name: "foo", linkageName: "foo", scope: !1,
3168 file: !2, line: 7, type: !3, isLocal: true,
3169 isDefinition: false, variable: i32* @foo,
3172 All global variables should be referenced by the `globals:` field of a
3173 :ref:`compile unit <MDCompileUnit>`.
3180 ``MDSubprogram`` nodes represent functions from the source language. The
3181 ``variables:`` field points at :ref:`variables <MDLocalVariable>` that must be
3182 retained, even if their IR counterparts are optimized out of the IR. The
3183 ``type:`` field must point at an :ref:`MDSubroutineType`.
3185 .. code-block:: llvm
3187 !0 = !MDSubprogram(name: "foo", linkageName: "_Zfoov", scope: !1,
3188 file: !2, line: 7, type: !3, isLocal: true,
3189 isDefinition: false, scopeLine: 8, containingType: !4,
3190 virtuality: DW_VIRTUALITY_pure_virtual, virtualIndex: 10,
3191 flags: DIFlagPrototyped, isOptimized: true,
3192 function: void ()* @_Z3foov,
3193 templateParams: !5, declaration: !6, variables: !7)
3200 ``MDLexicalBlock`` nodes describe nested blocks within a :ref:`subprogram
3201 <MDSubprogram>`. The line number and column numbers are used to dinstinguish
3202 two lexical blocks at same depth. They are valid targets for ``scope:``
3205 .. code-block:: llvm
3207 !0 = distinct !MDLexicalBlock(scope: !1, file: !2, line: 7, column: 35)
3209 Usually lexical blocks are ``distinct`` to prevent node merging based on
3212 .. _MDLexicalBlockFile:
3217 ``MDLexicalBlockFile`` nodes are used to discriminate between sections of a
3218 :ref:`lexical block <MDLexicalBlock>`. The ``file:`` field can be changed to
3219 indicate textual inclusion, or the ``discriminator:`` field can be used to
3220 discriminate between control flow within a single block in the source language.
3222 .. code-block:: llvm
3224 !0 = !MDLexicalBlock(scope: !3, file: !4, line: 7, column: 35)
3225 !1 = !MDLexicalBlockFile(scope: !0, file: !4, discriminator: 0)
3226 !2 = !MDLexicalBlockFile(scope: !0, file: !4, discriminator: 1)
3231 ``MDLocation`` nodes represent source debug locations. The ``scope:`` field is
3232 mandatory, and points at an :ref:`MDLexicalBlockFile`, an
3233 :ref:`MDLexicalBlock`, or an :ref:`MDSubprogram`.
3235 .. code-block:: llvm
3237 !0 = !MDLocation(line: 2900, column: 42, scope: !1, inlinedAt: !2)
3239 .. _MDLocalVariable:
3244 ``MDLocalVariable`` nodes represent local variables in the source language.
3245 Instead of ``DW_TAG_variable``, they use LLVM-specific fake tags to
3246 discriminate between local variables (``DW_TAG_auto_variable``) and subprogram
3247 arguments (``DW_TAG_arg_variable``). In the latter case, the ``arg:`` field
3248 specifies the argument position, and this variable will be included in the
3249 ``variables:`` field of its :ref:`MDSubprogram`.
3251 .. code-block:: llvm
3253 !0 = !MDLocalVariable(tag: DW_TAG_arg_variable, name: "this", arg: 0,
3254 scope: !3, file: !2, line: 7, type: !3,
3255 flags: DIFlagArtificial)
3256 !1 = !MDLocalVariable(tag: DW_TAG_arg_variable, name: "x", arg: 1,
3257 scope: !4, file: !2, line: 7, type: !3)
3258 !1 = !MDLocalVariable(tag: DW_TAG_auto_variable, name: "y",
3259 scope: !5, file: !2, line: 7, type: !3)
3264 ``MDExpression`` nodes represent DWARF expression sequences. They are used in
3265 :ref:`debug intrinsics<dbg_intrinsics>` (such as ``llvm.dbg.declare``) to
3266 describe how the referenced LLVM variable relates to the source language
3269 The current supported vocabulary is limited:
3271 - ``DW_OP_deref`` dereferences the working expression.
3272 - ``DW_OP_plus, 93`` adds ``93`` to the working expression.
3273 - ``DW_OP_bit_piece, 16, 8`` specifies the offset and size (``16`` and ``8``
3274 here, respectively) of the variable piece from the working expression.
3276 .. code-block:: llvm
3278 !0 = !MDExpression(DW_OP_deref)
3279 !1 = !MDExpression(DW_OP_plus, 3)
3280 !2 = !MDExpression(DW_OP_bit_piece, 3, 7)
3281 !3 = !MDExpression(DW_OP_deref, DW_OP_plus, 3, DW_OP_bit_piece, 3, 7)
3286 ``MDObjCProperty`` nodes represent Objective-C property nodes.
3288 .. code-block:: llvm
3290 !3 = !MDObjCProperty(name: "foo", file: !1, line: 7, setter: "setFoo",
3291 getter: "getFoo", attributes: 7, type: !2)
3296 ``MDImportedEntity`` nodes represent entities (such as modules) imported into a
3299 .. code-block:: llvm
3301 !2 = !MDImportedEntity(tag: DW_TAG_imported_module, name: "foo", scope: !0,
3302 entity: !1, line: 7)
3307 In LLVM IR, memory does not have types, so LLVM's own type system is not
3308 suitable for doing TBAA. Instead, metadata is added to the IR to
3309 describe a type system of a higher level language. This can be used to
3310 implement typical C/C++ TBAA, but it can also be used to implement
3311 custom alias analysis behavior for other languages.
3313 The current metadata format is very simple. TBAA metadata nodes have up
3314 to three fields, e.g.:
3316 .. code-block:: llvm
3318 !0 = !{ !"an example type tree" }
3319 !1 = !{ !"int", !0 }
3320 !2 = !{ !"float", !0 }
3321 !3 = !{ !"const float", !2, i64 1 }
3323 The first field is an identity field. It can be any value, usually a
3324 metadata string, which uniquely identifies the type. The most important
3325 name in the tree is the name of the root node. Two trees with different
3326 root node names are entirely disjoint, even if they have leaves with
3329 The second field identifies the type's parent node in the tree, or is
3330 null or omitted for a root node. A type is considered to alias all of
3331 its descendants and all of its ancestors in the tree. Also, a type is
3332 considered to alias all types in other trees, so that bitcode produced
3333 from multiple front-ends is handled conservatively.
3335 If the third field is present, it's an integer which if equal to 1
3336 indicates that the type is "constant" (meaning
3337 ``pointsToConstantMemory`` should return true; see `other useful
3338 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
3340 '``tbaa.struct``' Metadata
3341 ^^^^^^^^^^^^^^^^^^^^^^^^^^
3343 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
3344 aggregate assignment operations in C and similar languages, however it
3345 is defined to copy a contiguous region of memory, which is more than
3346 strictly necessary for aggregate types which contain holes due to
3347 padding. Also, it doesn't contain any TBAA information about the fields
3350 ``!tbaa.struct`` metadata can describe which memory subregions in a
3351 memcpy are padding and what the TBAA tags of the struct are.
3353 The current metadata format is very simple. ``!tbaa.struct`` metadata
3354 nodes are a list of operands which are in conceptual groups of three.
3355 For each group of three, the first operand gives the byte offset of a
3356 field in bytes, the second gives its size in bytes, and the third gives
3359 .. code-block:: llvm
3361 !4 = !{ i64 0, i64 4, !1, i64 8, i64 4, !2 }
3363 This describes a struct with two fields. The first is at offset 0 bytes
3364 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
3365 and has size 4 bytes and has tbaa tag !2.
3367 Note that the fields need not be contiguous. In this example, there is a
3368 4 byte gap between the two fields. This gap represents padding which
3369 does not carry useful data and need not be preserved.
3371 '``noalias``' and '``alias.scope``' Metadata
3372 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3374 ``noalias`` and ``alias.scope`` metadata provide the ability to specify generic
3375 noalias memory-access sets. This means that some collection of memory access
3376 instructions (loads, stores, memory-accessing calls, etc.) that carry
3377 ``noalias`` metadata can specifically be specified not to alias with some other
3378 collection of memory access instructions that carry ``alias.scope`` metadata.
3379 Each type of metadata specifies a list of scopes where each scope has an id and
3380 a domain. When evaluating an aliasing query, if for some domain, the set
3381 of scopes with that domain in one instruction's ``alias.scope`` list is a
3382 subset of (or equal to) the set of scopes for that domain in another
3383 instruction's ``noalias`` list, then the two memory accesses are assumed not to
3386 The metadata identifying each domain is itself a list containing one or two
3387 entries. The first entry is the name of the domain. Note that if the name is a
3388 string then it can be combined accross functions and translation units. A
3389 self-reference can be used to create globally unique domain names. A
3390 descriptive string may optionally be provided as a second list entry.
3392 The metadata identifying each scope is also itself a list containing two or
3393 three entries. The first entry is the name of the scope. Note that if the name
3394 is a string then it can be combined accross functions and translation units. A
3395 self-reference can be used to create globally unique scope names. A metadata
3396 reference to the scope's domain is the second entry. A descriptive string may
3397 optionally be provided as a third list entry.
3401 .. code-block:: llvm
3403 ; Two scope domains:
3407 ; Some scopes in these domains:
3413 !5 = !{!4} ; A list containing only scope !4
3417 ; These two instructions don't alias:
3418 %0 = load float, float* %c, align 4, !alias.scope !5
3419 store float %0, float* %arrayidx.i, align 4, !noalias !5
3421 ; These two instructions also don't alias (for domain !1, the set of scopes
3422 ; in the !alias.scope equals that in the !noalias list):
3423 %2 = load float, float* %c, align 4, !alias.scope !5
3424 store float %2, float* %arrayidx.i2, align 4, !noalias !6
3426 ; These two instructions don't alias (for domain !0, the set of scopes in
3427 ; the !noalias list is not a superset of, or equal to, the scopes in the
3428 ; !alias.scope list):
3429 %2 = load float, float* %c, align 4, !alias.scope !6
3430 store float %0, float* %arrayidx.i, align 4, !noalias !7
3432 '``fpmath``' Metadata
3433 ^^^^^^^^^^^^^^^^^^^^^
3435 ``fpmath`` metadata may be attached to any instruction of floating point
3436 type. It can be used to express the maximum acceptable error in the
3437 result of that instruction, in ULPs, thus potentially allowing the
3438 compiler to use a more efficient but less accurate method of computing
3439 it. ULP is defined as follows:
3441 If ``x`` is a real number that lies between two finite consecutive
3442 floating-point numbers ``a`` and ``b``, without being equal to one
3443 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
3444 distance between the two non-equal finite floating-point numbers
3445 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
3447 The metadata node shall consist of a single positive floating point
3448 number representing the maximum relative error, for example:
3450 .. code-block:: llvm
3452 !0 = !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
3456 '``range``' Metadata
3457 ^^^^^^^^^^^^^^^^^^^^
3459 ``range`` metadata may be attached only to ``load``, ``call`` and ``invoke`` of
3460 integer types. It expresses the possible ranges the loaded value or the value
3461 returned by the called function at this call site is in. The ranges are
3462 represented with a flattened list of integers. The loaded value or the value
3463 returned is known to be in the union of the ranges defined by each consecutive
3464 pair. Each pair has the following properties:
3466 - The type must match the type loaded by the instruction.
3467 - The pair ``a,b`` represents the range ``[a,b)``.
3468 - Both ``a`` and ``b`` are constants.
3469 - The range is allowed to wrap.
3470 - The range should not represent the full or empty set. That is,
3473 In addition, the pairs must be in signed order of the lower bound and
3474 they must be non-contiguous.
3478 .. code-block:: llvm
3480 %a = load i8, i8* %x, align 1, !range !0 ; Can only be 0 or 1
3481 %b = load i8, i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
3482 %c = call i8 @foo(), !range !2 ; Can only be 0, 1, 3, 4 or 5
3483 %d = invoke i8 @bar() to label %cont
3484 unwind label %lpad, !range !3 ; Can only be -2, -1, 3, 4 or 5
3486 !0 = !{ i8 0, i8 2 }
3487 !1 = !{ i8 255, i8 2 }
3488 !2 = !{ i8 0, i8 2, i8 3, i8 6 }
3489 !3 = !{ i8 -2, i8 0, i8 3, i8 6 }
3494 It is sometimes useful to attach information to loop constructs. Currently,
3495 loop metadata is implemented as metadata attached to the branch instruction
3496 in the loop latch block. This type of metadata refer to a metadata node that is
3497 guaranteed to be separate for each loop. The loop identifier metadata is
3498 specified with the name ``llvm.loop``.
3500 The loop identifier metadata is implemented using a metadata that refers to
3501 itself to avoid merging it with any other identifier metadata, e.g.,
3502 during module linkage or function inlining. That is, each loop should refer
3503 to their own identification metadata even if they reside in separate functions.
3504 The following example contains loop identifier metadata for two separate loop
3507 .. code-block:: llvm
3512 The loop identifier metadata can be used to specify additional
3513 per-loop metadata. Any operands after the first operand can be treated
3514 as user-defined metadata. For example the ``llvm.loop.unroll.count``
3515 suggests an unroll factor to the loop unroller:
3517 .. code-block:: llvm
3519 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
3522 !1 = !{!"llvm.loop.unroll.count", i32 4}
3524 '``llvm.loop.vectorize``' and '``llvm.loop.interleave``'
3525 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3527 Metadata prefixed with ``llvm.loop.vectorize`` or ``llvm.loop.interleave`` are
3528 used to control per-loop vectorization and interleaving parameters such as
3529 vectorization width and interleave count. These metadata should be used in
3530 conjunction with ``llvm.loop`` loop identification metadata. The
3531 ``llvm.loop.vectorize`` and ``llvm.loop.interleave`` metadata are only
3532 optimization hints and the optimizer will only interleave and vectorize loops if
3533 it believes it is safe to do so. The ``llvm.mem.parallel_loop_access`` metadata
3534 which contains information about loop-carried memory dependencies can be helpful
3535 in determining the safety of these transformations.
3537 '``llvm.loop.interleave.count``' Metadata
3538 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3540 This metadata suggests an interleave count to the loop interleaver.
3541 The first operand is the string ``llvm.loop.interleave.count`` and the
3542 second operand is an integer specifying the interleave count. For
3545 .. code-block:: llvm
3547 !0 = !{!"llvm.loop.interleave.count", i32 4}
3549 Note that setting ``llvm.loop.interleave.count`` to 1 disables interleaving
3550 multiple iterations of the loop. If ``llvm.loop.interleave.count`` is set to 0
3551 then the interleave count will be determined automatically.
3553 '``llvm.loop.vectorize.enable``' Metadata
3554 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3556 This metadata selectively enables or disables vectorization for the loop. The
3557 first operand is the string ``llvm.loop.vectorize.enable`` and the second operand
3558 is a bit. If the bit operand value is 1 vectorization is enabled. A value of
3559 0 disables vectorization:
3561 .. code-block:: llvm
3563 !0 = !{!"llvm.loop.vectorize.enable", i1 0}
3564 !1 = !{!"llvm.loop.vectorize.enable", i1 1}
3566 '``llvm.loop.vectorize.width``' Metadata
3567 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3569 This metadata sets the target width of the vectorizer. The first
3570 operand is the string ``llvm.loop.vectorize.width`` and the second
3571 operand is an integer specifying the width. For example:
3573 .. code-block:: llvm
3575 !0 = !{!"llvm.loop.vectorize.width", i32 4}
3577 Note that setting ``llvm.loop.vectorize.width`` to 1 disables
3578 vectorization of the loop. If ``llvm.loop.vectorize.width`` is set to
3579 0 or if the loop does not have this metadata the width will be
3580 determined automatically.
3582 '``llvm.loop.unroll``'
3583 ^^^^^^^^^^^^^^^^^^^^^^
3585 Metadata prefixed with ``llvm.loop.unroll`` are loop unrolling
3586 optimization hints such as the unroll factor. ``llvm.loop.unroll``
3587 metadata should be used in conjunction with ``llvm.loop`` loop
3588 identification metadata. The ``llvm.loop.unroll`` metadata are only
3589 optimization hints and the unrolling will only be performed if the
3590 optimizer believes it is safe to do so.
3592 '``llvm.loop.unroll.count``' Metadata
3593 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3595 This metadata suggests an unroll factor to the loop unroller. The
3596 first operand is the string ``llvm.loop.unroll.count`` and the second
3597 operand is a positive integer specifying the unroll factor. For
3600 .. code-block:: llvm
3602 !0 = !{!"llvm.loop.unroll.count", i32 4}
3604 If the trip count of the loop is less than the unroll count the loop
3605 will be partially unrolled.
3607 '``llvm.loop.unroll.disable``' Metadata
3608 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3610 This metadata either disables loop unrolling. The metadata has a single operand
3611 which is the string ``llvm.loop.unroll.disable``. For example:
3613 .. code-block:: llvm
3615 !0 = !{!"llvm.loop.unroll.disable"}
3617 '``llvm.loop.unroll.runtime.disable``' Metadata
3618 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3620 This metadata either disables runtime loop unrolling. The metadata has a single
3621 operand which is the string ``llvm.loop.unroll.runtime.disable``. For example:
3623 .. code-block:: llvm
3625 !0 = !{!"llvm.loop.unroll.runtime.disable"}
3627 '``llvm.loop.unroll.full``' Metadata
3628 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3630 This metadata either suggests that the loop should be unrolled fully. The
3631 metadata has a single operand which is the string ``llvm.loop.unroll.disable``.
3634 .. code-block:: llvm
3636 !0 = !{!"llvm.loop.unroll.full"}
3641 Metadata types used to annotate memory accesses with information helpful
3642 for optimizations are prefixed with ``llvm.mem``.
3644 '``llvm.mem.parallel_loop_access``' Metadata
3645 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3647 The ``llvm.mem.parallel_loop_access`` metadata refers to a loop identifier,
3648 or metadata containing a list of loop identifiers for nested loops.
3649 The metadata is attached to memory accessing instructions and denotes that
3650 no loop carried memory dependence exist between it and other instructions denoted
3651 with the same loop identifier.
3653 Precisely, given two instructions ``m1`` and ``m2`` that both have the
3654 ``llvm.mem.parallel_loop_access`` metadata, with ``L1`` and ``L2`` being the
3655 set of loops associated with that metadata, respectively, then there is no loop
3656 carried dependence between ``m1`` and ``m2`` for loops in both ``L1`` and
3659 As a special case, if all memory accessing instructions in a loop have
3660 ``llvm.mem.parallel_loop_access`` metadata that refers to that loop, then the
3661 loop has no loop carried memory dependences and is considered to be a parallel
3664 Note that if not all memory access instructions have such metadata referring to
3665 the loop, then the loop is considered not being trivially parallel. Additional
3666 memory dependence analysis is required to make that determination. As a fail
3667 safe mechanism, this causes loops that were originally parallel to be considered
3668 sequential (if optimization passes that are unaware of the parallel semantics
3669 insert new memory instructions into the loop body).
3671 Example of a loop that is considered parallel due to its correct use of
3672 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
3673 metadata types that refer to the same loop identifier metadata.
3675 .. code-block:: llvm
3679 %val0 = load i32, i32* %arrayidx, !llvm.mem.parallel_loop_access !0
3681 store i32 %val0, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3683 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
3689 It is also possible to have nested parallel loops. In that case the
3690 memory accesses refer to a list of loop identifier metadata nodes instead of
3691 the loop identifier metadata node directly:
3693 .. code-block:: llvm
3697 %val1 = load i32, i32* %arrayidx3, !llvm.mem.parallel_loop_access !2
3699 br label %inner.for.body
3703 %val0 = load i32, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3705 store i32 %val0, i32* %arrayidx2, !llvm.mem.parallel_loop_access !0
3707 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
3711 store i32 %val1, i32* %arrayidx4, !llvm.mem.parallel_loop_access !2
3713 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
3715 outer.for.end: ; preds = %for.body
3717 !0 = !{!1, !2} ; a list of loop identifiers
3718 !1 = !{!1} ; an identifier for the inner loop
3719 !2 = !{!2} ; an identifier for the outer loop
3724 The ``llvm.bitsets`` global metadata is used to implement
3725 :doc:`bitsets <BitSets>`.
3727 Module Flags Metadata
3728 =====================
3730 Information about the module as a whole is difficult to convey to LLVM's
3731 subsystems. The LLVM IR isn't sufficient to transmit this information.
3732 The ``llvm.module.flags`` named metadata exists in order to facilitate
3733 this. These flags are in the form of key / value pairs --- much like a
3734 dictionary --- making it easy for any subsystem who cares about a flag to
3737 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
3738 Each triplet has the following form:
3740 - The first element is a *behavior* flag, which specifies the behavior
3741 when two (or more) modules are merged together, and it encounters two
3742 (or more) metadata with the same ID. The supported behaviors are
3744 - The second element is a metadata string that is a unique ID for the
3745 metadata. Each module may only have one flag entry for each unique ID (not
3746 including entries with the **Require** behavior).
3747 - The third element is the value of the flag.
3749 When two (or more) modules are merged together, the resulting
3750 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
3751 each unique metadata ID string, there will be exactly one entry in the merged
3752 modules ``llvm.module.flags`` metadata table, and the value for that entry will
3753 be determined by the merge behavior flag, as described below. The only exception
3754 is that entries with the *Require* behavior are always preserved.
3756 The following behaviors are supported:
3767 Emits an error if two values disagree, otherwise the resulting value
3768 is that of the operands.
3772 Emits a warning if two values disagree. The result value will be the
3773 operand for the flag from the first module being linked.
3777 Adds a requirement that another module flag be present and have a
3778 specified value after linking is performed. The value must be a
3779 metadata pair, where the first element of the pair is the ID of the
3780 module flag to be restricted, and the second element of the pair is
3781 the value the module flag should be restricted to. This behavior can
3782 be used to restrict the allowable results (via triggering of an
3783 error) of linking IDs with the **Override** behavior.
3787 Uses the specified value, regardless of the behavior or value of the
3788 other module. If both modules specify **Override**, but the values
3789 differ, an error will be emitted.
3793 Appends the two values, which are required to be metadata nodes.
3797 Appends the two values, which are required to be metadata
3798 nodes. However, duplicate entries in the second list are dropped
3799 during the append operation.
3801 It is an error for a particular unique flag ID to have multiple behaviors,
3802 except in the case of **Require** (which adds restrictions on another metadata
3803 value) or **Override**.
3805 An example of module flags:
3807 .. code-block:: llvm
3809 !0 = !{ i32 1, !"foo", i32 1 }
3810 !1 = !{ i32 4, !"bar", i32 37 }
3811 !2 = !{ i32 2, !"qux", i32 42 }
3812 !3 = !{ i32 3, !"qux",
3817 !llvm.module.flags = !{ !0, !1, !2, !3 }
3819 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
3820 if two or more ``!"foo"`` flags are seen is to emit an error if their
3821 values are not equal.
3823 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
3824 behavior if two or more ``!"bar"`` flags are seen is to use the value
3827 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
3828 behavior if two or more ``!"qux"`` flags are seen is to emit a
3829 warning if their values are not equal.
3831 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
3837 The behavior is to emit an error if the ``llvm.module.flags`` does not
3838 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
3841 Objective-C Garbage Collection Module Flags Metadata
3842 ----------------------------------------------------
3844 On the Mach-O platform, Objective-C stores metadata about garbage
3845 collection in a special section called "image info". The metadata
3846 consists of a version number and a bitmask specifying what types of
3847 garbage collection are supported (if any) by the file. If two or more
3848 modules are linked together their garbage collection metadata needs to
3849 be merged rather than appended together.
3851 The Objective-C garbage collection module flags metadata consists of the
3852 following key-value pairs:
3861 * - ``Objective-C Version``
3862 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
3864 * - ``Objective-C Image Info Version``
3865 - **[Required]** --- The version of the image info section. Currently
3868 * - ``Objective-C Image Info Section``
3869 - **[Required]** --- The section to place the metadata. Valid values are
3870 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
3871 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
3872 Objective-C ABI version 2.
3874 * - ``Objective-C Garbage Collection``
3875 - **[Required]** --- Specifies whether garbage collection is supported or
3876 not. Valid values are 0, for no garbage collection, and 2, for garbage
3877 collection supported.
3879 * - ``Objective-C GC Only``
3880 - **[Optional]** --- Specifies that only garbage collection is supported.
3881 If present, its value must be 6. This flag requires that the
3882 ``Objective-C Garbage Collection`` flag have the value 2.
3884 Some important flag interactions:
3886 - If a module with ``Objective-C Garbage Collection`` set to 0 is
3887 merged with a module with ``Objective-C Garbage Collection`` set to
3888 2, then the resulting module has the
3889 ``Objective-C Garbage Collection`` flag set to 0.
3890 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
3891 merged with a module with ``Objective-C GC Only`` set to 6.
3893 Automatic Linker Flags Module Flags Metadata
3894 --------------------------------------------
3896 Some targets support embedding flags to the linker inside individual object
3897 files. Typically this is used in conjunction with language extensions which
3898 allow source files to explicitly declare the libraries they depend on, and have
3899 these automatically be transmitted to the linker via object files.
3901 These flags are encoded in the IR using metadata in the module flags section,
3902 using the ``Linker Options`` key. The merge behavior for this flag is required
3903 to be ``AppendUnique``, and the value for the key is expected to be a metadata
3904 node which should be a list of other metadata nodes, each of which should be a
3905 list of metadata strings defining linker options.
3907 For example, the following metadata section specifies two separate sets of
3908 linker options, presumably to link against ``libz`` and the ``Cocoa``
3911 !0 = !{ i32 6, !"Linker Options",
3914 !{ !"-framework", !"Cocoa" } } }
3915 !llvm.module.flags = !{ !0 }
3917 The metadata encoding as lists of lists of options, as opposed to a collapsed
3918 list of options, is chosen so that the IR encoding can use multiple option
3919 strings to specify e.g., a single library, while still having that specifier be
3920 preserved as an atomic element that can be recognized by a target specific
3921 assembly writer or object file emitter.
3923 Each individual option is required to be either a valid option for the target's
3924 linker, or an option that is reserved by the target specific assembly writer or
3925 object file emitter. No other aspect of these options is defined by the IR.
3927 C type width Module Flags Metadata
3928 ----------------------------------
3930 The ARM backend emits a section into each generated object file describing the
3931 options that it was compiled with (in a compiler-independent way) to prevent
3932 linking incompatible objects, and to allow automatic library selection. Some
3933 of these options are not visible at the IR level, namely wchar_t width and enum
3936 To pass this information to the backend, these options are encoded in module
3937 flags metadata, using the following key-value pairs:
3947 - * 0 --- sizeof(wchar_t) == 4
3948 * 1 --- sizeof(wchar_t) == 2
3951 - * 0 --- Enums are at least as large as an ``int``.
3952 * 1 --- Enums are stored in the smallest integer type which can
3953 represent all of its values.
3955 For example, the following metadata section specifies that the module was
3956 compiled with a ``wchar_t`` width of 4 bytes, and the underlying type of an
3957 enum is the smallest type which can represent all of its values::
3959 !llvm.module.flags = !{!0, !1}
3960 !0 = !{i32 1, !"short_wchar", i32 1}
3961 !1 = !{i32 1, !"short_enum", i32 0}
3963 .. _intrinsicglobalvariables:
3965 Intrinsic Global Variables
3966 ==========================
3968 LLVM has a number of "magic" global variables that contain data that
3969 affect code generation or other IR semantics. These are documented here.
3970 All globals of this sort should have a section specified as
3971 "``llvm.metadata``". This section and all globals that start with
3972 "``llvm.``" are reserved for use by LLVM.
3976 The '``llvm.used``' Global Variable
3977 -----------------------------------
3979 The ``@llvm.used`` global is an array which has
3980 :ref:`appending linkage <linkage_appending>`. This array contains a list of
3981 pointers to named global variables, functions and aliases which may optionally
3982 have a pointer cast formed of bitcast or getelementptr. For example, a legal
3985 .. code-block:: llvm
3990 @llvm.used = appending global [2 x i8*] [
3992 i8* bitcast (i32* @Y to i8*)
3993 ], section "llvm.metadata"
3995 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
3996 and linker are required to treat the symbol as if there is a reference to the
3997 symbol that it cannot see (which is why they have to be named). For example, if
3998 a variable has internal linkage and no references other than that from the
3999 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
4000 references from inline asms and other things the compiler cannot "see", and
4001 corresponds to "``attribute((used))``" in GNU C.
4003 On some targets, the code generator must emit a directive to the
4004 assembler or object file to prevent the assembler and linker from
4005 molesting the symbol.
4007 .. _gv_llvmcompilerused:
4009 The '``llvm.compiler.used``' Global Variable
4010 --------------------------------------------
4012 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
4013 directive, except that it only prevents the compiler from touching the
4014 symbol. On targets that support it, this allows an intelligent linker to
4015 optimize references to the symbol without being impeded as it would be
4018 This is a rare construct that should only be used in rare circumstances,
4019 and should not be exposed to source languages.
4021 .. _gv_llvmglobalctors:
4023 The '``llvm.global_ctors``' Global Variable
4024 -------------------------------------------
4026 .. code-block:: llvm
4028 %0 = type { i32, void ()*, i8* }
4029 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor, i8* @data }]
4031 The ``@llvm.global_ctors`` array contains a list of constructor
4032 functions, priorities, and an optional associated global or function.
4033 The functions referenced by this array will be called in ascending order
4034 of priority (i.e. lowest first) when the module is loaded. The order of
4035 functions with the same priority is not defined.
4037 If the third field is present, non-null, and points to a global variable
4038 or function, the initializer function will only run if the associated
4039 data from the current module is not discarded.
4041 .. _llvmglobaldtors:
4043 The '``llvm.global_dtors``' Global Variable
4044 -------------------------------------------
4046 .. code-block:: llvm
4048 %0 = type { i32, void ()*, i8* }
4049 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor, i8* @data }]
4051 The ``@llvm.global_dtors`` array contains a list of destructor
4052 functions, priorities, and an optional associated global or function.
4053 The functions referenced by this array will be called in descending
4054 order of priority (i.e. highest first) when the module is unloaded. The
4055 order of functions with the same priority is not defined.
4057 If the third field is present, non-null, and points to a global variable
4058 or function, the destructor function will only run if the associated
4059 data from the current module is not discarded.
4061 Instruction Reference
4062 =====================
4064 The LLVM instruction set consists of several different classifications
4065 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
4066 instructions <binaryops>`, :ref:`bitwise binary
4067 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
4068 :ref:`other instructions <otherops>`.
4072 Terminator Instructions
4073 -----------------------
4075 As mentioned :ref:`previously <functionstructure>`, every basic block in a
4076 program ends with a "Terminator" instruction, which indicates which
4077 block should be executed after the current block is finished. These
4078 terminator instructions typically yield a '``void``' value: they produce
4079 control flow, not values (the one exception being the
4080 ':ref:`invoke <i_invoke>`' instruction).
4082 The terminator instructions are: ':ref:`ret <i_ret>`',
4083 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
4084 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
4085 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
4089 '``ret``' Instruction
4090 ^^^^^^^^^^^^^^^^^^^^^
4097 ret <type> <value> ; Return a value from a non-void function
4098 ret void ; Return from void function
4103 The '``ret``' instruction is used to return control flow (and optionally
4104 a value) from a function back to the caller.
4106 There are two forms of the '``ret``' instruction: one that returns a
4107 value and then causes control flow, and one that just causes control
4113 The '``ret``' instruction optionally accepts a single argument, the
4114 return value. The type of the return value must be a ':ref:`first
4115 class <t_firstclass>`' type.
4117 A function is not :ref:`well formed <wellformed>` if it it has a non-void
4118 return type and contains a '``ret``' instruction with no return value or
4119 a return value with a type that does not match its type, or if it has a
4120 void return type and contains a '``ret``' instruction with a return
4126 When the '``ret``' instruction is executed, control flow returns back to
4127 the calling function's context. If the caller is a
4128 ":ref:`call <i_call>`" instruction, execution continues at the
4129 instruction after the call. If the caller was an
4130 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
4131 beginning of the "normal" destination block. If the instruction returns
4132 a value, that value shall set the call or invoke instruction's return
4138 .. code-block:: llvm
4140 ret i32 5 ; Return an integer value of 5
4141 ret void ; Return from a void function
4142 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
4146 '``br``' Instruction
4147 ^^^^^^^^^^^^^^^^^^^^
4154 br i1 <cond>, label <iftrue>, label <iffalse>
4155 br label <dest> ; Unconditional branch
4160 The '``br``' instruction is used to cause control flow to transfer to a
4161 different basic block in the current function. There are two forms of
4162 this instruction, corresponding to a conditional branch and an
4163 unconditional branch.
4168 The conditional branch form of the '``br``' instruction takes a single
4169 '``i1``' value and two '``label``' values. The unconditional form of the
4170 '``br``' instruction takes a single '``label``' value as a target.
4175 Upon execution of a conditional '``br``' instruction, the '``i1``'
4176 argument is evaluated. If the value is ``true``, control flows to the
4177 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
4178 to the '``iffalse``' ``label`` argument.
4183 .. code-block:: llvm
4186 %cond = icmp eq i32 %a, %b
4187 br i1 %cond, label %IfEqual, label %IfUnequal
4195 '``switch``' Instruction
4196 ^^^^^^^^^^^^^^^^^^^^^^^^
4203 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
4208 The '``switch``' instruction is used to transfer control flow to one of
4209 several different places. It is a generalization of the '``br``'
4210 instruction, allowing a branch to occur to one of many possible
4216 The '``switch``' instruction uses three parameters: an integer
4217 comparison value '``value``', a default '``label``' destination, and an
4218 array of pairs of comparison value constants and '``label``'s. The table
4219 is not allowed to contain duplicate constant entries.
4224 The ``switch`` instruction specifies a table of values and destinations.
4225 When the '``switch``' instruction is executed, this table is searched
4226 for the given value. If the value is found, control flow is transferred
4227 to the corresponding destination; otherwise, control flow is transferred
4228 to the default destination.
4233 Depending on properties of the target machine and the particular
4234 ``switch`` instruction, this instruction may be code generated in
4235 different ways. For example, it could be generated as a series of
4236 chained conditional branches or with a lookup table.
4241 .. code-block:: llvm
4243 ; Emulate a conditional br instruction
4244 %Val = zext i1 %value to i32
4245 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
4247 ; Emulate an unconditional br instruction
4248 switch i32 0, label %dest [ ]
4250 ; Implement a jump table:
4251 switch i32 %val, label %otherwise [ i32 0, label %onzero
4253 i32 2, label %ontwo ]
4257 '``indirectbr``' Instruction
4258 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4265 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
4270 The '``indirectbr``' instruction implements an indirect branch to a
4271 label within the current function, whose address is specified by
4272 "``address``". Address must be derived from a
4273 :ref:`blockaddress <blockaddress>` constant.
4278 The '``address``' argument is the address of the label to jump to. The
4279 rest of the arguments indicate the full set of possible destinations
4280 that the address may point to. Blocks are allowed to occur multiple
4281 times in the destination list, though this isn't particularly useful.
4283 This destination list is required so that dataflow analysis has an
4284 accurate understanding of the CFG.
4289 Control transfers to the block specified in the address argument. All
4290 possible destination blocks must be listed in the label list, otherwise
4291 this instruction has undefined behavior. This implies that jumps to
4292 labels defined in other functions have undefined behavior as well.
4297 This is typically implemented with a jump through a register.
4302 .. code-block:: llvm
4304 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
4308 '``invoke``' Instruction
4309 ^^^^^^^^^^^^^^^^^^^^^^^^
4316 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
4317 to label <normal label> unwind label <exception label>
4322 The '``invoke``' instruction causes control to transfer to a specified
4323 function, with the possibility of control flow transfer to either the
4324 '``normal``' label or the '``exception``' label. If the callee function
4325 returns with the "``ret``" instruction, control flow will return to the
4326 "normal" label. If the callee (or any indirect callees) returns via the
4327 ":ref:`resume <i_resume>`" instruction or other exception handling
4328 mechanism, control is interrupted and continued at the dynamically
4329 nearest "exception" label.
4331 The '``exception``' label is a `landing
4332 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
4333 '``exception``' label is required to have the
4334 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
4335 information about the behavior of the program after unwinding happens,
4336 as its first non-PHI instruction. The restrictions on the
4337 "``landingpad``" instruction's tightly couples it to the "``invoke``"
4338 instruction, so that the important information contained within the
4339 "``landingpad``" instruction can't be lost through normal code motion.
4344 This instruction requires several arguments:
4346 #. The optional "cconv" marker indicates which :ref:`calling
4347 convention <callingconv>` the call should use. If none is
4348 specified, the call defaults to using C calling conventions.
4349 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
4350 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
4352 #. '``ptr to function ty``': shall be the signature of the pointer to
4353 function value being invoked. In most cases, this is a direct
4354 function invocation, but indirect ``invoke``'s are just as possible,
4355 branching off an arbitrary pointer to function value.
4356 #. '``function ptr val``': An LLVM value containing a pointer to a
4357 function to be invoked.
4358 #. '``function args``': argument list whose types match the function
4359 signature argument types and parameter attributes. All arguments must
4360 be of :ref:`first class <t_firstclass>` type. If the function signature
4361 indicates the function accepts a variable number of arguments, the
4362 extra arguments can be specified.
4363 #. '``normal label``': the label reached when the called function
4364 executes a '``ret``' instruction.
4365 #. '``exception label``': the label reached when a callee returns via
4366 the :ref:`resume <i_resume>` instruction or other exception handling
4368 #. The optional :ref:`function attributes <fnattrs>` list. Only
4369 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
4370 attributes are valid here.
4375 This instruction is designed to operate as a standard '``call``'
4376 instruction in most regards. The primary difference is that it
4377 establishes an association with a label, which is used by the runtime
4378 library to unwind the stack.
4380 This instruction is used in languages with destructors to ensure that
4381 proper cleanup is performed in the case of either a ``longjmp`` or a
4382 thrown exception. Additionally, this is important for implementation of
4383 '``catch``' clauses in high-level languages that support them.
4385 For the purposes of the SSA form, the definition of the value returned
4386 by the '``invoke``' instruction is deemed to occur on the edge from the
4387 current block to the "normal" label. If the callee unwinds then no
4388 return value is available.
4393 .. code-block:: llvm
4395 %retval = invoke i32 @Test(i32 15) to label %Continue
4396 unwind label %TestCleanup ; i32:retval set
4397 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
4398 unwind label %TestCleanup ; i32:retval set
4402 '``resume``' Instruction
4403 ^^^^^^^^^^^^^^^^^^^^^^^^
4410 resume <type> <value>
4415 The '``resume``' instruction is a terminator instruction that has no
4421 The '``resume``' instruction requires one argument, which must have the
4422 same type as the result of any '``landingpad``' instruction in the same
4428 The '``resume``' instruction resumes propagation of an existing
4429 (in-flight) exception whose unwinding was interrupted with a
4430 :ref:`landingpad <i_landingpad>` instruction.
4435 .. code-block:: llvm
4437 resume { i8*, i32 } %exn
4441 '``unreachable``' Instruction
4442 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4454 The '``unreachable``' instruction has no defined semantics. This
4455 instruction is used to inform the optimizer that a particular portion of
4456 the code is not reachable. This can be used to indicate that the code
4457 after a no-return function cannot be reached, and other facts.
4462 The '``unreachable``' instruction has no defined semantics.
4469 Binary operators are used to do most of the computation in a program.
4470 They require two operands of the same type, execute an operation on
4471 them, and produce a single value. The operands might represent multiple
4472 data, as is the case with the :ref:`vector <t_vector>` data type. The
4473 result value has the same type as its operands.
4475 There are several different binary operators:
4479 '``add``' Instruction
4480 ^^^^^^^^^^^^^^^^^^^^^
4487 <result> = add <ty> <op1>, <op2> ; yields ty:result
4488 <result> = add nuw <ty> <op1>, <op2> ; yields ty:result
4489 <result> = add nsw <ty> <op1>, <op2> ; yields ty:result
4490 <result> = add nuw nsw <ty> <op1>, <op2> ; yields ty:result
4495 The '``add``' instruction returns the sum of its two operands.
4500 The two arguments to the '``add``' instruction must be
4501 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4502 arguments must have identical types.
4507 The value produced is the integer sum of the two operands.
4509 If the sum has unsigned overflow, the result returned is the
4510 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4513 Because LLVM integers use a two's complement representation, this
4514 instruction is appropriate for both signed and unsigned integers.
4516 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4517 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4518 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
4519 unsigned and/or signed overflow, respectively, occurs.
4524 .. code-block:: llvm
4526 <result> = add i32 4, %var ; yields i32:result = 4 + %var
4530 '``fadd``' Instruction
4531 ^^^^^^^^^^^^^^^^^^^^^^
4538 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4543 The '``fadd``' instruction returns the sum of its two operands.
4548 The two arguments to the '``fadd``' instruction must be :ref:`floating
4549 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4550 Both arguments must have identical types.
4555 The value produced is the floating point sum of the two operands. This
4556 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
4557 which are optimization hints to enable otherwise unsafe floating point
4563 .. code-block:: llvm
4565 <result> = fadd float 4.0, %var ; yields float:result = 4.0 + %var
4567 '``sub``' Instruction
4568 ^^^^^^^^^^^^^^^^^^^^^
4575 <result> = sub <ty> <op1>, <op2> ; yields ty:result
4576 <result> = sub nuw <ty> <op1>, <op2> ; yields ty:result
4577 <result> = sub nsw <ty> <op1>, <op2> ; yields ty:result
4578 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields ty:result
4583 The '``sub``' instruction returns the difference of its two operands.
4585 Note that the '``sub``' instruction is used to represent the '``neg``'
4586 instruction present in most other intermediate representations.
4591 The two arguments to the '``sub``' instruction must be
4592 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4593 arguments must have identical types.
4598 The value produced is the integer difference of the two operands.
4600 If the difference has unsigned overflow, the result returned is the
4601 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4604 Because LLVM integers use a two's complement representation, this
4605 instruction is appropriate for both signed and unsigned integers.
4607 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4608 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4609 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
4610 unsigned and/or signed overflow, respectively, occurs.
4615 .. code-block:: llvm
4617 <result> = sub i32 4, %var ; yields i32:result = 4 - %var
4618 <result> = sub i32 0, %val ; yields i32:result = -%var
4622 '``fsub``' Instruction
4623 ^^^^^^^^^^^^^^^^^^^^^^
4630 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4635 The '``fsub``' instruction returns the difference of its two operands.
4637 Note that the '``fsub``' instruction is used to represent the '``fneg``'
4638 instruction present in most other intermediate representations.
4643 The two arguments to the '``fsub``' instruction must be :ref:`floating
4644 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4645 Both arguments must have identical types.
4650 The value produced is the floating point difference of the two operands.
4651 This instruction can also take any number of :ref:`fast-math
4652 flags <fastmath>`, which are optimization hints to enable otherwise
4653 unsafe floating point optimizations:
4658 .. code-block:: llvm
4660 <result> = fsub float 4.0, %var ; yields float:result = 4.0 - %var
4661 <result> = fsub float -0.0, %val ; yields float:result = -%var
4663 '``mul``' Instruction
4664 ^^^^^^^^^^^^^^^^^^^^^
4671 <result> = mul <ty> <op1>, <op2> ; yields ty:result
4672 <result> = mul nuw <ty> <op1>, <op2> ; yields ty:result
4673 <result> = mul nsw <ty> <op1>, <op2> ; yields ty:result
4674 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields ty:result
4679 The '``mul``' instruction returns the product of its two operands.
4684 The two arguments to the '``mul``' instruction must be
4685 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4686 arguments must have identical types.
4691 The value produced is the integer product of the two operands.
4693 If the result of the multiplication has unsigned overflow, the result
4694 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
4695 bit width of the result.
4697 Because LLVM integers use a two's complement representation, and the
4698 result is the same width as the operands, this instruction returns the
4699 correct result for both signed and unsigned integers. If a full product
4700 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
4701 sign-extended or zero-extended as appropriate to the width of the full
4704 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4705 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4706 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
4707 unsigned and/or signed overflow, respectively, occurs.
4712 .. code-block:: llvm
4714 <result> = mul i32 4, %var ; yields i32:result = 4 * %var
4718 '``fmul``' Instruction
4719 ^^^^^^^^^^^^^^^^^^^^^^
4726 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4731 The '``fmul``' instruction returns the product of its two operands.
4736 The two arguments to the '``fmul``' instruction must be :ref:`floating
4737 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4738 Both arguments must have identical types.
4743 The value produced is the floating point product of the two operands.
4744 This instruction can also take any number of :ref:`fast-math
4745 flags <fastmath>`, which are optimization hints to enable otherwise
4746 unsafe floating point optimizations:
4751 .. code-block:: llvm
4753 <result> = fmul float 4.0, %var ; yields float:result = 4.0 * %var
4755 '``udiv``' Instruction
4756 ^^^^^^^^^^^^^^^^^^^^^^
4763 <result> = udiv <ty> <op1>, <op2> ; yields ty:result
4764 <result> = udiv exact <ty> <op1>, <op2> ; yields ty:result
4769 The '``udiv``' instruction returns the quotient of its two operands.
4774 The two arguments to the '``udiv``' instruction must be
4775 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4776 arguments must have identical types.
4781 The value produced is the unsigned integer quotient of the two operands.
4783 Note that unsigned integer division and signed integer division are
4784 distinct operations; for signed integer division, use '``sdiv``'.
4786 Division by zero leads to undefined behavior.
4788 If the ``exact`` keyword is present, the result value of the ``udiv`` is
4789 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
4790 such, "((a udiv exact b) mul b) == a").
4795 .. code-block:: llvm
4797 <result> = udiv i32 4, %var ; yields i32:result = 4 / %var
4799 '``sdiv``' Instruction
4800 ^^^^^^^^^^^^^^^^^^^^^^
4807 <result> = sdiv <ty> <op1>, <op2> ; yields ty:result
4808 <result> = sdiv exact <ty> <op1>, <op2> ; yields ty:result
4813 The '``sdiv``' instruction returns the quotient of its two operands.
4818 The two arguments to the '``sdiv``' instruction must be
4819 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4820 arguments must have identical types.
4825 The value produced is the signed integer quotient of the two operands
4826 rounded towards zero.
4828 Note that signed integer division and unsigned integer division are
4829 distinct operations; for unsigned integer division, use '``udiv``'.
4831 Division by zero leads to undefined behavior. Overflow also leads to
4832 undefined behavior; this is a rare case, but can occur, for example, by
4833 doing a 32-bit division of -2147483648 by -1.
4835 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
4836 a :ref:`poison value <poisonvalues>` if the result would be rounded.
4841 .. code-block:: llvm
4843 <result> = sdiv i32 4, %var ; yields i32:result = 4 / %var
4847 '``fdiv``' Instruction
4848 ^^^^^^^^^^^^^^^^^^^^^^
4855 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4860 The '``fdiv``' instruction returns the quotient of its two operands.
4865 The two arguments to the '``fdiv``' instruction must be :ref:`floating
4866 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4867 Both arguments must have identical types.
4872 The value produced is the floating point quotient of the two operands.
4873 This instruction can also take any number of :ref:`fast-math
4874 flags <fastmath>`, which are optimization hints to enable otherwise
4875 unsafe floating point optimizations:
4880 .. code-block:: llvm
4882 <result> = fdiv float 4.0, %var ; yields float:result = 4.0 / %var
4884 '``urem``' Instruction
4885 ^^^^^^^^^^^^^^^^^^^^^^
4892 <result> = urem <ty> <op1>, <op2> ; yields ty:result
4897 The '``urem``' instruction returns the remainder from the unsigned
4898 division of its two arguments.
4903 The two arguments to the '``urem``' instruction must be
4904 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4905 arguments must have identical types.
4910 This instruction returns the unsigned integer *remainder* of a division.
4911 This instruction always performs an unsigned division to get the
4914 Note that unsigned integer remainder and signed integer remainder are
4915 distinct operations; for signed integer remainder, use '``srem``'.
4917 Taking the remainder of a division by zero leads to undefined behavior.
4922 .. code-block:: llvm
4924 <result> = urem i32 4, %var ; yields i32:result = 4 % %var
4926 '``srem``' Instruction
4927 ^^^^^^^^^^^^^^^^^^^^^^
4934 <result> = srem <ty> <op1>, <op2> ; yields ty:result
4939 The '``srem``' instruction returns the remainder from the signed
4940 division of its two operands. This instruction can also take
4941 :ref:`vector <t_vector>` versions of the values in which case the elements
4947 The two arguments to the '``srem``' instruction must be
4948 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4949 arguments must have identical types.
4954 This instruction returns the *remainder* of a division (where the result
4955 is either zero or has the same sign as the dividend, ``op1``), not the
4956 *modulo* operator (where the result is either zero or has the same sign
4957 as the divisor, ``op2``) of a value. For more information about the
4958 difference, see `The Math
4959 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
4960 table of how this is implemented in various languages, please see
4962 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
4964 Note that signed integer remainder and unsigned integer remainder are
4965 distinct operations; for unsigned integer remainder, use '``urem``'.
4967 Taking the remainder of a division by zero leads to undefined behavior.
4968 Overflow also leads to undefined behavior; this is a rare case, but can
4969 occur, for example, by taking the remainder of a 32-bit division of
4970 -2147483648 by -1. (The remainder doesn't actually overflow, but this
4971 rule lets srem be implemented using instructions that return both the
4972 result of the division and the remainder.)
4977 .. code-block:: llvm
4979 <result> = srem i32 4, %var ; yields i32:result = 4 % %var
4983 '``frem``' Instruction
4984 ^^^^^^^^^^^^^^^^^^^^^^
4991 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4996 The '``frem``' instruction returns the remainder from the division of
5002 The two arguments to the '``frem``' instruction must be :ref:`floating
5003 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
5004 Both arguments must have identical types.
5009 This instruction returns the *remainder* of a division. The remainder
5010 has the same sign as the dividend. This instruction can also take any
5011 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
5012 to enable otherwise unsafe floating point optimizations:
5017 .. code-block:: llvm
5019 <result> = frem float 4.0, %var ; yields float:result = 4.0 % %var
5023 Bitwise Binary Operations
5024 -------------------------
5026 Bitwise binary operators are used to do various forms of bit-twiddling
5027 in a program. They are generally very efficient instructions and can
5028 commonly be strength reduced from other instructions. They require two
5029 operands of the same type, execute an operation on them, and produce a
5030 single value. The resulting value is the same type as its operands.
5032 '``shl``' Instruction
5033 ^^^^^^^^^^^^^^^^^^^^^
5040 <result> = shl <ty> <op1>, <op2> ; yields ty:result
5041 <result> = shl nuw <ty> <op1>, <op2> ; yields ty:result
5042 <result> = shl nsw <ty> <op1>, <op2> ; yields ty:result
5043 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields ty:result
5048 The '``shl``' instruction returns the first operand shifted to the left
5049 a specified number of bits.
5054 Both arguments to the '``shl``' instruction must be the same
5055 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
5056 '``op2``' is treated as an unsigned value.
5061 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
5062 where ``n`` is the width of the result. If ``op2`` is (statically or
5063 dynamically) equal to or larger than the number of bits in
5064 ``op1``, the result is undefined. If the arguments are vectors, each
5065 vector element of ``op1`` is shifted by the corresponding shift amount
5068 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
5069 value <poisonvalues>` if it shifts out any non-zero bits. If the
5070 ``nsw`` keyword is present, then the shift produces a :ref:`poison
5071 value <poisonvalues>` if it shifts out any bits that disagree with the
5072 resultant sign bit. As such, NUW/NSW have the same semantics as they
5073 would if the shift were expressed as a mul instruction with the same
5074 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
5079 .. code-block:: llvm
5081 <result> = shl i32 4, %var ; yields i32: 4 << %var
5082 <result> = shl i32 4, 2 ; yields i32: 16
5083 <result> = shl i32 1, 10 ; yields i32: 1024
5084 <result> = shl i32 1, 32 ; undefined
5085 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
5087 '``lshr``' Instruction
5088 ^^^^^^^^^^^^^^^^^^^^^^
5095 <result> = lshr <ty> <op1>, <op2> ; yields ty:result
5096 <result> = lshr exact <ty> <op1>, <op2> ; yields ty:result
5101 The '``lshr``' instruction (logical shift right) returns the first
5102 operand shifted to the right a specified number of bits with zero fill.
5107 Both arguments to the '``lshr``' instruction must be the same
5108 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
5109 '``op2``' is treated as an unsigned value.
5114 This instruction always performs a logical shift right operation. The
5115 most significant bits of the result will be filled with zero bits after
5116 the shift. If ``op2`` is (statically or dynamically) equal to or larger
5117 than the number of bits in ``op1``, the result is undefined. If the
5118 arguments are vectors, each vector element of ``op1`` is shifted by the
5119 corresponding shift amount in ``op2``.
5121 If the ``exact`` keyword is present, the result value of the ``lshr`` is
5122 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
5128 .. code-block:: llvm
5130 <result> = lshr i32 4, 1 ; yields i32:result = 2
5131 <result> = lshr i32 4, 2 ; yields i32:result = 1
5132 <result> = lshr i8 4, 3 ; yields i8:result = 0
5133 <result> = lshr i8 -2, 1 ; yields i8:result = 0x7F
5134 <result> = lshr i32 1, 32 ; undefined
5135 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
5137 '``ashr``' Instruction
5138 ^^^^^^^^^^^^^^^^^^^^^^
5145 <result> = ashr <ty> <op1>, <op2> ; yields ty:result
5146 <result> = ashr exact <ty> <op1>, <op2> ; yields ty:result
5151 The '``ashr``' instruction (arithmetic shift right) returns the first
5152 operand shifted to the right a specified number of bits with sign
5158 Both arguments to the '``ashr``' instruction must be the same
5159 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
5160 '``op2``' is treated as an unsigned value.
5165 This instruction always performs an arithmetic shift right operation,
5166 The most significant bits of the result will be filled with the sign bit
5167 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
5168 than the number of bits in ``op1``, the result is undefined. If the
5169 arguments are vectors, each vector element of ``op1`` is shifted by the
5170 corresponding shift amount in ``op2``.
5172 If the ``exact`` keyword is present, the result value of the ``ashr`` is
5173 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
5179 .. code-block:: llvm
5181 <result> = ashr i32 4, 1 ; yields i32:result = 2
5182 <result> = ashr i32 4, 2 ; yields i32:result = 1
5183 <result> = ashr i8 4, 3 ; yields i8:result = 0
5184 <result> = ashr i8 -2, 1 ; yields i8:result = -1
5185 <result> = ashr i32 1, 32 ; undefined
5186 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
5188 '``and``' Instruction
5189 ^^^^^^^^^^^^^^^^^^^^^
5196 <result> = and <ty> <op1>, <op2> ; yields ty:result
5201 The '``and``' instruction returns the bitwise logical and of its two
5207 The two arguments to the '``and``' instruction must be
5208 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5209 arguments must have identical types.
5214 The truth table used for the '``and``' instruction is:
5231 .. code-block:: llvm
5233 <result> = and i32 4, %var ; yields i32:result = 4 & %var
5234 <result> = and i32 15, 40 ; yields i32:result = 8
5235 <result> = and i32 4, 8 ; yields i32:result = 0
5237 '``or``' Instruction
5238 ^^^^^^^^^^^^^^^^^^^^
5245 <result> = or <ty> <op1>, <op2> ; yields ty:result
5250 The '``or``' instruction returns the bitwise logical inclusive or of its
5256 The two arguments to the '``or``' instruction must be
5257 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5258 arguments must have identical types.
5263 The truth table used for the '``or``' instruction is:
5282 <result> = or i32 4, %var ; yields i32:result = 4 | %var
5283 <result> = or i32 15, 40 ; yields i32:result = 47
5284 <result> = or i32 4, 8 ; yields i32:result = 12
5286 '``xor``' Instruction
5287 ^^^^^^^^^^^^^^^^^^^^^
5294 <result> = xor <ty> <op1>, <op2> ; yields ty:result
5299 The '``xor``' instruction returns the bitwise logical exclusive or of
5300 its two operands. The ``xor`` is used to implement the "one's
5301 complement" operation, which is the "~" operator in C.
5306 The two arguments to the '``xor``' instruction must be
5307 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5308 arguments must have identical types.
5313 The truth table used for the '``xor``' instruction is:
5330 .. code-block:: llvm
5332 <result> = xor i32 4, %var ; yields i32:result = 4 ^ %var
5333 <result> = xor i32 15, 40 ; yields i32:result = 39
5334 <result> = xor i32 4, 8 ; yields i32:result = 12
5335 <result> = xor i32 %V, -1 ; yields i32:result = ~%V
5340 LLVM supports several instructions to represent vector operations in a
5341 target-independent manner. These instructions cover the element-access
5342 and vector-specific operations needed to process vectors effectively.
5343 While LLVM does directly support these vector operations, many
5344 sophisticated algorithms will want to use target-specific intrinsics to
5345 take full advantage of a specific target.
5347 .. _i_extractelement:
5349 '``extractelement``' Instruction
5350 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5357 <result> = extractelement <n x <ty>> <val>, <ty2> <idx> ; yields <ty>
5362 The '``extractelement``' instruction extracts a single scalar element
5363 from a vector at a specified index.
5368 The first operand of an '``extractelement``' instruction is a value of
5369 :ref:`vector <t_vector>` type. The second operand is an index indicating
5370 the position from which to extract the element. The index may be a
5371 variable of any integer type.
5376 The result is a scalar of the same type as the element type of ``val``.
5377 Its value is the value at position ``idx`` of ``val``. If ``idx``
5378 exceeds the length of ``val``, the results are undefined.
5383 .. code-block:: llvm
5385 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
5387 .. _i_insertelement:
5389 '``insertelement``' Instruction
5390 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5397 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, <ty2> <idx> ; yields <n x <ty>>
5402 The '``insertelement``' instruction inserts a scalar element into a
5403 vector at a specified index.
5408 The first operand of an '``insertelement``' instruction is a value of
5409 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
5410 type must equal the element type of the first operand. The third operand
5411 is an index indicating the position at which to insert the value. The
5412 index may be a variable of any integer type.
5417 The result is a vector of the same type as ``val``. Its element values
5418 are those of ``val`` except at position ``idx``, where it gets the value
5419 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
5425 .. code-block:: llvm
5427 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
5429 .. _i_shufflevector:
5431 '``shufflevector``' Instruction
5432 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5439 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
5444 The '``shufflevector``' instruction constructs a permutation of elements
5445 from two input vectors, returning a vector with the same element type as
5446 the input and length that is the same as the shuffle mask.
5451 The first two operands of a '``shufflevector``' instruction are vectors
5452 with the same type. The third argument is a shuffle mask whose element
5453 type is always 'i32'. The result of the instruction is a vector whose
5454 length is the same as the shuffle mask and whose element type is the
5455 same as the element type of the first two operands.
5457 The shuffle mask operand is required to be a constant vector with either
5458 constant integer or undef values.
5463 The elements of the two input vectors are numbered from left to right
5464 across both of the vectors. The shuffle mask operand specifies, for each
5465 element of the result vector, which element of the two input vectors the
5466 result element gets. The element selector may be undef (meaning "don't
5467 care") and the second operand may be undef if performing a shuffle from
5473 .. code-block:: llvm
5475 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
5476 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
5477 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
5478 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
5479 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
5480 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
5481 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
5482 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
5484 Aggregate Operations
5485 --------------------
5487 LLVM supports several instructions for working with
5488 :ref:`aggregate <t_aggregate>` values.
5492 '``extractvalue``' Instruction
5493 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5500 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
5505 The '``extractvalue``' instruction extracts the value of a member field
5506 from an :ref:`aggregate <t_aggregate>` value.
5511 The first operand of an '``extractvalue``' instruction is a value of
5512 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
5513 constant indices to specify which value to extract in a similar manner
5514 as indices in a '``getelementptr``' instruction.
5516 The major differences to ``getelementptr`` indexing are:
5518 - Since the value being indexed is not a pointer, the first index is
5519 omitted and assumed to be zero.
5520 - At least one index must be specified.
5521 - Not only struct indices but also array indices must be in bounds.
5526 The result is the value at the position in the aggregate specified by
5532 .. code-block:: llvm
5534 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
5538 '``insertvalue``' Instruction
5539 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5546 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
5551 The '``insertvalue``' instruction inserts a value into a member field in
5552 an :ref:`aggregate <t_aggregate>` value.
5557 The first operand of an '``insertvalue``' instruction is a value of
5558 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
5559 a first-class value to insert. The following operands are constant
5560 indices indicating the position at which to insert the value in a
5561 similar manner as indices in a '``extractvalue``' instruction. The value
5562 to insert must have the same type as the value identified by the
5568 The result is an aggregate of the same type as ``val``. Its value is
5569 that of ``val`` except that the value at the position specified by the
5570 indices is that of ``elt``.
5575 .. code-block:: llvm
5577 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
5578 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
5579 %agg3 = insertvalue {i32, {float}} undef, float %val, 1, 0 ; yields {i32 undef, {float %val}}
5583 Memory Access and Addressing Operations
5584 ---------------------------------------
5586 A key design point of an SSA-based representation is how it represents
5587 memory. In LLVM, no memory locations are in SSA form, which makes things
5588 very simple. This section describes how to read, write, and allocate
5593 '``alloca``' Instruction
5594 ^^^^^^^^^^^^^^^^^^^^^^^^
5601 <result> = alloca [inalloca] <type> [, <ty> <NumElements>] [, align <alignment>] ; yields type*:result
5606 The '``alloca``' instruction allocates memory on the stack frame of the
5607 currently executing function, to be automatically released when this
5608 function returns to its caller. The object is always allocated in the
5609 generic address space (address space zero).
5614 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
5615 bytes of memory on the runtime stack, returning a pointer of the
5616 appropriate type to the program. If "NumElements" is specified, it is
5617 the number of elements allocated, otherwise "NumElements" is defaulted
5618 to be one. If a constant alignment is specified, the value result of the
5619 allocation is guaranteed to be aligned to at least that boundary. The
5620 alignment may not be greater than ``1 << 29``. If not specified, or if
5621 zero, the target can choose to align the allocation on any convenient
5622 boundary compatible with the type.
5624 '``type``' may be any sized type.
5629 Memory is allocated; a pointer is returned. The operation is undefined
5630 if there is insufficient stack space for the allocation. '``alloca``'d
5631 memory is automatically released when the function returns. The
5632 '``alloca``' instruction is commonly used to represent automatic
5633 variables that must have an address available. When the function returns
5634 (either with the ``ret`` or ``resume`` instructions), the memory is
5635 reclaimed. Allocating zero bytes is legal, but the result is undefined.
5636 The order in which memory is allocated (ie., which way the stack grows)
5642 .. code-block:: llvm
5644 %ptr = alloca i32 ; yields i32*:ptr
5645 %ptr = alloca i32, i32 4 ; yields i32*:ptr
5646 %ptr = alloca i32, i32 4, align 1024 ; yields i32*:ptr
5647 %ptr = alloca i32, align 1024 ; yields i32*:ptr
5651 '``load``' Instruction
5652 ^^^^^^^^^^^^^^^^^^^^^^
5659 <result> = load [volatile] <ty>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>][, !nonnull !<index>]
5660 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
5661 !<index> = !{ i32 1 }
5666 The '``load``' instruction is used to read from memory.
5671 The argument to the ``load`` instruction specifies the memory address
5672 from which to load. The type specified must be a :ref:`first
5673 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
5674 then the optimizer is not allowed to modify the number or order of
5675 execution of this ``load`` with other :ref:`volatile
5676 operations <volatile>`.
5678 If the ``load`` is marked as ``atomic``, it takes an extra
5679 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5680 ``release`` and ``acq_rel`` orderings are not valid on ``load``
5681 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5682 when they may see multiple atomic stores. The type of the pointee must
5683 be an integer type whose bit width is a power of two greater than or
5684 equal to eight and less than or equal to a target-specific size limit.
5685 ``align`` must be explicitly specified on atomic loads, and the load has
5686 undefined behavior if the alignment is not set to a value which is at
5687 least the size in bytes of the pointee. ``!nontemporal`` does not have
5688 any defined semantics for atomic loads.
5690 The optional constant ``align`` argument specifies the alignment of the
5691 operation (that is, the alignment of the memory address). A value of 0
5692 or an omitted ``align`` argument means that the operation has the ABI
5693 alignment for the target. It is the responsibility of the code emitter
5694 to ensure that the alignment information is correct. Overestimating the
5695 alignment results in undefined behavior. Underestimating the alignment
5696 may produce less efficient code. An alignment of 1 is always safe. The
5697 maximum possible alignment is ``1 << 29``.
5699 The optional ``!nontemporal`` metadata must reference a single
5700 metadata name ``<index>`` corresponding to a metadata node with one
5701 ``i32`` entry of value 1. The existence of the ``!nontemporal``
5702 metadata on the instruction tells the optimizer and code generator
5703 that this load is not expected to be reused in the cache. The code
5704 generator may select special instructions to save cache bandwidth, such
5705 as the ``MOVNT`` instruction on x86.
5707 The optional ``!invariant.load`` metadata must reference a single
5708 metadata name ``<index>`` corresponding to a metadata node with no
5709 entries. The existence of the ``!invariant.load`` metadata on the
5710 instruction tells the optimizer and code generator that the address
5711 operand to this load points to memory which can be assumed unchanged.
5712 Being invariant does not imply that a location is dereferenceable,
5713 but it does imply that once the location is known dereferenceable
5714 its value is henceforth unchanging.
5716 The optional ``!nonnull`` metadata must reference a single
5717 metadata name ``<index>`` corresponding to a metadata node with no
5718 entries. The existence of the ``!nonnull`` metadata on the
5719 instruction tells the optimizer that the value loaded is known to
5720 never be null. This is analogous to the ''nonnull'' attribute
5721 on parameters and return values. This metadata can only be applied
5722 to loads of a pointer type.
5727 The location of memory pointed to is loaded. If the value being loaded
5728 is of scalar type then the number of bytes read does not exceed the
5729 minimum number of bytes needed to hold all bits of the type. For
5730 example, loading an ``i24`` reads at most three bytes. When loading a
5731 value of a type like ``i20`` with a size that is not an integral number
5732 of bytes, the result is undefined if the value was not originally
5733 written using a store of the same type.
5738 .. code-block:: llvm
5740 %ptr = alloca i32 ; yields i32*:ptr
5741 store i32 3, i32* %ptr ; yields void
5742 %val = load i32, i32* %ptr ; yields i32:val = i32 3
5746 '``store``' Instruction
5747 ^^^^^^^^^^^^^^^^^^^^^^^
5754 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields void
5755 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields void
5760 The '``store``' instruction is used to write to memory.
5765 There are two arguments to the ``store`` instruction: a value to store
5766 and an address at which to store it. The type of the ``<pointer>``
5767 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
5768 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
5769 then the optimizer is not allowed to modify the number or order of
5770 execution of this ``store`` with other :ref:`volatile
5771 operations <volatile>`.
5773 If the ``store`` is marked as ``atomic``, it takes an extra
5774 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5775 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
5776 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5777 when they may see multiple atomic stores. The type of the pointee must
5778 be an integer type whose bit width is a power of two greater than or
5779 equal to eight and less than or equal to a target-specific size limit.
5780 ``align`` must be explicitly specified on atomic stores, and the store
5781 has undefined behavior if the alignment is not set to a value which is
5782 at least the size in bytes of the pointee. ``!nontemporal`` does not
5783 have any defined semantics for atomic stores.
5785 The optional constant ``align`` argument specifies the alignment of the
5786 operation (that is, the alignment of the memory address). A value of 0
5787 or an omitted ``align`` argument means that the operation has the ABI
5788 alignment for the target. It is the responsibility of the code emitter
5789 to ensure that the alignment information is correct. Overestimating the
5790 alignment results in undefined behavior. Underestimating the
5791 alignment may produce less efficient code. An alignment of 1 is always
5792 safe. The maximum possible alignment is ``1 << 29``.
5794 The optional ``!nontemporal`` metadata must reference a single metadata
5795 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
5796 value 1. The existence of the ``!nontemporal`` metadata on the instruction
5797 tells the optimizer and code generator that this load is not expected to
5798 be reused in the cache. The code generator may select special
5799 instructions to save cache bandwidth, such as the MOVNT instruction on
5805 The contents of memory are updated to contain ``<value>`` at the
5806 location specified by the ``<pointer>`` operand. If ``<value>`` is
5807 of scalar type then the number of bytes written does not exceed the
5808 minimum number of bytes needed to hold all bits of the type. For
5809 example, storing an ``i24`` writes at most three bytes. When writing a
5810 value of a type like ``i20`` with a size that is not an integral number
5811 of bytes, it is unspecified what happens to the extra bits that do not
5812 belong to the type, but they will typically be overwritten.
5817 .. code-block:: llvm
5819 %ptr = alloca i32 ; yields i32*:ptr
5820 store i32 3, i32* %ptr ; yields void
5821 %val = load i32* %ptr ; yields i32:val = i32 3
5825 '``fence``' Instruction
5826 ^^^^^^^^^^^^^^^^^^^^^^^
5833 fence [singlethread] <ordering> ; yields void
5838 The '``fence``' instruction is used to introduce happens-before edges
5844 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
5845 defines what *synchronizes-with* edges they add. They can only be given
5846 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
5851 A fence A which has (at least) ``release`` ordering semantics
5852 *synchronizes with* a fence B with (at least) ``acquire`` ordering
5853 semantics if and only if there exist atomic operations X and Y, both
5854 operating on some atomic object M, such that A is sequenced before X, X
5855 modifies M (either directly or through some side effect of a sequence
5856 headed by X), Y is sequenced before B, and Y observes M. This provides a
5857 *happens-before* dependency between A and B. Rather than an explicit
5858 ``fence``, one (but not both) of the atomic operations X or Y might
5859 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
5860 still *synchronize-with* the explicit ``fence`` and establish the
5861 *happens-before* edge.
5863 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
5864 ``acquire`` and ``release`` semantics specified above, participates in
5865 the global program order of other ``seq_cst`` operations and/or fences.
5867 The optional ":ref:`singlethread <singlethread>`" argument specifies
5868 that the fence only synchronizes with other fences in the same thread.
5869 (This is useful for interacting with signal handlers.)
5874 .. code-block:: llvm
5876 fence acquire ; yields void
5877 fence singlethread seq_cst ; yields void
5881 '``cmpxchg``' Instruction
5882 ^^^^^^^^^^^^^^^^^^^^^^^^^
5889 cmpxchg [weak] [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <success ordering> <failure ordering> ; yields { ty, i1 }
5894 The '``cmpxchg``' instruction is used to atomically modify memory. It
5895 loads a value in memory and compares it to a given value. If they are
5896 equal, it tries to store a new value into the memory.
5901 There are three arguments to the '``cmpxchg``' instruction: an address
5902 to operate on, a value to compare to the value currently be at that
5903 address, and a new value to place at that address if the compared values
5904 are equal. The type of '<cmp>' must be an integer type whose bit width
5905 is a power of two greater than or equal to eight and less than or equal
5906 to a target-specific size limit. '<cmp>' and '<new>' must have the same
5907 type, and the type of '<pointer>' must be a pointer to that type. If the
5908 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
5909 to modify the number or order of execution of this ``cmpxchg`` with
5910 other :ref:`volatile operations <volatile>`.
5912 The success and failure :ref:`ordering <ordering>` arguments specify how this
5913 ``cmpxchg`` synchronizes with other atomic operations. Both ordering parameters
5914 must be at least ``monotonic``, the ordering constraint on failure must be no
5915 stronger than that on success, and the failure ordering cannot be either
5916 ``release`` or ``acq_rel``.
5918 The optional "``singlethread``" argument declares that the ``cmpxchg``
5919 is only atomic with respect to code (usually signal handlers) running in
5920 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
5921 respect to all other code in the system.
5923 The pointer passed into cmpxchg must have alignment greater than or
5924 equal to the size in memory of the operand.
5929 The contents of memory at the location specified by the '``<pointer>``' operand
5930 is read and compared to '``<cmp>``'; if the read value is the equal, the
5931 '``<new>``' is written. The original value at the location is returned, together
5932 with a flag indicating success (true) or failure (false).
5934 If the cmpxchg operation is marked as ``weak`` then a spurious failure is
5935 permitted: the operation may not write ``<new>`` even if the comparison
5938 If the cmpxchg operation is strong (the default), the i1 value is 1 if and only
5939 if the value loaded equals ``cmp``.
5941 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose of
5942 identifying release sequences. A failed ``cmpxchg`` is equivalent to an atomic
5943 load with an ordering parameter determined the second ordering parameter.
5948 .. code-block:: llvm
5951 %orig = atomic load i32, i32* %ptr unordered ; yields i32
5955 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
5956 %squared = mul i32 %cmp, %cmp
5957 %val_success = cmpxchg i32* %ptr, i32 %cmp, i32 %squared acq_rel monotonic ; yields { i32, i1 }
5958 %value_loaded = extractvalue { i32, i1 } %val_success, 0
5959 %success = extractvalue { i32, i1 } %val_success, 1
5960 br i1 %success, label %done, label %loop
5967 '``atomicrmw``' Instruction
5968 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
5975 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields ty
5980 The '``atomicrmw``' instruction is used to atomically modify memory.
5985 There are three arguments to the '``atomicrmw``' instruction: an
5986 operation to apply, an address whose value to modify, an argument to the
5987 operation. The operation must be one of the following keywords:
6001 The type of '<value>' must be an integer type whose bit width is a power
6002 of two greater than or equal to eight and less than or equal to a
6003 target-specific size limit. The type of the '``<pointer>``' operand must
6004 be a pointer to that type. If the ``atomicrmw`` is marked as
6005 ``volatile``, then the optimizer is not allowed to modify the number or
6006 order of execution of this ``atomicrmw`` with other :ref:`volatile
6007 operations <volatile>`.
6012 The contents of memory at the location specified by the '``<pointer>``'
6013 operand are atomically read, modified, and written back. The original
6014 value at the location is returned. The modification is specified by the
6017 - xchg: ``*ptr = val``
6018 - add: ``*ptr = *ptr + val``
6019 - sub: ``*ptr = *ptr - val``
6020 - and: ``*ptr = *ptr & val``
6021 - nand: ``*ptr = ~(*ptr & val)``
6022 - or: ``*ptr = *ptr | val``
6023 - xor: ``*ptr = *ptr ^ val``
6024 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
6025 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
6026 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
6028 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
6034 .. code-block:: llvm
6036 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields i32
6038 .. _i_getelementptr:
6040 '``getelementptr``' Instruction
6041 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6048 <result> = getelementptr <ty>, <ty>* <ptrval>{, <ty> <idx>}*
6049 <result> = getelementptr inbounds <ty>, <ty>* <ptrval>{, <ty> <idx>}*
6050 <result> = getelementptr <ty>, <ptr vector> <ptrval>, <vector index type> <idx>
6055 The '``getelementptr``' instruction is used to get the address of a
6056 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
6057 address calculation only and does not access memory.
6062 The first argument is always a type used as the basis for the calculations.
6063 The second argument is always a pointer or a vector of pointers, and is the
6064 base address to start from. The remaining arguments are indices
6065 that indicate which of the elements of the aggregate object are indexed.
6066 The interpretation of each index is dependent on the type being indexed
6067 into. The first index always indexes the pointer value given as the
6068 first argument, the second index indexes a value of the type pointed to
6069 (not necessarily the value directly pointed to, since the first index
6070 can be non-zero), etc. The first type indexed into must be a pointer
6071 value, subsequent types can be arrays, vectors, and structs. Note that
6072 subsequent types being indexed into can never be pointers, since that
6073 would require loading the pointer before continuing calculation.
6075 The type of each index argument depends on the type it is indexing into.
6076 When indexing into a (optionally packed) structure, only ``i32`` integer
6077 **constants** are allowed (when using a vector of indices they must all
6078 be the **same** ``i32`` integer constant). When indexing into an array,
6079 pointer or vector, integers of any width are allowed, and they are not
6080 required to be constant. These integers are treated as signed values
6083 For example, let's consider a C code fragment and how it gets compiled
6099 int *foo(struct ST *s) {
6100 return &s[1].Z.B[5][13];
6103 The LLVM code generated by Clang is:
6105 .. code-block:: llvm
6107 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
6108 %struct.ST = type { i32, double, %struct.RT }
6110 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
6112 %arrayidx = getelementptr inbounds %struct.ST, %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
6119 In the example above, the first index is indexing into the
6120 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
6121 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
6122 indexes into the third element of the structure, yielding a
6123 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
6124 structure. The third index indexes into the second element of the
6125 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
6126 dimensions of the array are subscripted into, yielding an '``i32``'
6127 type. The '``getelementptr``' instruction returns a pointer to this
6128 element, thus computing a value of '``i32*``' type.
6130 Note that it is perfectly legal to index partially through a structure,
6131 returning a pointer to an inner element. Because of this, the LLVM code
6132 for the given testcase is equivalent to:
6134 .. code-block:: llvm
6136 define i32* @foo(%struct.ST* %s) {
6137 %t1 = getelementptr %struct.ST, %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
6138 %t2 = getelementptr %struct.ST, %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
6139 %t3 = getelementptr %struct.RT, %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
6140 %t4 = getelementptr [10 x [20 x i32]], [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
6141 %t5 = getelementptr [20 x i32], [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
6145 If the ``inbounds`` keyword is present, the result value of the
6146 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
6147 pointer is not an *in bounds* address of an allocated object, or if any
6148 of the addresses that would be formed by successive addition of the
6149 offsets implied by the indices to the base address with infinitely
6150 precise signed arithmetic are not an *in bounds* address of that
6151 allocated object. The *in bounds* addresses for an allocated object are
6152 all the addresses that point into the object, plus the address one byte
6153 past the end. In cases where the base is a vector of pointers the
6154 ``inbounds`` keyword applies to each of the computations element-wise.
6156 If the ``inbounds`` keyword is not present, the offsets are added to the
6157 base address with silently-wrapping two's complement arithmetic. If the
6158 offsets have a different width from the pointer, they are sign-extended
6159 or truncated to the width of the pointer. The result value of the
6160 ``getelementptr`` may be outside the object pointed to by the base
6161 pointer. The result value may not necessarily be used to access memory
6162 though, even if it happens to point into allocated storage. See the
6163 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
6166 The getelementptr instruction is often confusing. For some more insight
6167 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
6172 .. code-block:: llvm
6174 ; yields [12 x i8]*:aptr
6175 %aptr = getelementptr {i32, [12 x i8]}, {i32, [12 x i8]}* %saptr, i64 0, i32 1
6177 %vptr = getelementptr {i32, <2 x i8>}, {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
6179 %eptr = getelementptr [12 x i8], [12 x i8]* %aptr, i64 0, i32 1
6181 %iptr = getelementptr [10 x i32], [10 x i32]* @arr, i16 0, i16 0
6183 In cases where the pointer argument is a vector of pointers, each index
6184 must be a vector with the same number of elements. For example:
6186 .. code-block:: llvm
6188 %A = getelementptr i8, <4 x i8*> %ptrs, <4 x i64> %offsets,
6190 Conversion Operations
6191 ---------------------
6193 The instructions in this category are the conversion instructions
6194 (casting) which all take a single operand and a type. They perform
6195 various bit conversions on the operand.
6197 '``trunc .. to``' Instruction
6198 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6205 <result> = trunc <ty> <value> to <ty2> ; yields ty2
6210 The '``trunc``' instruction truncates its operand to the type ``ty2``.
6215 The '``trunc``' instruction takes a value to trunc, and a type to trunc
6216 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
6217 of the same number of integers. The bit size of the ``value`` must be
6218 larger than the bit size of the destination type, ``ty2``. Equal sized
6219 types are not allowed.
6224 The '``trunc``' instruction truncates the high order bits in ``value``
6225 and converts the remaining bits to ``ty2``. Since the source size must
6226 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
6227 It will always truncate bits.
6232 .. code-block:: llvm
6234 %X = trunc i32 257 to i8 ; yields i8:1
6235 %Y = trunc i32 123 to i1 ; yields i1:true
6236 %Z = trunc i32 122 to i1 ; yields i1:false
6237 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
6239 '``zext .. to``' Instruction
6240 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6247 <result> = zext <ty> <value> to <ty2> ; yields ty2
6252 The '``zext``' instruction zero extends its operand to type ``ty2``.
6257 The '``zext``' instruction takes a value to cast, and a type to cast it
6258 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
6259 the same number of integers. The bit size of the ``value`` must be
6260 smaller than the bit size of the destination type, ``ty2``.
6265 The ``zext`` fills the high order bits of the ``value`` with zero bits
6266 until it reaches the size of the destination type, ``ty2``.
6268 When zero extending from i1, the result will always be either 0 or 1.
6273 .. code-block:: llvm
6275 %X = zext i32 257 to i64 ; yields i64:257
6276 %Y = zext i1 true to i32 ; yields i32:1
6277 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
6279 '``sext .. to``' Instruction
6280 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6287 <result> = sext <ty> <value> to <ty2> ; yields ty2
6292 The '``sext``' sign extends ``value`` to the type ``ty2``.
6297 The '``sext``' instruction takes a value to cast, and a type to cast it
6298 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
6299 the same number of integers. The bit size of the ``value`` must be
6300 smaller than the bit size of the destination type, ``ty2``.
6305 The '``sext``' instruction performs a sign extension by copying the sign
6306 bit (highest order bit) of the ``value`` until it reaches the bit size
6307 of the type ``ty2``.
6309 When sign extending from i1, the extension always results in -1 or 0.
6314 .. code-block:: llvm
6316 %X = sext i8 -1 to i16 ; yields i16 :65535
6317 %Y = sext i1 true to i32 ; yields i32:-1
6318 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
6320 '``fptrunc .. to``' Instruction
6321 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6328 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
6333 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
6338 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
6339 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
6340 The size of ``value`` must be larger than the size of ``ty2``. This
6341 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
6346 The '``fptrunc``' instruction truncates a ``value`` from a larger
6347 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
6348 point <t_floating>` type. If the value cannot fit within the
6349 destination type, ``ty2``, then the results are undefined.
6354 .. code-block:: llvm
6356 %X = fptrunc double 123.0 to float ; yields float:123.0
6357 %Y = fptrunc double 1.0E+300 to float ; yields undefined
6359 '``fpext .. to``' Instruction
6360 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6367 <result> = fpext <ty> <value> to <ty2> ; yields ty2
6372 The '``fpext``' extends a floating point ``value`` to a larger floating
6378 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
6379 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
6380 to. The source type must be smaller than the destination type.
6385 The '``fpext``' instruction extends the ``value`` from a smaller
6386 :ref:`floating point <t_floating>` type to a larger :ref:`floating
6387 point <t_floating>` type. The ``fpext`` cannot be used to make a
6388 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
6389 *no-op cast* for a floating point cast.
6394 .. code-block:: llvm
6396 %X = fpext float 3.125 to double ; yields double:3.125000e+00
6397 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
6399 '``fptoui .. to``' Instruction
6400 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6407 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
6412 The '``fptoui``' converts a floating point ``value`` to its unsigned
6413 integer equivalent of type ``ty2``.
6418 The '``fptoui``' instruction takes a value to cast, which must be a
6419 scalar or vector :ref:`floating point <t_floating>` value, and a type to
6420 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
6421 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
6422 type with the same number of elements as ``ty``
6427 The '``fptoui``' instruction converts its :ref:`floating
6428 point <t_floating>` operand into the nearest (rounding towards zero)
6429 unsigned integer value. If the value cannot fit in ``ty2``, the results
6435 .. code-block:: llvm
6437 %X = fptoui double 123.0 to i32 ; yields i32:123
6438 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
6439 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
6441 '``fptosi .. to``' Instruction
6442 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6449 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
6454 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
6455 ``value`` to type ``ty2``.
6460 The '``fptosi``' instruction takes a value to cast, which must be a
6461 scalar or vector :ref:`floating point <t_floating>` value, and a type to
6462 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
6463 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
6464 type with the same number of elements as ``ty``
6469 The '``fptosi``' instruction converts its :ref:`floating
6470 point <t_floating>` operand into the nearest (rounding towards zero)
6471 signed integer value. If the value cannot fit in ``ty2``, the results
6477 .. code-block:: llvm
6479 %X = fptosi double -123.0 to i32 ; yields i32:-123
6480 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
6481 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
6483 '``uitofp .. to``' Instruction
6484 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6491 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
6496 The '``uitofp``' instruction regards ``value`` as an unsigned integer
6497 and converts that value to the ``ty2`` type.
6502 The '``uitofp``' instruction takes a value to cast, which must be a
6503 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
6504 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
6505 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
6506 type with the same number of elements as ``ty``
6511 The '``uitofp``' instruction interprets its operand as an unsigned
6512 integer quantity and converts it to the corresponding floating point
6513 value. If the value cannot fit in the floating point value, the results
6519 .. code-block:: llvm
6521 %X = uitofp i32 257 to float ; yields float:257.0
6522 %Y = uitofp i8 -1 to double ; yields double:255.0
6524 '``sitofp .. to``' Instruction
6525 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6532 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
6537 The '``sitofp``' instruction regards ``value`` as a signed integer and
6538 converts that value to the ``ty2`` type.
6543 The '``sitofp``' instruction takes a value to cast, which must be a
6544 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
6545 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
6546 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
6547 type with the same number of elements as ``ty``
6552 The '``sitofp``' instruction interprets its operand as a signed integer
6553 quantity and converts it to the corresponding floating point value. If
6554 the value cannot fit in the floating point value, the results are
6560 .. code-block:: llvm
6562 %X = sitofp i32 257 to float ; yields float:257.0
6563 %Y = sitofp i8 -1 to double ; yields double:-1.0
6567 '``ptrtoint .. to``' Instruction
6568 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6575 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
6580 The '``ptrtoint``' instruction converts the pointer or a vector of
6581 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
6586 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
6587 a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
6588 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
6589 a vector of integers type.
6594 The '``ptrtoint``' instruction converts ``value`` to integer type
6595 ``ty2`` by interpreting the pointer value as an integer and either
6596 truncating or zero extending that value to the size of the integer type.
6597 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
6598 ``value`` is larger than ``ty2`` then a truncation is done. If they are
6599 the same size, then nothing is done (*no-op cast*) other than a type
6605 .. code-block:: llvm
6607 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
6608 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
6609 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
6613 '``inttoptr .. to``' Instruction
6614 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6621 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
6626 The '``inttoptr``' instruction converts an integer ``value`` to a
6627 pointer type, ``ty2``.
6632 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
6633 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
6639 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
6640 applying either a zero extension or a truncation depending on the size
6641 of the integer ``value``. If ``value`` is larger than the size of a
6642 pointer then a truncation is done. If ``value`` is smaller than the size
6643 of a pointer then a zero extension is done. If they are the same size,
6644 nothing is done (*no-op cast*).
6649 .. code-block:: llvm
6651 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
6652 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
6653 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
6654 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
6658 '``bitcast .. to``' Instruction
6659 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6666 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
6671 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
6677 The '``bitcast``' instruction takes a value to cast, which must be a
6678 non-aggregate first class value, and a type to cast it to, which must
6679 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
6680 bit sizes of ``value`` and the destination type, ``ty2``, must be
6681 identical. If the source type is a pointer, the destination type must
6682 also be a pointer of the same size. This instruction supports bitwise
6683 conversion of vectors to integers and to vectors of other types (as
6684 long as they have the same size).
6689 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
6690 is always a *no-op cast* because no bits change with this
6691 conversion. The conversion is done as if the ``value`` had been stored
6692 to memory and read back as type ``ty2``. Pointer (or vector of
6693 pointers) types may only be converted to other pointer (or vector of
6694 pointers) types with the same address space through this instruction.
6695 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
6696 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
6701 .. code-block:: llvm
6703 %X = bitcast i8 255 to i8 ; yields i8 :-1
6704 %Y = bitcast i32* %x to sint* ; yields sint*:%x
6705 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
6706 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
6708 .. _i_addrspacecast:
6710 '``addrspacecast .. to``' Instruction
6711 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6718 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
6723 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
6724 address space ``n`` to type ``pty2`` in address space ``m``.
6729 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
6730 to cast and a pointer type to cast it to, which must have a different
6736 The '``addrspacecast``' instruction converts the pointer value
6737 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
6738 value modification, depending on the target and the address space
6739 pair. Pointer conversions within the same address space must be
6740 performed with the ``bitcast`` instruction. Note that if the address space
6741 conversion is legal then both result and operand refer to the same memory
6747 .. code-block:: llvm
6749 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
6750 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
6751 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
6758 The instructions in this category are the "miscellaneous" instructions,
6759 which defy better classification.
6763 '``icmp``' Instruction
6764 ^^^^^^^^^^^^^^^^^^^^^^
6771 <result> = icmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6776 The '``icmp``' instruction returns a boolean value or a vector of
6777 boolean values based on comparison of its two integer, integer vector,
6778 pointer, or pointer vector operands.
6783 The '``icmp``' instruction takes three operands. The first operand is
6784 the condition code indicating the kind of comparison to perform. It is
6785 not a value, just a keyword. The possible condition code are:
6788 #. ``ne``: not equal
6789 #. ``ugt``: unsigned greater than
6790 #. ``uge``: unsigned greater or equal
6791 #. ``ult``: unsigned less than
6792 #. ``ule``: unsigned less or equal
6793 #. ``sgt``: signed greater than
6794 #. ``sge``: signed greater or equal
6795 #. ``slt``: signed less than
6796 #. ``sle``: signed less or equal
6798 The remaining two arguments must be :ref:`integer <t_integer>` or
6799 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
6800 must also be identical types.
6805 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
6806 code given as ``cond``. The comparison performed always yields either an
6807 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
6809 #. ``eq``: yields ``true`` if the operands are equal, ``false``
6810 otherwise. No sign interpretation is necessary or performed.
6811 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
6812 otherwise. No sign interpretation is necessary or performed.
6813 #. ``ugt``: interprets the operands as unsigned values and yields
6814 ``true`` if ``op1`` is greater than ``op2``.
6815 #. ``uge``: interprets the operands as unsigned values and yields
6816 ``true`` if ``op1`` is greater than or equal to ``op2``.
6817 #. ``ult``: interprets the operands as unsigned values and yields
6818 ``true`` if ``op1`` is less than ``op2``.
6819 #. ``ule``: interprets the operands as unsigned values and yields
6820 ``true`` if ``op1`` is less than or equal to ``op2``.
6821 #. ``sgt``: interprets the operands as signed values and yields ``true``
6822 if ``op1`` is greater than ``op2``.
6823 #. ``sge``: interprets the operands as signed values and yields ``true``
6824 if ``op1`` is greater than or equal to ``op2``.
6825 #. ``slt``: interprets the operands as signed values and yields ``true``
6826 if ``op1`` is less than ``op2``.
6827 #. ``sle``: interprets the operands as signed values and yields ``true``
6828 if ``op1`` is less than or equal to ``op2``.
6830 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
6831 are compared as if they were integers.
6833 If the operands are integer vectors, then they are compared element by
6834 element. The result is an ``i1`` vector with the same number of elements
6835 as the values being compared. Otherwise, the result is an ``i1``.
6840 .. code-block:: llvm
6842 <result> = icmp eq i32 4, 5 ; yields: result=false
6843 <result> = icmp ne float* %X, %X ; yields: result=false
6844 <result> = icmp ult i16 4, 5 ; yields: result=true
6845 <result> = icmp sgt i16 4, 5 ; yields: result=false
6846 <result> = icmp ule i16 -4, 5 ; yields: result=false
6847 <result> = icmp sge i16 4, 5 ; yields: result=false
6849 Note that the code generator does not yet support vector types with the
6850 ``icmp`` instruction.
6854 '``fcmp``' Instruction
6855 ^^^^^^^^^^^^^^^^^^^^^^
6862 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6867 The '``fcmp``' instruction returns a boolean value or vector of boolean
6868 values based on comparison of its operands.
6870 If the operands are floating point scalars, then the result type is a
6871 boolean (:ref:`i1 <t_integer>`).
6873 If the operands are floating point vectors, then the result type is a
6874 vector of boolean with the same number of elements as the operands being
6880 The '``fcmp``' instruction takes three operands. The first operand is
6881 the condition code indicating the kind of comparison to perform. It is
6882 not a value, just a keyword. The possible condition code are:
6884 #. ``false``: no comparison, always returns false
6885 #. ``oeq``: ordered and equal
6886 #. ``ogt``: ordered and greater than
6887 #. ``oge``: ordered and greater than or equal
6888 #. ``olt``: ordered and less than
6889 #. ``ole``: ordered and less than or equal
6890 #. ``one``: ordered and not equal
6891 #. ``ord``: ordered (no nans)
6892 #. ``ueq``: unordered or equal
6893 #. ``ugt``: unordered or greater than
6894 #. ``uge``: unordered or greater than or equal
6895 #. ``ult``: unordered or less than
6896 #. ``ule``: unordered or less than or equal
6897 #. ``une``: unordered or not equal
6898 #. ``uno``: unordered (either nans)
6899 #. ``true``: no comparison, always returns true
6901 *Ordered* means that neither operand is a QNAN while *unordered* means
6902 that either operand may be a QNAN.
6904 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
6905 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
6906 type. They must have identical types.
6911 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
6912 condition code given as ``cond``. If the operands are vectors, then the
6913 vectors are compared element by element. Each comparison performed
6914 always yields an :ref:`i1 <t_integer>` result, as follows:
6916 #. ``false``: always yields ``false``, regardless of operands.
6917 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
6918 is equal to ``op2``.
6919 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
6920 is greater than ``op2``.
6921 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
6922 is greater than or equal to ``op2``.
6923 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
6924 is less than ``op2``.
6925 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
6926 is less than or equal to ``op2``.
6927 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
6928 is not equal to ``op2``.
6929 #. ``ord``: yields ``true`` if both operands are not a QNAN.
6930 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
6932 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
6933 greater than ``op2``.
6934 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
6935 greater than or equal to ``op2``.
6936 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
6938 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
6939 less than or equal to ``op2``.
6940 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
6941 not equal to ``op2``.
6942 #. ``uno``: yields ``true`` if either operand is a QNAN.
6943 #. ``true``: always yields ``true``, regardless of operands.
6948 .. code-block:: llvm
6950 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
6951 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
6952 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
6953 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
6955 Note that the code generator does not yet support vector types with the
6956 ``fcmp`` instruction.
6960 '``phi``' Instruction
6961 ^^^^^^^^^^^^^^^^^^^^^
6968 <result> = phi <ty> [ <val0>, <label0>], ...
6973 The '``phi``' instruction is used to implement the φ node in the SSA
6974 graph representing the function.
6979 The type of the incoming values is specified with the first type field.
6980 After this, the '``phi``' instruction takes a list of pairs as
6981 arguments, with one pair for each predecessor basic block of the current
6982 block. Only values of :ref:`first class <t_firstclass>` type may be used as
6983 the value arguments to the PHI node. Only labels may be used as the
6986 There must be no non-phi instructions between the start of a basic block
6987 and the PHI instructions: i.e. PHI instructions must be first in a basic
6990 For the purposes of the SSA form, the use of each incoming value is
6991 deemed to occur on the edge from the corresponding predecessor block to
6992 the current block (but after any definition of an '``invoke``'
6993 instruction's return value on the same edge).
6998 At runtime, the '``phi``' instruction logically takes on the value
6999 specified by the pair corresponding to the predecessor basic block that
7000 executed just prior to the current block.
7005 .. code-block:: llvm
7007 Loop: ; Infinite loop that counts from 0 on up...
7008 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
7009 %nextindvar = add i32 %indvar, 1
7014 '``select``' Instruction
7015 ^^^^^^^^^^^^^^^^^^^^^^^^
7022 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
7024 selty is either i1 or {<N x i1>}
7029 The '``select``' instruction is used to choose one value based on a
7030 condition, without IR-level branching.
7035 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
7036 values indicating the condition, and two values of the same :ref:`first
7037 class <t_firstclass>` type.
7042 If the condition is an i1 and it evaluates to 1, the instruction returns
7043 the first value argument; otherwise, it returns the second value
7046 If the condition is a vector of i1, then the value arguments must be
7047 vectors of the same size, and the selection is done element by element.
7049 If the condition is an i1 and the value arguments are vectors of the
7050 same size, then an entire vector is selected.
7055 .. code-block:: llvm
7057 %X = select i1 true, i8 17, i8 42 ; yields i8:17
7061 '``call``' Instruction
7062 ^^^^^^^^^^^^^^^^^^^^^^
7069 <result> = [tail | musttail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
7074 The '``call``' instruction represents a simple function call.
7079 This instruction requires several arguments:
7081 #. The optional ``tail`` and ``musttail`` markers indicate that the optimizers
7082 should perform tail call optimization. The ``tail`` marker is a hint that
7083 `can be ignored <CodeGenerator.html#sibcallopt>`_. The ``musttail`` marker
7084 means that the call must be tail call optimized in order for the program to
7085 be correct. The ``musttail`` marker provides these guarantees:
7087 #. The call will not cause unbounded stack growth if it is part of a
7088 recursive cycle in the call graph.
7089 #. Arguments with the :ref:`inalloca <attr_inalloca>` attribute are
7092 Both markers imply that the callee does not access allocas or varargs from
7093 the caller. Calls marked ``musttail`` must obey the following additional
7096 - The call must immediately precede a :ref:`ret <i_ret>` instruction,
7097 or a pointer bitcast followed by a ret instruction.
7098 - The ret instruction must return the (possibly bitcasted) value
7099 produced by the call or void.
7100 - The caller and callee prototypes must match. Pointer types of
7101 parameters or return types may differ in pointee type, but not
7103 - The calling conventions of the caller and callee must match.
7104 - All ABI-impacting function attributes, such as sret, byval, inreg,
7105 returned, and inalloca, must match.
7106 - The callee must be varargs iff the caller is varargs. Bitcasting a
7107 non-varargs function to the appropriate varargs type is legal so
7108 long as the non-varargs prefixes obey the other rules.
7110 Tail call optimization for calls marked ``tail`` is guaranteed to occur if
7111 the following conditions are met:
7113 - Caller and callee both have the calling convention ``fastcc``.
7114 - The call is in tail position (ret immediately follows call and ret
7115 uses value of call or is void).
7116 - Option ``-tailcallopt`` is enabled, or
7117 ``llvm::GuaranteedTailCallOpt`` is ``true``.
7118 - `Platform-specific constraints are
7119 met. <CodeGenerator.html#tailcallopt>`_
7121 #. The optional "cconv" marker indicates which :ref:`calling
7122 convention <callingconv>` the call should use. If none is
7123 specified, the call defaults to using C calling conventions. The
7124 calling convention of the call must match the calling convention of
7125 the target function, or else the behavior is undefined.
7126 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
7127 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
7129 #. '``ty``': the type of the call instruction itself which is also the
7130 type of the return value. Functions that return no value are marked
7132 #. '``fnty``': shall be the signature of the pointer to function value
7133 being invoked. The argument types must match the types implied by
7134 this signature. This type can be omitted if the function is not
7135 varargs and if the function type does not return a pointer to a
7137 #. '``fnptrval``': An LLVM value containing a pointer to a function to
7138 be invoked. In most cases, this is a direct function invocation, but
7139 indirect ``call``'s are just as possible, calling an arbitrary pointer
7141 #. '``function args``': argument list whose types match the function
7142 signature argument types and parameter attributes. All arguments must
7143 be of :ref:`first class <t_firstclass>` type. If the function signature
7144 indicates the function accepts a variable number of arguments, the
7145 extra arguments can be specified.
7146 #. The optional :ref:`function attributes <fnattrs>` list. Only
7147 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
7148 attributes are valid here.
7153 The '``call``' instruction is used to cause control flow to transfer to
7154 a specified function, with its incoming arguments bound to the specified
7155 values. Upon a '``ret``' instruction in the called function, control
7156 flow continues with the instruction after the function call, and the
7157 return value of the function is bound to the result argument.
7162 .. code-block:: llvm
7164 %retval = call i32 @test(i32 %argc)
7165 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
7166 %X = tail call i32 @foo() ; yields i32
7167 %Y = tail call fastcc i32 @foo() ; yields i32
7168 call void %foo(i8 97 signext)
7170 %struct.A = type { i32, i8 }
7171 %r = call %struct.A @foo() ; yields { i32, i8 }
7172 %gr = extractvalue %struct.A %r, 0 ; yields i32
7173 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
7174 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
7175 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
7177 llvm treats calls to some functions with names and arguments that match
7178 the standard C99 library as being the C99 library functions, and may
7179 perform optimizations or generate code for them under that assumption.
7180 This is something we'd like to change in the future to provide better
7181 support for freestanding environments and non-C-based languages.
7185 '``va_arg``' Instruction
7186 ^^^^^^^^^^^^^^^^^^^^^^^^
7193 <resultval> = va_arg <va_list*> <arglist>, <argty>
7198 The '``va_arg``' instruction is used to access arguments passed through
7199 the "variable argument" area of a function call. It is used to implement
7200 the ``va_arg`` macro in C.
7205 This instruction takes a ``va_list*`` value and the type of the
7206 argument. It returns a value of the specified argument type and
7207 increments the ``va_list`` to point to the next argument. The actual
7208 type of ``va_list`` is target specific.
7213 The '``va_arg``' instruction loads an argument of the specified type
7214 from the specified ``va_list`` and causes the ``va_list`` to point to
7215 the next argument. For more information, see the variable argument
7216 handling :ref:`Intrinsic Functions <int_varargs>`.
7218 It is legal for this instruction to be called in a function which does
7219 not take a variable number of arguments, for example, the ``vfprintf``
7222 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
7223 function <intrinsics>` because it takes a type as an argument.
7228 See the :ref:`variable argument processing <int_varargs>` section.
7230 Note that the code generator does not yet fully support va\_arg on many
7231 targets. Also, it does not currently support va\_arg with aggregate
7232 types on any target.
7236 '``landingpad``' Instruction
7237 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7244 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
7245 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
7247 <clause> := catch <type> <value>
7248 <clause> := filter <array constant type> <array constant>
7253 The '``landingpad``' instruction is used by `LLVM's exception handling
7254 system <ExceptionHandling.html#overview>`_ to specify that a basic block
7255 is a landing pad --- one where the exception lands, and corresponds to the
7256 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
7257 defines values supplied by the personality function (``pers_fn``) upon
7258 re-entry to the function. The ``resultval`` has the type ``resultty``.
7263 This instruction takes a ``pers_fn`` value. This is the personality
7264 function associated with the unwinding mechanism. The optional
7265 ``cleanup`` flag indicates that the landing pad block is a cleanup.
7267 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
7268 contains the global variable representing the "type" that may be caught
7269 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
7270 clause takes an array constant as its argument. Use
7271 "``[0 x i8**] undef``" for a filter which cannot throw. The
7272 '``landingpad``' instruction must contain *at least* one ``clause`` or
7273 the ``cleanup`` flag.
7278 The '``landingpad``' instruction defines the values which are set by the
7279 personality function (``pers_fn``) upon re-entry to the function, and
7280 therefore the "result type" of the ``landingpad`` instruction. As with
7281 calling conventions, how the personality function results are
7282 represented in LLVM IR is target specific.
7284 The clauses are applied in order from top to bottom. If two
7285 ``landingpad`` instructions are merged together through inlining, the
7286 clauses from the calling function are appended to the list of clauses.
7287 When the call stack is being unwound due to an exception being thrown,
7288 the exception is compared against each ``clause`` in turn. If it doesn't
7289 match any of the clauses, and the ``cleanup`` flag is not set, then
7290 unwinding continues further up the call stack.
7292 The ``landingpad`` instruction has several restrictions:
7294 - A landing pad block is a basic block which is the unwind destination
7295 of an '``invoke``' instruction.
7296 - A landing pad block must have a '``landingpad``' instruction as its
7297 first non-PHI instruction.
7298 - There can be only one '``landingpad``' instruction within the landing
7300 - A basic block that is not a landing pad block may not include a
7301 '``landingpad``' instruction.
7302 - All '``landingpad``' instructions in a function must have the same
7303 personality function.
7308 .. code-block:: llvm
7310 ;; A landing pad which can catch an integer.
7311 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
7313 ;; A landing pad that is a cleanup.
7314 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
7316 ;; A landing pad which can catch an integer and can only throw a double.
7317 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
7319 filter [1 x i8**] [@_ZTId]
7326 LLVM supports the notion of an "intrinsic function". These functions
7327 have well known names and semantics and are required to follow certain
7328 restrictions. Overall, these intrinsics represent an extension mechanism
7329 for the LLVM language that does not require changing all of the
7330 transformations in LLVM when adding to the language (or the bitcode
7331 reader/writer, the parser, etc...).
7333 Intrinsic function names must all start with an "``llvm.``" prefix. This
7334 prefix is reserved in LLVM for intrinsic names; thus, function names may
7335 not begin with this prefix. Intrinsic functions must always be external
7336 functions: you cannot define the body of intrinsic functions. Intrinsic
7337 functions may only be used in call or invoke instructions: it is illegal
7338 to take the address of an intrinsic function. Additionally, because
7339 intrinsic functions are part of the LLVM language, it is required if any
7340 are added that they be documented here.
7342 Some intrinsic functions can be overloaded, i.e., the intrinsic
7343 represents a family of functions that perform the same operation but on
7344 different data types. Because LLVM can represent over 8 million
7345 different integer types, overloading is used commonly to allow an
7346 intrinsic function to operate on any integer type. One or more of the
7347 argument types or the result type can be overloaded to accept any
7348 integer type. Argument types may also be defined as exactly matching a
7349 previous argument's type or the result type. This allows an intrinsic
7350 function which accepts multiple arguments, but needs all of them to be
7351 of the same type, to only be overloaded with respect to a single
7352 argument or the result.
7354 Overloaded intrinsics will have the names of its overloaded argument
7355 types encoded into its function name, each preceded by a period. Only
7356 those types which are overloaded result in a name suffix. Arguments
7357 whose type is matched against another type do not. For example, the
7358 ``llvm.ctpop`` function can take an integer of any width and returns an
7359 integer of exactly the same integer width. This leads to a family of
7360 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
7361 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
7362 overloaded, and only one type suffix is required. Because the argument's
7363 type is matched against the return type, it does not require its own
7366 To learn how to add an intrinsic function, please see the `Extending
7367 LLVM Guide <ExtendingLLVM.html>`_.
7371 Variable Argument Handling Intrinsics
7372 -------------------------------------
7374 Variable argument support is defined in LLVM with the
7375 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
7376 functions. These functions are related to the similarly named macros
7377 defined in the ``<stdarg.h>`` header file.
7379 All of these functions operate on arguments that use a target-specific
7380 value type "``va_list``". The LLVM assembly language reference manual
7381 does not define what this type is, so all transformations should be
7382 prepared to handle these functions regardless of the type used.
7384 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
7385 variable argument handling intrinsic functions are used.
7387 .. code-block:: llvm
7389 ; This struct is different for every platform. For most platforms,
7390 ; it is merely an i8*.
7391 %struct.va_list = type { i8* }
7393 ; For Unix x86_64 platforms, va_list is the following struct:
7394 ; %struct.va_list = type { i32, i32, i8*, i8* }
7396 define i32 @test(i32 %X, ...) {
7397 ; Initialize variable argument processing
7398 %ap = alloca %struct.va_list
7399 %ap2 = bitcast %struct.va_list* %ap to i8*
7400 call void @llvm.va_start(i8* %ap2)
7402 ; Read a single integer argument
7403 %tmp = va_arg i8* %ap2, i32
7405 ; Demonstrate usage of llvm.va_copy and llvm.va_end
7407 %aq2 = bitcast i8** %aq to i8*
7408 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
7409 call void @llvm.va_end(i8* %aq2)
7411 ; Stop processing of arguments.
7412 call void @llvm.va_end(i8* %ap2)
7416 declare void @llvm.va_start(i8*)
7417 declare void @llvm.va_copy(i8*, i8*)
7418 declare void @llvm.va_end(i8*)
7422 '``llvm.va_start``' Intrinsic
7423 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7430 declare void @llvm.va_start(i8* <arglist>)
7435 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
7436 subsequent use by ``va_arg``.
7441 The argument is a pointer to a ``va_list`` element to initialize.
7446 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
7447 available in C. In a target-dependent way, it initializes the
7448 ``va_list`` element to which the argument points, so that the next call
7449 to ``va_arg`` will produce the first variable argument passed to the
7450 function. Unlike the C ``va_start`` macro, this intrinsic does not need
7451 to know the last argument of the function as the compiler can figure
7454 '``llvm.va_end``' Intrinsic
7455 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7462 declare void @llvm.va_end(i8* <arglist>)
7467 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
7468 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
7473 The argument is a pointer to a ``va_list`` to destroy.
7478 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
7479 available in C. In a target-dependent way, it destroys the ``va_list``
7480 element to which the argument points. Calls to
7481 :ref:`llvm.va_start <int_va_start>` and
7482 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
7487 '``llvm.va_copy``' Intrinsic
7488 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7495 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
7500 The '``llvm.va_copy``' intrinsic copies the current argument position
7501 from the source argument list to the destination argument list.
7506 The first argument is a pointer to a ``va_list`` element to initialize.
7507 The second argument is a pointer to a ``va_list`` element to copy from.
7512 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
7513 available in C. In a target-dependent way, it copies the source
7514 ``va_list`` element into the destination ``va_list`` element. This
7515 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
7516 arbitrarily complex and require, for example, memory allocation.
7518 Accurate Garbage Collection Intrinsics
7519 --------------------------------------
7521 LLVM's support for `Accurate Garbage Collection <GarbageCollection.html>`_
7522 (GC) requires the frontend to generate code containing appropriate intrinsic
7523 calls and select an appropriate GC strategy which knows how to lower these
7524 intrinsics in a manner which is appropriate for the target collector.
7526 These intrinsics allow identification of :ref:`GC roots on the
7527 stack <int_gcroot>`, as well as garbage collector implementations that
7528 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
7529 Frontends for type-safe garbage collected languages should generate
7530 these intrinsics to make use of the LLVM garbage collectors. For more
7531 details, see `Garbage Collection with LLVM <GarbageCollection.html>`_.
7533 Experimental Statepoint Intrinsics
7534 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7536 LLVM provides an second experimental set of intrinsics for describing garbage
7537 collection safepoints in compiled code. These intrinsics are an alternative
7538 to the ``llvm.gcroot`` intrinsics, but are compatible with the ones for
7539 :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers. The
7540 differences in approach are covered in the `Garbage Collection with LLVM
7541 <GarbageCollection.html>`_ documentation. The intrinsics themselves are
7542 described in :doc:`Statepoints`.
7546 '``llvm.gcroot``' Intrinsic
7547 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7554 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
7559 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
7560 the code generator, and allows some metadata to be associated with it.
7565 The first argument specifies the address of a stack object that contains
7566 the root pointer. The second pointer (which must be either a constant or
7567 a global value address) contains the meta-data to be associated with the
7573 At runtime, a call to this intrinsic stores a null pointer into the
7574 "ptrloc" location. At compile-time, the code generator generates
7575 information to allow the runtime to find the pointer at GC safe points.
7576 The '``llvm.gcroot``' intrinsic may only be used in a function which
7577 :ref:`specifies a GC algorithm <gc>`.
7581 '``llvm.gcread``' Intrinsic
7582 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7589 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
7594 The '``llvm.gcread``' intrinsic identifies reads of references from heap
7595 locations, allowing garbage collector implementations that require read
7601 The second argument is the address to read from, which should be an
7602 address allocated from the garbage collector. The first object is a
7603 pointer to the start of the referenced object, if needed by the language
7604 runtime (otherwise null).
7609 The '``llvm.gcread``' intrinsic has the same semantics as a load
7610 instruction, but may be replaced with substantially more complex code by
7611 the garbage collector runtime, as needed. The '``llvm.gcread``'
7612 intrinsic may only be used in a function which :ref:`specifies a GC
7617 '``llvm.gcwrite``' Intrinsic
7618 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7625 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
7630 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
7631 locations, allowing garbage collector implementations that require write
7632 barriers (such as generational or reference counting collectors).
7637 The first argument is the reference to store, the second is the start of
7638 the object to store it to, and the third is the address of the field of
7639 Obj to store to. If the runtime does not require a pointer to the
7640 object, Obj may be null.
7645 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
7646 instruction, but may be replaced with substantially more complex code by
7647 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
7648 intrinsic may only be used in a function which :ref:`specifies a GC
7651 Code Generator Intrinsics
7652 -------------------------
7654 These intrinsics are provided by LLVM to expose special features that
7655 may only be implemented with code generator support.
7657 '``llvm.returnaddress``' Intrinsic
7658 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7665 declare i8 *@llvm.returnaddress(i32 <level>)
7670 The '``llvm.returnaddress``' intrinsic attempts to compute a
7671 target-specific value indicating the return address of the current
7672 function or one of its callers.
7677 The argument to this intrinsic indicates which function to return the
7678 address for. Zero indicates the calling function, one indicates its
7679 caller, etc. The argument is **required** to be a constant integer
7685 The '``llvm.returnaddress``' intrinsic either returns a pointer
7686 indicating the return address of the specified call frame, or zero if it
7687 cannot be identified. The value returned by this intrinsic is likely to
7688 be incorrect or 0 for arguments other than zero, so it should only be
7689 used for debugging purposes.
7691 Note that calling this intrinsic does not prevent function inlining or
7692 other aggressive transformations, so the value returned may not be that
7693 of the obvious source-language caller.
7695 '``llvm.frameaddress``' Intrinsic
7696 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7703 declare i8* @llvm.frameaddress(i32 <level>)
7708 The '``llvm.frameaddress``' intrinsic attempts to return the
7709 target-specific frame pointer value for the specified stack frame.
7714 The argument to this intrinsic indicates which function to return the
7715 frame pointer for. Zero indicates the calling function, one indicates
7716 its caller, etc. The argument is **required** to be a constant integer
7722 The '``llvm.frameaddress``' intrinsic either returns a pointer
7723 indicating the frame address of the specified call frame, or zero if it
7724 cannot be identified. The value returned by this intrinsic is likely to
7725 be incorrect or 0 for arguments other than zero, so it should only be
7726 used for debugging purposes.
7728 Note that calling this intrinsic does not prevent function inlining or
7729 other aggressive transformations, so the value returned may not be that
7730 of the obvious source-language caller.
7732 '``llvm.frameescape``' and '``llvm.framerecover``' Intrinsics
7733 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7740 declare void @llvm.frameescape(...)
7741 declare i8* @llvm.framerecover(i8* %func, i8* %fp, i32 %idx)
7746 The '``llvm.frameescape``' intrinsic escapes offsets of a collection of static
7747 allocas, and the '``llvm.framerecover``' intrinsic applies those offsets to a
7748 live frame pointer to recover the address of the allocation. The offset is
7749 computed during frame layout of the caller of ``llvm.frameescape``.
7754 All arguments to '``llvm.frameescape``' must be pointers to static allocas or
7755 casts of static allocas. Each function can only call '``llvm.frameescape``'
7756 once, and it can only do so from the entry block.
7758 The ``func`` argument to '``llvm.framerecover``' must be a constant
7759 bitcasted pointer to a function defined in the current module. The code
7760 generator cannot determine the frame allocation offset of functions defined in
7763 The ``fp`` argument to '``llvm.framerecover``' must be a frame
7764 pointer of a call frame that is currently live. The return value of
7765 '``llvm.frameaddress``' is one way to produce such a value, but most platforms
7766 also expose the frame pointer through stack unwinding mechanisms.
7768 The ``idx`` argument to '``llvm.framerecover``' indicates which alloca passed to
7769 '``llvm.frameescape``' to recover. It is zero-indexed.
7774 These intrinsics allow a group of functions to access one stack memory
7775 allocation in an ancestor stack frame. The memory returned from
7776 '``llvm.frameallocate``' may be allocated prior to stack realignment, so the
7777 memory is only aligned to the ABI-required stack alignment. Each function may
7778 only call '``llvm.frameallocate``' one or zero times from the function entry
7779 block. The frame allocation intrinsic inhibits inlining, as any frame
7780 allocations in the inlined function frame are likely to be at a different
7781 offset from the one used by '``llvm.framerecover``' called with the
7784 .. _int_read_register:
7785 .. _int_write_register:
7787 '``llvm.read_register``' and '``llvm.write_register``' Intrinsics
7788 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7795 declare i32 @llvm.read_register.i32(metadata)
7796 declare i64 @llvm.read_register.i64(metadata)
7797 declare void @llvm.write_register.i32(metadata, i32 @value)
7798 declare void @llvm.write_register.i64(metadata, i64 @value)
7804 The '``llvm.read_register``' and '``llvm.write_register``' intrinsics
7805 provides access to the named register. The register must be valid on
7806 the architecture being compiled to. The type needs to be compatible
7807 with the register being read.
7812 The '``llvm.read_register``' intrinsic returns the current value of the
7813 register, where possible. The '``llvm.write_register``' intrinsic sets
7814 the current value of the register, where possible.
7816 This is useful to implement named register global variables that need
7817 to always be mapped to a specific register, as is common practice on
7818 bare-metal programs including OS kernels.
7820 The compiler doesn't check for register availability or use of the used
7821 register in surrounding code, including inline assembly. Because of that,
7822 allocatable registers are not supported.
7824 Warning: So far it only works with the stack pointer on selected
7825 architectures (ARM, AArch64, PowerPC and x86_64). Significant amount of
7826 work is needed to support other registers and even more so, allocatable
7831 '``llvm.stacksave``' Intrinsic
7832 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7839 declare i8* @llvm.stacksave()
7844 The '``llvm.stacksave``' intrinsic is used to remember the current state
7845 of the function stack, for use with
7846 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
7847 implementing language features like scoped automatic variable sized
7853 This intrinsic returns a opaque pointer value that can be passed to
7854 :ref:`llvm.stackrestore <int_stackrestore>`. When an
7855 ``llvm.stackrestore`` intrinsic is executed with a value saved from
7856 ``llvm.stacksave``, it effectively restores the state of the stack to
7857 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
7858 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
7859 were allocated after the ``llvm.stacksave`` was executed.
7861 .. _int_stackrestore:
7863 '``llvm.stackrestore``' Intrinsic
7864 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7871 declare void @llvm.stackrestore(i8* %ptr)
7876 The '``llvm.stackrestore``' intrinsic is used to restore the state of
7877 the function stack to the state it was in when the corresponding
7878 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
7879 useful for implementing language features like scoped automatic variable
7880 sized arrays in C99.
7885 See the description for :ref:`llvm.stacksave <int_stacksave>`.
7887 '``llvm.prefetch``' Intrinsic
7888 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7895 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
7900 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
7901 insert a prefetch instruction if supported; otherwise, it is a noop.
7902 Prefetches have no effect on the behavior of the program but can change
7903 its performance characteristics.
7908 ``address`` is the address to be prefetched, ``rw`` is the specifier
7909 determining if the fetch should be for a read (0) or write (1), and
7910 ``locality`` is a temporal locality specifier ranging from (0) - no
7911 locality, to (3) - extremely local keep in cache. The ``cache type``
7912 specifies whether the prefetch is performed on the data (1) or
7913 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
7914 arguments must be constant integers.
7919 This intrinsic does not modify the behavior of the program. In
7920 particular, prefetches cannot trap and do not produce a value. On
7921 targets that support this intrinsic, the prefetch can provide hints to
7922 the processor cache for better performance.
7924 '``llvm.pcmarker``' Intrinsic
7925 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7932 declare void @llvm.pcmarker(i32 <id>)
7937 The '``llvm.pcmarker``' intrinsic is a method to export a Program
7938 Counter (PC) in a region of code to simulators and other tools. The
7939 method is target specific, but it is expected that the marker will use
7940 exported symbols to transmit the PC of the marker. The marker makes no
7941 guarantees that it will remain with any specific instruction after
7942 optimizations. It is possible that the presence of a marker will inhibit
7943 optimizations. The intended use is to be inserted after optimizations to
7944 allow correlations of simulation runs.
7949 ``id`` is a numerical id identifying the marker.
7954 This intrinsic does not modify the behavior of the program. Backends
7955 that do not support this intrinsic may ignore it.
7957 '``llvm.readcyclecounter``' Intrinsic
7958 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7965 declare i64 @llvm.readcyclecounter()
7970 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
7971 counter register (or similar low latency, high accuracy clocks) on those
7972 targets that support it. On X86, it should map to RDTSC. On Alpha, it
7973 should map to RPCC. As the backing counters overflow quickly (on the
7974 order of 9 seconds on alpha), this should only be used for small
7980 When directly supported, reading the cycle counter should not modify any
7981 memory. Implementations are allowed to either return a application
7982 specific value or a system wide value. On backends without support, this
7983 is lowered to a constant 0.
7985 Note that runtime support may be conditional on the privilege-level code is
7986 running at and the host platform.
7988 '``llvm.clear_cache``' Intrinsic
7989 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7996 declare void @llvm.clear_cache(i8*, i8*)
8001 The '``llvm.clear_cache``' intrinsic ensures visibility of modifications
8002 in the specified range to the execution unit of the processor. On
8003 targets with non-unified instruction and data cache, the implementation
8004 flushes the instruction cache.
8009 On platforms with coherent instruction and data caches (e.g. x86), this
8010 intrinsic is a nop. On platforms with non-coherent instruction and data
8011 cache (e.g. ARM, MIPS), the intrinsic is lowered either to appropriate
8012 instructions or a system call, if cache flushing requires special
8015 The default behavior is to emit a call to ``__clear_cache`` from the run
8018 This instrinsic does *not* empty the instruction pipeline. Modifications
8019 of the current function are outside the scope of the intrinsic.
8021 '``llvm.instrprof_increment``' Intrinsic
8022 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8029 declare void @llvm.instrprof_increment(i8* <name>, i64 <hash>,
8030 i32 <num-counters>, i32 <index>)
8035 The '``llvm.instrprof_increment``' intrinsic can be emitted by a
8036 frontend for use with instrumentation based profiling. These will be
8037 lowered by the ``-instrprof`` pass to generate execution counts of a
8043 The first argument is a pointer to a global variable containing the
8044 name of the entity being instrumented. This should generally be the
8045 (mangled) function name for a set of counters.
8047 The second argument is a hash value that can be used by the consumer
8048 of the profile data to detect changes to the instrumented source, and
8049 the third is the number of counters associated with ``name``. It is an
8050 error if ``hash`` or ``num-counters`` differ between two instances of
8051 ``instrprof_increment`` that refer to the same name.
8053 The last argument refers to which of the counters for ``name`` should
8054 be incremented. It should be a value between 0 and ``num-counters``.
8059 This intrinsic represents an increment of a profiling counter. It will
8060 cause the ``-instrprof`` pass to generate the appropriate data
8061 structures and the code to increment the appropriate value, in a
8062 format that can be written out by a compiler runtime and consumed via
8063 the ``llvm-profdata`` tool.
8065 Standard C Library Intrinsics
8066 -----------------------------
8068 LLVM provides intrinsics for a few important standard C library
8069 functions. These intrinsics allow source-language front-ends to pass
8070 information about the alignment of the pointer arguments to the code
8071 generator, providing opportunity for more efficient code generation.
8075 '``llvm.memcpy``' Intrinsic
8076 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8081 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
8082 integer bit width and for different address spaces. Not all targets
8083 support all bit widths however.
8087 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
8088 i32 <len>, i32 <align>, i1 <isvolatile>)
8089 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
8090 i64 <len>, i32 <align>, i1 <isvolatile>)
8095 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
8096 source location to the destination location.
8098 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
8099 intrinsics do not return a value, takes extra alignment/isvolatile
8100 arguments and the pointers can be in specified address spaces.
8105 The first argument is a pointer to the destination, the second is a
8106 pointer to the source. The third argument is an integer argument
8107 specifying the number of bytes to copy, the fourth argument is the
8108 alignment of the source and destination locations, and the fifth is a
8109 boolean indicating a volatile access.
8111 If the call to this intrinsic has an alignment value that is not 0 or 1,
8112 then the caller guarantees that both the source and destination pointers
8113 are aligned to that boundary.
8115 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
8116 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
8117 very cleanly specified and it is unwise to depend on it.
8122 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
8123 source location to the destination location, which are not allowed to
8124 overlap. It copies "len" bytes of memory over. If the argument is known
8125 to be aligned to some boundary, this can be specified as the fourth
8126 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
8128 '``llvm.memmove``' Intrinsic
8129 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8134 This is an overloaded intrinsic. You can use llvm.memmove on any integer
8135 bit width and for different address space. Not all targets support all
8140 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
8141 i32 <len>, i32 <align>, i1 <isvolatile>)
8142 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
8143 i64 <len>, i32 <align>, i1 <isvolatile>)
8148 The '``llvm.memmove.*``' intrinsics move a block of memory from the
8149 source location to the destination location. It is similar to the
8150 '``llvm.memcpy``' intrinsic but allows the two memory locations to
8153 Note that, unlike the standard libc function, the ``llvm.memmove.*``
8154 intrinsics do not return a value, takes extra alignment/isvolatile
8155 arguments and the pointers can be in specified address spaces.
8160 The first argument is a pointer to the destination, the second is a
8161 pointer to the source. The third argument is an integer argument
8162 specifying the number of bytes to copy, the fourth argument is the
8163 alignment of the source and destination locations, and the fifth is a
8164 boolean indicating a volatile access.
8166 If the call to this intrinsic has an alignment value that is not 0 or 1,
8167 then the caller guarantees that the source and destination pointers are
8168 aligned to that boundary.
8170 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
8171 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
8172 not very cleanly specified and it is unwise to depend on it.
8177 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
8178 source location to the destination location, which may overlap. It
8179 copies "len" bytes of memory over. If the argument is known to be
8180 aligned to some boundary, this can be specified as the fourth argument,
8181 otherwise it should be set to 0 or 1 (both meaning no alignment).
8183 '``llvm.memset.*``' Intrinsics
8184 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8189 This is an overloaded intrinsic. You can use llvm.memset on any integer
8190 bit width and for different address spaces. However, not all targets
8191 support all bit widths.
8195 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
8196 i32 <len>, i32 <align>, i1 <isvolatile>)
8197 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
8198 i64 <len>, i32 <align>, i1 <isvolatile>)
8203 The '``llvm.memset.*``' intrinsics fill a block of memory with a
8204 particular byte value.
8206 Note that, unlike the standard libc function, the ``llvm.memset``
8207 intrinsic does not return a value and takes extra alignment/volatile
8208 arguments. Also, the destination can be in an arbitrary address space.
8213 The first argument is a pointer to the destination to fill, the second
8214 is the byte value with which to fill it, the third argument is an
8215 integer argument specifying the number of bytes to fill, and the fourth
8216 argument is the known alignment of the destination location.
8218 If the call to this intrinsic has an alignment value that is not 0 or 1,
8219 then the caller guarantees that the destination pointer is aligned to
8222 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
8223 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
8224 very cleanly specified and it is unwise to depend on it.
8229 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
8230 at the destination location. If the argument is known to be aligned to
8231 some boundary, this can be specified as the fourth argument, otherwise
8232 it should be set to 0 or 1 (both meaning no alignment).
8234 '``llvm.sqrt.*``' Intrinsic
8235 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8240 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
8241 floating point or vector of floating point type. Not all targets support
8246 declare float @llvm.sqrt.f32(float %Val)
8247 declare double @llvm.sqrt.f64(double %Val)
8248 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
8249 declare fp128 @llvm.sqrt.f128(fp128 %Val)
8250 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
8255 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
8256 returning the same value as the libm '``sqrt``' functions would. Unlike
8257 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
8258 negative numbers other than -0.0 (which allows for better optimization,
8259 because there is no need to worry about errno being set).
8260 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
8265 The argument and return value are floating point numbers of the same
8271 This function returns the sqrt of the specified operand if it is a
8272 nonnegative floating point number.
8274 '``llvm.powi.*``' Intrinsic
8275 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8280 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
8281 floating point or vector of floating point type. Not all targets support
8286 declare float @llvm.powi.f32(float %Val, i32 %power)
8287 declare double @llvm.powi.f64(double %Val, i32 %power)
8288 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
8289 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
8290 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
8295 The '``llvm.powi.*``' intrinsics return the first operand raised to the
8296 specified (positive or negative) power. The order of evaluation of
8297 multiplications is not defined. When a vector of floating point type is
8298 used, the second argument remains a scalar integer value.
8303 The second argument is an integer power, and the first is a value to
8304 raise to that power.
8309 This function returns the first value raised to the second power with an
8310 unspecified sequence of rounding operations.
8312 '``llvm.sin.*``' Intrinsic
8313 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8318 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
8319 floating point or vector of floating point type. Not all targets support
8324 declare float @llvm.sin.f32(float %Val)
8325 declare double @llvm.sin.f64(double %Val)
8326 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
8327 declare fp128 @llvm.sin.f128(fp128 %Val)
8328 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
8333 The '``llvm.sin.*``' intrinsics return the sine of the operand.
8338 The argument and return value are floating point numbers of the same
8344 This function returns the sine of the specified operand, returning the
8345 same values as the libm ``sin`` functions would, and handles error
8346 conditions in the same way.
8348 '``llvm.cos.*``' Intrinsic
8349 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8354 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
8355 floating point or vector of floating point type. Not all targets support
8360 declare float @llvm.cos.f32(float %Val)
8361 declare double @llvm.cos.f64(double %Val)
8362 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
8363 declare fp128 @llvm.cos.f128(fp128 %Val)
8364 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
8369 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
8374 The argument and return value are floating point numbers of the same
8380 This function returns the cosine of the specified operand, returning the
8381 same values as the libm ``cos`` functions would, and handles error
8382 conditions in the same way.
8384 '``llvm.pow.*``' Intrinsic
8385 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8390 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
8391 floating point or vector of floating point type. Not all targets support
8396 declare float @llvm.pow.f32(float %Val, float %Power)
8397 declare double @llvm.pow.f64(double %Val, double %Power)
8398 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
8399 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
8400 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
8405 The '``llvm.pow.*``' intrinsics return the first operand raised to the
8406 specified (positive or negative) power.
8411 The second argument is a floating point power, and the first is a value
8412 to raise to that power.
8417 This function returns the first value raised to the second power,
8418 returning the same values as the libm ``pow`` functions would, and
8419 handles error conditions in the same way.
8421 '``llvm.exp.*``' Intrinsic
8422 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8427 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
8428 floating point or vector of floating point type. Not all targets support
8433 declare float @llvm.exp.f32(float %Val)
8434 declare double @llvm.exp.f64(double %Val)
8435 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
8436 declare fp128 @llvm.exp.f128(fp128 %Val)
8437 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
8442 The '``llvm.exp.*``' intrinsics perform the exp function.
8447 The argument and return value are floating point numbers of the same
8453 This function returns the same values as the libm ``exp`` functions
8454 would, and handles error conditions in the same way.
8456 '``llvm.exp2.*``' Intrinsic
8457 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8462 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
8463 floating point or vector of floating point type. Not all targets support
8468 declare float @llvm.exp2.f32(float %Val)
8469 declare double @llvm.exp2.f64(double %Val)
8470 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
8471 declare fp128 @llvm.exp2.f128(fp128 %Val)
8472 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
8477 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
8482 The argument and return value are floating point numbers of the same
8488 This function returns the same values as the libm ``exp2`` functions
8489 would, and handles error conditions in the same way.
8491 '``llvm.log.*``' Intrinsic
8492 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8497 This is an overloaded intrinsic. You can use ``llvm.log`` on any
8498 floating point or vector of floating point type. Not all targets support
8503 declare float @llvm.log.f32(float %Val)
8504 declare double @llvm.log.f64(double %Val)
8505 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
8506 declare fp128 @llvm.log.f128(fp128 %Val)
8507 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
8512 The '``llvm.log.*``' intrinsics perform the log function.
8517 The argument and return value are floating point numbers of the same
8523 This function returns the same values as the libm ``log`` functions
8524 would, and handles error conditions in the same way.
8526 '``llvm.log10.*``' Intrinsic
8527 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8532 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
8533 floating point or vector of floating point type. Not all targets support
8538 declare float @llvm.log10.f32(float %Val)
8539 declare double @llvm.log10.f64(double %Val)
8540 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
8541 declare fp128 @llvm.log10.f128(fp128 %Val)
8542 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
8547 The '``llvm.log10.*``' intrinsics perform the log10 function.
8552 The argument and return value are floating point numbers of the same
8558 This function returns the same values as the libm ``log10`` functions
8559 would, and handles error conditions in the same way.
8561 '``llvm.log2.*``' Intrinsic
8562 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8567 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
8568 floating point or vector of floating point type. Not all targets support
8573 declare float @llvm.log2.f32(float %Val)
8574 declare double @llvm.log2.f64(double %Val)
8575 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
8576 declare fp128 @llvm.log2.f128(fp128 %Val)
8577 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
8582 The '``llvm.log2.*``' intrinsics perform the log2 function.
8587 The argument and return value are floating point numbers of the same
8593 This function returns the same values as the libm ``log2`` functions
8594 would, and handles error conditions in the same way.
8596 '``llvm.fma.*``' Intrinsic
8597 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8602 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
8603 floating point or vector of floating point type. Not all targets support
8608 declare float @llvm.fma.f32(float %a, float %b, float %c)
8609 declare double @llvm.fma.f64(double %a, double %b, double %c)
8610 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
8611 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
8612 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
8617 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
8623 The argument and return value are floating point numbers of the same
8629 This function returns the same values as the libm ``fma`` functions
8630 would, and does not set errno.
8632 '``llvm.fabs.*``' Intrinsic
8633 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8638 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
8639 floating point or vector of floating point type. Not all targets support
8644 declare float @llvm.fabs.f32(float %Val)
8645 declare double @llvm.fabs.f64(double %Val)
8646 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
8647 declare fp128 @llvm.fabs.f128(fp128 %Val)
8648 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
8653 The '``llvm.fabs.*``' intrinsics return the absolute value of the
8659 The argument and return value are floating point numbers of the same
8665 This function returns the same values as the libm ``fabs`` functions
8666 would, and handles error conditions in the same way.
8668 '``llvm.minnum.*``' Intrinsic
8669 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8674 This is an overloaded intrinsic. You can use ``llvm.minnum`` on any
8675 floating point or vector of floating point type. Not all targets support
8680 declare float @llvm.minnum.f32(float %Val0, float %Val1)
8681 declare double @llvm.minnum.f64(double %Val0, double %Val1)
8682 declare x86_fp80 @llvm.minnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
8683 declare fp128 @llvm.minnum.f128(fp128 %Val0, fp128 %Val1)
8684 declare ppc_fp128 @llvm.minnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
8689 The '``llvm.minnum.*``' intrinsics return the minimum of the two
8696 The arguments and return value are floating point numbers of the same
8702 Follows the IEEE-754 semantics for minNum, which also match for libm's
8705 If either operand is a NaN, returns the other non-NaN operand. Returns
8706 NaN only if both operands are NaN. If the operands compare equal,
8707 returns a value that compares equal to both operands. This means that
8708 fmin(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
8710 '``llvm.maxnum.*``' Intrinsic
8711 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8716 This is an overloaded intrinsic. You can use ``llvm.maxnum`` on any
8717 floating point or vector of floating point type. Not all targets support
8722 declare float @llvm.maxnum.f32(float %Val0, float %Val1l)
8723 declare double @llvm.maxnum.f64(double %Val0, double %Val1)
8724 declare x86_fp80 @llvm.maxnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
8725 declare fp128 @llvm.maxnum.f128(fp128 %Val0, fp128 %Val1)
8726 declare ppc_fp128 @llvm.maxnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
8731 The '``llvm.maxnum.*``' intrinsics return the maximum of the two
8738 The arguments and return value are floating point numbers of the same
8743 Follows the IEEE-754 semantics for maxNum, which also match for libm's
8746 If either operand is a NaN, returns the other non-NaN operand. Returns
8747 NaN only if both operands are NaN. If the operands compare equal,
8748 returns a value that compares equal to both operands. This means that
8749 fmax(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
8751 '``llvm.copysign.*``' Intrinsic
8752 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8757 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
8758 floating point or vector of floating point type. Not all targets support
8763 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
8764 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
8765 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
8766 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
8767 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
8772 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
8773 first operand and the sign of the second operand.
8778 The arguments and return value are floating point numbers of the same
8784 This function returns the same values as the libm ``copysign``
8785 functions would, and handles error conditions in the same way.
8787 '``llvm.floor.*``' Intrinsic
8788 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8793 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
8794 floating point or vector of floating point type. Not all targets support
8799 declare float @llvm.floor.f32(float %Val)
8800 declare double @llvm.floor.f64(double %Val)
8801 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
8802 declare fp128 @llvm.floor.f128(fp128 %Val)
8803 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
8808 The '``llvm.floor.*``' intrinsics return the floor of the operand.
8813 The argument and return value are floating point numbers of the same
8819 This function returns the same values as the libm ``floor`` functions
8820 would, and handles error conditions in the same way.
8822 '``llvm.ceil.*``' Intrinsic
8823 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8828 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
8829 floating point or vector of floating point type. Not all targets support
8834 declare float @llvm.ceil.f32(float %Val)
8835 declare double @llvm.ceil.f64(double %Val)
8836 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
8837 declare fp128 @llvm.ceil.f128(fp128 %Val)
8838 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
8843 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
8848 The argument and return value are floating point numbers of the same
8854 This function returns the same values as the libm ``ceil`` functions
8855 would, and handles error conditions in the same way.
8857 '``llvm.trunc.*``' Intrinsic
8858 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8863 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
8864 floating point or vector of floating point type. Not all targets support
8869 declare float @llvm.trunc.f32(float %Val)
8870 declare double @llvm.trunc.f64(double %Val)
8871 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
8872 declare fp128 @llvm.trunc.f128(fp128 %Val)
8873 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
8878 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
8879 nearest integer not larger in magnitude than the operand.
8884 The argument and return value are floating point numbers of the same
8890 This function returns the same values as the libm ``trunc`` functions
8891 would, and handles error conditions in the same way.
8893 '``llvm.rint.*``' Intrinsic
8894 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8899 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
8900 floating point or vector of floating point type. Not all targets support
8905 declare float @llvm.rint.f32(float %Val)
8906 declare double @llvm.rint.f64(double %Val)
8907 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
8908 declare fp128 @llvm.rint.f128(fp128 %Val)
8909 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
8914 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
8915 nearest integer. It may raise an inexact floating-point exception if the
8916 operand isn't an integer.
8921 The argument and return value are floating point numbers of the same
8927 This function returns the same values as the libm ``rint`` functions
8928 would, and handles error conditions in the same way.
8930 '``llvm.nearbyint.*``' Intrinsic
8931 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8936 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
8937 floating point or vector of floating point type. Not all targets support
8942 declare float @llvm.nearbyint.f32(float %Val)
8943 declare double @llvm.nearbyint.f64(double %Val)
8944 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
8945 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
8946 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
8951 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
8957 The argument and return value are floating point numbers of the same
8963 This function returns the same values as the libm ``nearbyint``
8964 functions would, and handles error conditions in the same way.
8966 '``llvm.round.*``' Intrinsic
8967 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8972 This is an overloaded intrinsic. You can use ``llvm.round`` on any
8973 floating point or vector of floating point type. Not all targets support
8978 declare float @llvm.round.f32(float %Val)
8979 declare double @llvm.round.f64(double %Val)
8980 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
8981 declare fp128 @llvm.round.f128(fp128 %Val)
8982 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
8987 The '``llvm.round.*``' intrinsics returns the operand rounded to the
8993 The argument and return value are floating point numbers of the same
8999 This function returns the same values as the libm ``round``
9000 functions would, and handles error conditions in the same way.
9002 Bit Manipulation Intrinsics
9003 ---------------------------
9005 LLVM provides intrinsics for a few important bit manipulation
9006 operations. These allow efficient code generation for some algorithms.
9008 '``llvm.bswap.*``' Intrinsics
9009 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9014 This is an overloaded intrinsic function. You can use bswap on any
9015 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
9019 declare i16 @llvm.bswap.i16(i16 <id>)
9020 declare i32 @llvm.bswap.i32(i32 <id>)
9021 declare i64 @llvm.bswap.i64(i64 <id>)
9026 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
9027 values with an even number of bytes (positive multiple of 16 bits).
9028 These are useful for performing operations on data that is not in the
9029 target's native byte order.
9034 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
9035 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
9036 intrinsic returns an i32 value that has the four bytes of the input i32
9037 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
9038 returned i32 will have its bytes in 3, 2, 1, 0 order. The
9039 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
9040 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
9043 '``llvm.ctpop.*``' Intrinsic
9044 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9049 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
9050 bit width, or on any vector with integer elements. Not all targets
9051 support all bit widths or vector types, however.
9055 declare i8 @llvm.ctpop.i8(i8 <src>)
9056 declare i16 @llvm.ctpop.i16(i16 <src>)
9057 declare i32 @llvm.ctpop.i32(i32 <src>)
9058 declare i64 @llvm.ctpop.i64(i64 <src>)
9059 declare i256 @llvm.ctpop.i256(i256 <src>)
9060 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
9065 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
9071 The only argument is the value to be counted. The argument may be of any
9072 integer type, or a vector with integer elements. The return type must
9073 match the argument type.
9078 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
9079 each element of a vector.
9081 '``llvm.ctlz.*``' Intrinsic
9082 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9087 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
9088 integer bit width, or any vector whose elements are integers. Not all
9089 targets support all bit widths or vector types, however.
9093 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
9094 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
9095 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
9096 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
9097 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
9098 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
9103 The '``llvm.ctlz``' family of intrinsic functions counts the number of
9104 leading zeros in a variable.
9109 The first argument is the value to be counted. This argument may be of
9110 any integer type, or a vector with integer element type. The return
9111 type must match the first argument type.
9113 The second argument must be a constant and is a flag to indicate whether
9114 the intrinsic should ensure that a zero as the first argument produces a
9115 defined result. Historically some architectures did not provide a
9116 defined result for zero values as efficiently, and many algorithms are
9117 now predicated on avoiding zero-value inputs.
9122 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
9123 zeros in a variable, or within each element of the vector. If
9124 ``src == 0`` then the result is the size in bits of the type of ``src``
9125 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
9126 ``llvm.ctlz(i32 2) = 30``.
9128 '``llvm.cttz.*``' Intrinsic
9129 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9134 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
9135 integer bit width, or any vector of integer elements. Not all targets
9136 support all bit widths or vector types, however.
9140 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
9141 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
9142 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
9143 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
9144 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
9145 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
9150 The '``llvm.cttz``' family of intrinsic functions counts the number of
9156 The first argument is the value to be counted. This argument may be of
9157 any integer type, or a vector with integer element type. The return
9158 type must match the first argument type.
9160 The second argument must be a constant and is a flag to indicate whether
9161 the intrinsic should ensure that a zero as the first argument produces a
9162 defined result. Historically some architectures did not provide a
9163 defined result for zero values as efficiently, and many algorithms are
9164 now predicated on avoiding zero-value inputs.
9169 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
9170 zeros in a variable, or within each element of a vector. If ``src == 0``
9171 then the result is the size in bits of the type of ``src`` if
9172 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
9173 ``llvm.cttz(2) = 1``.
9177 Arithmetic with Overflow Intrinsics
9178 -----------------------------------
9180 LLVM provides intrinsics for some arithmetic with overflow operations.
9182 '``llvm.sadd.with.overflow.*``' Intrinsics
9183 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9188 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
9189 on any integer bit width.
9193 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
9194 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
9195 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
9200 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
9201 a signed addition of the two arguments, and indicate whether an overflow
9202 occurred during the signed summation.
9207 The arguments (%a and %b) and the first element of the result structure
9208 may be of integer types of any bit width, but they must have the same
9209 bit width. The second element of the result structure must be of type
9210 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
9216 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
9217 a signed addition of the two variables. They return a structure --- the
9218 first element of which is the signed summation, and the second element
9219 of which is a bit specifying if the signed summation resulted in an
9225 .. code-block:: llvm
9227 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
9228 %sum = extractvalue {i32, i1} %res, 0
9229 %obit = extractvalue {i32, i1} %res, 1
9230 br i1 %obit, label %overflow, label %normal
9232 '``llvm.uadd.with.overflow.*``' Intrinsics
9233 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9238 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
9239 on any integer bit width.
9243 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
9244 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
9245 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
9250 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
9251 an unsigned addition of the two arguments, and indicate whether a carry
9252 occurred during the unsigned summation.
9257 The arguments (%a and %b) and the first element of the result structure
9258 may be of integer types of any bit width, but they must have the same
9259 bit width. The second element of the result structure must be of type
9260 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
9266 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
9267 an unsigned addition of the two arguments. They return a structure --- the
9268 first element of which is the sum, and the second element of which is a
9269 bit specifying if the unsigned summation resulted in a carry.
9274 .. code-block:: llvm
9276 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
9277 %sum = extractvalue {i32, i1} %res, 0
9278 %obit = extractvalue {i32, i1} %res, 1
9279 br i1 %obit, label %carry, label %normal
9281 '``llvm.ssub.with.overflow.*``' Intrinsics
9282 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9287 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
9288 on any integer bit width.
9292 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
9293 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
9294 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
9299 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
9300 a signed subtraction of the two arguments, and indicate whether an
9301 overflow occurred during the signed subtraction.
9306 The arguments (%a and %b) and the first element of the result structure
9307 may be of integer types of any bit width, but they must have the same
9308 bit width. The second element of the result structure must be of type
9309 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
9315 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
9316 a signed subtraction of the two arguments. They return a structure --- the
9317 first element of which is the subtraction, and the second element of
9318 which is a bit specifying if the signed subtraction resulted in an
9324 .. code-block:: llvm
9326 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
9327 %sum = extractvalue {i32, i1} %res, 0
9328 %obit = extractvalue {i32, i1} %res, 1
9329 br i1 %obit, label %overflow, label %normal
9331 '``llvm.usub.with.overflow.*``' Intrinsics
9332 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9337 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
9338 on any integer bit width.
9342 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
9343 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
9344 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
9349 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
9350 an unsigned subtraction of the two arguments, and indicate whether an
9351 overflow occurred during the unsigned subtraction.
9356 The arguments (%a and %b) and the first element of the result structure
9357 may be of integer types of any bit width, but they must have the same
9358 bit width. The second element of the result structure must be of type
9359 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
9365 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
9366 an unsigned subtraction of the two arguments. They return a structure ---
9367 the first element of which is the subtraction, and the second element of
9368 which is a bit specifying if the unsigned subtraction resulted in an
9374 .. code-block:: llvm
9376 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
9377 %sum = extractvalue {i32, i1} %res, 0
9378 %obit = extractvalue {i32, i1} %res, 1
9379 br i1 %obit, label %overflow, label %normal
9381 '``llvm.smul.with.overflow.*``' Intrinsics
9382 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9387 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
9388 on any integer bit width.
9392 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
9393 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
9394 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
9399 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
9400 a signed multiplication of the two arguments, and indicate whether an
9401 overflow occurred during the signed multiplication.
9406 The arguments (%a and %b) and the first element of the result structure
9407 may be of integer types of any bit width, but they must have the same
9408 bit width. The second element of the result structure must be of type
9409 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
9415 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
9416 a signed multiplication of the two arguments. They return a structure ---
9417 the first element of which is the multiplication, and the second element
9418 of which is a bit specifying if the signed multiplication resulted in an
9424 .. code-block:: llvm
9426 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
9427 %sum = extractvalue {i32, i1} %res, 0
9428 %obit = extractvalue {i32, i1} %res, 1
9429 br i1 %obit, label %overflow, label %normal
9431 '``llvm.umul.with.overflow.*``' Intrinsics
9432 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9437 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
9438 on any integer bit width.
9442 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
9443 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
9444 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
9449 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
9450 a unsigned multiplication of the two arguments, and indicate whether an
9451 overflow occurred during the unsigned multiplication.
9456 The arguments (%a and %b) and the first element of the result structure
9457 may be of integer types of any bit width, but they must have the same
9458 bit width. The second element of the result structure must be of type
9459 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
9465 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
9466 an unsigned multiplication of the two arguments. They return a structure ---
9467 the first element of which is the multiplication, and the second
9468 element of which is a bit specifying if the unsigned multiplication
9469 resulted in an overflow.
9474 .. code-block:: llvm
9476 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
9477 %sum = extractvalue {i32, i1} %res, 0
9478 %obit = extractvalue {i32, i1} %res, 1
9479 br i1 %obit, label %overflow, label %normal
9481 Specialised Arithmetic Intrinsics
9482 ---------------------------------
9484 '``llvm.fmuladd.*``' Intrinsic
9485 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9492 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
9493 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
9498 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
9499 expressions that can be fused if the code generator determines that (a) the
9500 target instruction set has support for a fused operation, and (b) that the
9501 fused operation is more efficient than the equivalent, separate pair of mul
9502 and add instructions.
9507 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
9508 multiplicands, a and b, and an addend c.
9517 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
9519 is equivalent to the expression a \* b + c, except that rounding will
9520 not be performed between the multiplication and addition steps if the
9521 code generator fuses the operations. Fusion is not guaranteed, even if
9522 the target platform supports it. If a fused multiply-add is required the
9523 corresponding llvm.fma.\* intrinsic function should be used
9524 instead. This never sets errno, just as '``llvm.fma.*``'.
9529 .. code-block:: llvm
9531 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields float:r2 = (a * b) + c
9533 Half Precision Floating Point Intrinsics
9534 ----------------------------------------
9536 For most target platforms, half precision floating point is a
9537 storage-only format. This means that it is a dense encoding (in memory)
9538 but does not support computation in the format.
9540 This means that code must first load the half-precision floating point
9541 value as an i16, then convert it to float with
9542 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
9543 then be performed on the float value (including extending to double
9544 etc). To store the value back to memory, it is first converted to float
9545 if needed, then converted to i16 with
9546 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
9549 .. _int_convert_to_fp16:
9551 '``llvm.convert.to.fp16``' Intrinsic
9552 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9559 declare i16 @llvm.convert.to.fp16.f32(float %a)
9560 declare i16 @llvm.convert.to.fp16.f64(double %a)
9565 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
9566 conventional floating point type to half precision floating point format.
9571 The intrinsic function contains single argument - the value to be
9577 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
9578 conventional floating point format to half precision floating point format. The
9579 return value is an ``i16`` which contains the converted number.
9584 .. code-block:: llvm
9586 %res = call i16 @llvm.convert.to.fp16.f32(float %a)
9587 store i16 %res, i16* @x, align 2
9589 .. _int_convert_from_fp16:
9591 '``llvm.convert.from.fp16``' Intrinsic
9592 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9599 declare float @llvm.convert.from.fp16.f32(i16 %a)
9600 declare double @llvm.convert.from.fp16.f64(i16 %a)
9605 The '``llvm.convert.from.fp16``' intrinsic function performs a
9606 conversion from half precision floating point format to single precision
9607 floating point format.
9612 The intrinsic function contains single argument - the value to be
9618 The '``llvm.convert.from.fp16``' intrinsic function performs a
9619 conversion from half single precision floating point format to single
9620 precision floating point format. The input half-float value is
9621 represented by an ``i16`` value.
9626 .. code-block:: llvm
9628 %a = load i16, i16* @x, align 2
9629 %res = call float @llvm.convert.from.fp16(i16 %a)
9636 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
9637 prefix), are described in the `LLVM Source Level
9638 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
9641 Exception Handling Intrinsics
9642 -----------------------------
9644 The LLVM exception handling intrinsics (which all start with
9645 ``llvm.eh.`` prefix), are described in the `LLVM Exception
9646 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
9650 Trampoline Intrinsics
9651 ---------------------
9653 These intrinsics make it possible to excise one parameter, marked with
9654 the :ref:`nest <nest>` attribute, from a function. The result is a
9655 callable function pointer lacking the nest parameter - the caller does
9656 not need to provide a value for it. Instead, the value to use is stored
9657 in advance in a "trampoline", a block of memory usually allocated on the
9658 stack, which also contains code to splice the nest value into the
9659 argument list. This is used to implement the GCC nested function address
9662 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
9663 then the resulting function pointer has signature ``i32 (i32, i32)*``.
9664 It can be created as follows:
9666 .. code-block:: llvm
9668 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
9669 %tramp1 = getelementptr [10 x i8], [10 x i8]* %tramp, i32 0, i32 0
9670 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
9671 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
9672 %fp = bitcast i8* %p to i32 (i32, i32)*
9674 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
9675 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
9679 '``llvm.init.trampoline``' Intrinsic
9680 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9687 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
9692 This fills the memory pointed to by ``tramp`` with executable code,
9693 turning it into a trampoline.
9698 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
9699 pointers. The ``tramp`` argument must point to a sufficiently large and
9700 sufficiently aligned block of memory; this memory is written to by the
9701 intrinsic. Note that the size and the alignment are target-specific -
9702 LLVM currently provides no portable way of determining them, so a
9703 front-end that generates this intrinsic needs to have some
9704 target-specific knowledge. The ``func`` argument must hold a function
9705 bitcast to an ``i8*``.
9710 The block of memory pointed to by ``tramp`` is filled with target
9711 dependent code, turning it into a function. Then ``tramp`` needs to be
9712 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
9713 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
9714 function's signature is the same as that of ``func`` with any arguments
9715 marked with the ``nest`` attribute removed. At most one such ``nest``
9716 argument is allowed, and it must be of pointer type. Calling the new
9717 function is equivalent to calling ``func`` with the same argument list,
9718 but with ``nval`` used for the missing ``nest`` argument. If, after
9719 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
9720 modified, then the effect of any later call to the returned function
9721 pointer is undefined.
9725 '``llvm.adjust.trampoline``' Intrinsic
9726 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9733 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
9738 This performs any required machine-specific adjustment to the address of
9739 a trampoline (passed as ``tramp``).
9744 ``tramp`` must point to a block of memory which already has trampoline
9745 code filled in by a previous call to
9746 :ref:`llvm.init.trampoline <int_it>`.
9751 On some architectures the address of the code to be executed needs to be
9752 different than the address where the trampoline is actually stored. This
9753 intrinsic returns the executable address corresponding to ``tramp``
9754 after performing the required machine specific adjustments. The pointer
9755 returned can then be :ref:`bitcast and executed <int_trampoline>`.
9757 Masked Vector Load and Store Intrinsics
9758 ---------------------------------------
9760 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.
9764 '``llvm.masked.load.*``' Intrinsics
9765 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9769 This is an overloaded intrinsic. The loaded data is a vector of any integer or floating point data type.
9773 declare <16 x float> @llvm.masked.load.v16f32 (<16 x float>* <ptr>, i32 <alignment>, <16 x i1> <mask>, <16 x float> <passthru>)
9774 declare <2 x double> @llvm.masked.load.v2f64 (<2 x double>* <ptr>, i32 <alignment>, <2 x i1> <mask>, <2 x double> <passthru>)
9779 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.
9785 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.
9791 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.
9792 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.
9797 %res = call <16 x float> @llvm.masked.load.v16f32 (<16 x float>* %ptr, i32 4, <16 x i1>%mask, <16 x float> %passthru)
9799 ;; The result of the two following instructions is identical aside from potential memory access exception
9800 %loadlal = load <16 x float>, <16 x float>* %ptr, align 4
9801 %res = select <16 x i1> %mask, <16 x float> %loadlal, <16 x float> %passthru
9805 '``llvm.masked.store.*``' Intrinsics
9806 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9810 This is an overloaded intrinsic. The data stored in memory is a vector of any integer or floating point data type.
9814 declare void @llvm.masked.store.v8i32 (<8 x i32> <value>, <8 x i32> * <ptr>, i32 <alignment>, <8 x i1> <mask>)
9815 declare void @llvm.masked.store.v16f32(<16 x i32> <value>, <16 x i32>* <ptr>, i32 <alignment>, <16 x i1> <mask>)
9820 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.
9825 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.
9831 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.
9832 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.
9836 call void @llvm.masked.store.v16f32(<16 x float> %value, <16 x float>* %ptr, i32 4, <16 x i1> %mask)
9838 ;; The result of the following instructions is identical aside from potential data races and memory access exceptions
9839 %oldval = load <16 x float>, <16 x float>* %ptr, align 4
9840 %res = select <16 x i1> %mask, <16 x float> %value, <16 x float> %oldval
9841 store <16 x float> %res, <16 x float>* %ptr, align 4
9847 This class of intrinsics provides information about the lifetime of
9848 memory objects and ranges where variables are immutable.
9852 '``llvm.lifetime.start``' Intrinsic
9853 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9860 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
9865 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
9871 The first argument is a constant integer representing the size of the
9872 object, or -1 if it is variable sized. The second argument is a pointer
9878 This intrinsic indicates that before this point in the code, the value
9879 of the memory pointed to by ``ptr`` is dead. This means that it is known
9880 to never be used and has an undefined value. A load from the pointer
9881 that precedes this intrinsic can be replaced with ``'undef'``.
9885 '``llvm.lifetime.end``' Intrinsic
9886 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9893 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
9898 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
9904 The first argument is a constant integer representing the size of the
9905 object, or -1 if it is variable sized. The second argument is a pointer
9911 This intrinsic indicates that after this point in the code, the value of
9912 the memory pointed to by ``ptr`` is dead. This means that it is known to
9913 never be used and has an undefined value. Any stores into the memory
9914 object following this intrinsic may be removed as dead.
9916 '``llvm.invariant.start``' Intrinsic
9917 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9924 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
9929 The '``llvm.invariant.start``' intrinsic specifies that the contents of
9930 a memory object will not change.
9935 The first argument is a constant integer representing the size of the
9936 object, or -1 if it is variable sized. The second argument is a pointer
9942 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
9943 the return value, the referenced memory location is constant and
9946 '``llvm.invariant.end``' Intrinsic
9947 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9954 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
9959 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
9960 memory object are mutable.
9965 The first argument is the matching ``llvm.invariant.start`` intrinsic.
9966 The second argument is a constant integer representing the size of the
9967 object, or -1 if it is variable sized and the third argument is a
9968 pointer to the object.
9973 This intrinsic indicates that the memory is mutable again.
9978 This class of intrinsics is designed to be generic and has no specific
9981 '``llvm.var.annotation``' Intrinsic
9982 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9989 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9994 The '``llvm.var.annotation``' intrinsic.
9999 The first argument is a pointer to a value, the second is a pointer to a
10000 global string, the third is a pointer to a global string which is the
10001 source file name, and the last argument is the line number.
10006 This intrinsic allows annotation of local variables with arbitrary
10007 strings. This can be useful for special purpose optimizations that want
10008 to look for these annotations. These have no other defined use; they are
10009 ignored by code generation and optimization.
10011 '``llvm.ptr.annotation.*``' Intrinsic
10012 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10017 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
10018 pointer to an integer of any width. *NOTE* you must specify an address space for
10019 the pointer. The identifier for the default address space is the integer
10024 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
10025 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
10026 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
10027 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
10028 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
10033 The '``llvm.ptr.annotation``' intrinsic.
10038 The first argument is a pointer to an integer value of arbitrary bitwidth
10039 (result of some expression), the second is a pointer to a global string, the
10040 third is a pointer to a global string which is the source file name, and the
10041 last argument is the line number. It returns the value of the first argument.
10046 This intrinsic allows annotation of a pointer to an integer with arbitrary
10047 strings. This can be useful for special purpose optimizations that want to look
10048 for these annotations. These have no other defined use; they are ignored by code
10049 generation and optimization.
10051 '``llvm.annotation.*``' Intrinsic
10052 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10057 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
10058 any integer bit width.
10062 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
10063 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
10064 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
10065 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
10066 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
10071 The '``llvm.annotation``' intrinsic.
10076 The first argument is an integer value (result of some expression), the
10077 second is a pointer to a global string, the third is a pointer to a
10078 global string which is the source file name, and the last argument is
10079 the line number. It returns the value of the first argument.
10084 This intrinsic allows annotations to be put on arbitrary expressions
10085 with arbitrary strings. This can be useful for special purpose
10086 optimizations that want to look for these annotations. These have no
10087 other defined use; they are ignored by code generation and optimization.
10089 '``llvm.trap``' Intrinsic
10090 ^^^^^^^^^^^^^^^^^^^^^^^^^
10097 declare void @llvm.trap() noreturn nounwind
10102 The '``llvm.trap``' intrinsic.
10112 This intrinsic is lowered to the target dependent trap instruction. If
10113 the target does not have a trap instruction, this intrinsic will be
10114 lowered to a call of the ``abort()`` function.
10116 '``llvm.debugtrap``' Intrinsic
10117 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10124 declare void @llvm.debugtrap() nounwind
10129 The '``llvm.debugtrap``' intrinsic.
10139 This intrinsic is lowered to code which is intended to cause an
10140 execution trap with the intention of requesting the attention of a
10143 '``llvm.stackprotector``' Intrinsic
10144 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10151 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
10156 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
10157 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
10158 is placed on the stack before local variables.
10163 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
10164 The first argument is the value loaded from the stack guard
10165 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
10166 enough space to hold the value of the guard.
10171 This intrinsic causes the prologue/epilogue inserter to force the position of
10172 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
10173 to ensure that if a local variable on the stack is overwritten, it will destroy
10174 the value of the guard. When the function exits, the guard on the stack is
10175 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
10176 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
10177 calling the ``__stack_chk_fail()`` function.
10179 '``llvm.stackprotectorcheck``' Intrinsic
10180 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10187 declare void @llvm.stackprotectorcheck(i8** <guard>)
10192 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
10193 created stack protector and if they are not equal calls the
10194 ``__stack_chk_fail()`` function.
10199 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
10200 the variable ``@__stack_chk_guard``.
10205 This intrinsic is provided to perform the stack protector check by comparing
10206 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
10207 values do not match call the ``__stack_chk_fail()`` function.
10209 The reason to provide this as an IR level intrinsic instead of implementing it
10210 via other IR operations is that in order to perform this operation at the IR
10211 level without an intrinsic, one would need to create additional basic blocks to
10212 handle the success/failure cases. This makes it difficult to stop the stack
10213 protector check from disrupting sibling tail calls in Codegen. With this
10214 intrinsic, we are able to generate the stack protector basic blocks late in
10215 codegen after the tail call decision has occurred.
10217 '``llvm.objectsize``' Intrinsic
10218 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10225 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
10226 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
10231 The ``llvm.objectsize`` intrinsic is designed to provide information to
10232 the optimizers to determine at compile time whether a) an operation
10233 (like memcpy) will overflow a buffer that corresponds to an object, or
10234 b) that a runtime check for overflow isn't necessary. An object in this
10235 context means an allocation of a specific class, structure, array, or
10241 The ``llvm.objectsize`` intrinsic takes two arguments. The first
10242 argument is a pointer to or into the ``object``. The second argument is
10243 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
10244 or -1 (if false) when the object size is unknown. The second argument
10245 only accepts constants.
10250 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
10251 the size of the object concerned. If the size cannot be determined at
10252 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
10253 on the ``min`` argument).
10255 '``llvm.expect``' Intrinsic
10256 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10261 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
10266 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
10267 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
10268 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
10273 The ``llvm.expect`` intrinsic provides information about expected (the
10274 most probable) value of ``val``, which can be used by optimizers.
10279 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
10280 a value. The second argument is an expected value, this needs to be a
10281 constant value, variables are not allowed.
10286 This intrinsic is lowered to the ``val``.
10288 '``llvm.assume``' Intrinsic
10289 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10296 declare void @llvm.assume(i1 %cond)
10301 The ``llvm.assume`` allows the optimizer to assume that the provided
10302 condition is true. This information can then be used in simplifying other parts
10308 The condition which the optimizer may assume is always true.
10313 The intrinsic allows the optimizer to assume that the provided condition is
10314 always true whenever the control flow reaches the intrinsic call. No code is
10315 generated for this intrinsic, and instructions that contribute only to the
10316 provided condition are not used for code generation. If the condition is
10317 violated during execution, the behavior is undefined.
10319 Note that the optimizer might limit the transformations performed on values
10320 used by the ``llvm.assume`` intrinsic in order to preserve the instructions
10321 only used to form the intrinsic's input argument. This might prove undesirable
10322 if the extra information provided by the ``llvm.assume`` intrinsic does not cause
10323 sufficient overall improvement in code quality. For this reason,
10324 ``llvm.assume`` should not be used to document basic mathematical invariants
10325 that the optimizer can otherwise deduce or facts that are of little use to the
10330 '``llvm.bitset.test``' Intrinsic
10331 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10338 declare i1 @llvm.bitset.test(i8* %ptr, metadata %bitset) nounwind readnone
10344 The first argument is a pointer to be tested. The second argument is a
10345 metadata string containing the name of a :doc:`bitset <BitSets>`.
10350 The ``llvm.bitset.test`` intrinsic tests whether the given pointer is a
10351 member of the given bitset.
10353 '``llvm.donothing``' Intrinsic
10354 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10361 declare void @llvm.donothing() nounwind readnone
10366 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's one of only
10367 two intrinsics (besides ``llvm.experimental.patchpoint``) that can be called
10368 with an invoke instruction.
10378 This intrinsic does nothing, and it's removed by optimizers and ignored
10381 Stack Map Intrinsics
10382 --------------------
10384 LLVM provides experimental intrinsics to support runtime patching
10385 mechanisms commonly desired in dynamic language JITs. These intrinsics
10386 are described in :doc:`StackMaps`.