1 ==============================
2 LLVM Language Reference Manual
3 ==============================
12 This document is a reference manual for the LLVM assembly language. LLVM
13 is a Static Single Assignment (SSA) based representation that provides
14 type safety, low-level operations, flexibility, and the capability of
15 representing 'all' high-level languages cleanly. It is the common code
16 representation used throughout all phases of the LLVM compilation
22 The LLVM code representation is designed to be used in three different
23 forms: as an in-memory compiler IR, as an on-disk bitcode representation
24 (suitable for fast loading by a Just-In-Time compiler), and as a human
25 readable assembly language representation. This allows LLVM to provide a
26 powerful intermediate representation for efficient compiler
27 transformations and analysis, while providing a natural means to debug
28 and visualize the transformations. The three different forms of LLVM are
29 all equivalent. This document describes the human readable
30 representation and notation.
32 The LLVM representation aims to be light-weight and low-level while
33 being expressive, typed, and extensible at the same time. It aims to be
34 a "universal IR" of sorts, by being at a low enough level that
35 high-level ideas may be cleanly mapped to it (similar to how
36 microprocessors are "universal IR's", allowing many source languages to
37 be mapped to them). By providing type information, LLVM can be used as
38 the target of optimizations: for example, through pointer analysis, it
39 can be proven that a C automatic variable is never accessed outside of
40 the current function, allowing it to be promoted to a simple SSA value
41 instead of a memory location.
48 It is important to note that this document describes 'well formed' LLVM
49 assembly language. There is a difference between what the parser accepts
50 and what is considered 'well formed'. For example, the following
51 instruction is syntactically okay, but not well formed:
57 because the definition of ``%x`` does not dominate all of its uses. The
58 LLVM infrastructure provides a verification pass that may be used to
59 verify that an LLVM module is well formed. This pass is automatically
60 run by the parser after parsing input assembly and by the optimizer
61 before it outputs bitcode. The violations pointed out by the verifier
62 pass indicate bugs in transformation passes or input to the parser.
69 LLVM identifiers come in two basic types: global and local. Global
70 identifiers (functions, global variables) begin with the ``'@'``
71 character. Local identifiers (register names, types) begin with the
72 ``'%'`` character. Additionally, there are three different formats for
73 identifiers, for different purposes:
75 #. Named values are represented as a string of characters with their
76 prefix. For example, ``%foo``, ``@DivisionByZero``,
77 ``%a.really.long.identifier``. The actual regular expression used is
78 '``[%@][a-zA-Z$._][a-zA-Z$._0-9]*``'. Identifiers which require other
79 characters in their names can be surrounded with quotes. Special
80 characters may be escaped using ``"\xx"`` where ``xx`` is the ASCII
81 code for the character in hexadecimal. In this way, any character can
82 be used in a name value, even quotes themselves.
83 #. Unnamed values are represented as an unsigned numeric value with
84 their prefix. For example, ``%12``, ``@2``, ``%44``.
85 #. Constants, which are described in the section Constants_ below.
87 LLVM requires that values start with a prefix for two reasons: Compilers
88 don't need to worry about name clashes with reserved words, and the set
89 of reserved words may be expanded in the future without penalty.
90 Additionally, unnamed identifiers allow a compiler to quickly come up
91 with a temporary variable without having to avoid symbol table
94 Reserved words in LLVM are very similar to reserved words in other
95 languages. There are keywords for different opcodes ('``add``',
96 '``bitcast``', '``ret``', etc...), for primitive type names ('``void``',
97 '``i32``', etc...), and others. These reserved words cannot conflict
98 with variable names, because none of them start with a prefix character
101 Here is an example of LLVM code to multiply the integer variable
108 %result = mul i32 %X, 8
110 After strength reduction:
114 %result = shl i32 %X, 3
120 %0 = add i32 %X, %X ; yields i32:%0
121 %1 = add i32 %0, %0 ; yields i32:%1
122 %result = add i32 %1, %1
124 This last way of multiplying ``%X`` by 8 illustrates several important
125 lexical features of LLVM:
127 #. Comments are delimited with a '``;``' and go until the end of line.
128 #. Unnamed temporaries are created when the result of a computation is
129 not assigned to a named value.
130 #. Unnamed temporaries are numbered sequentially (using a per-function
131 incrementing counter, starting with 0). Note that basic blocks are
132 included in this numbering. For example, if the entry basic block is not
133 given a label name, then it will get number 0.
135 It also shows a convention that we follow in this document. When
136 demonstrating instructions, we will follow an instruction with a comment
137 that defines the type and name of value produced.
145 LLVM programs are composed of ``Module``'s, each of which is a
146 translation unit of the input programs. Each module consists of
147 functions, global variables, and symbol table entries. Modules may be
148 combined together with the LLVM linker, which merges function (and
149 global variable) definitions, resolves forward declarations, and merges
150 symbol table entries. Here is an example of the "hello world" module:
154 ; Declare the string constant as a global constant.
155 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
157 ; External declaration of the puts function
158 declare i32 @puts(i8* nocapture) nounwind
160 ; Definition of main function
161 define i32 @main() { ; i32()*
162 ; Convert [13 x i8]* to i8 *...
163 %cast210 = getelementptr [13 x i8]* @.str, i64 0, i64 0
165 ; Call puts function to write out the string to stdout.
166 call i32 @puts(i8* %cast210)
171 !1 = metadata !{i32 42}
174 This example is made up of a :ref:`global variable <globalvars>` named
175 "``.str``", an external declaration of the "``puts``" function, a
176 :ref:`function definition <functionstructure>` for "``main``" and
177 :ref:`named metadata <namedmetadatastructure>` "``foo``".
179 In general, a module is made up of a list of global values (where both
180 functions and global variables are global values). Global values are
181 represented by a pointer to a memory location (in this case, a pointer
182 to an array of char, and a pointer to a function), and have one of the
183 following :ref:`linkage types <linkage>`.
190 All Global Variables and Functions have one of the following types of
194 Global values with "``private``" linkage are only directly
195 accessible by objects in the current module. In particular, linking
196 code into a module with an private global value may cause the
197 private to be renamed as necessary to avoid collisions. Because the
198 symbol is private to the module, all references can be updated. This
199 doesn't show up in any symbol table in the object file.
201 Similar to private, but the value shows as a local symbol
202 (``STB_LOCAL`` in the case of ELF) in the object file. This
203 corresponds to the notion of the '``static``' keyword in C.
204 ``available_externally``
205 Globals with "``available_externally``" linkage are never emitted
206 into the object file corresponding to the LLVM module. They exist to
207 allow inlining and other optimizations to take place given knowledge
208 of the definition of the global, which is known to be somewhere
209 outside the module. Globals with ``available_externally`` linkage
210 are allowed to be discarded at will, and are otherwise the same as
211 ``linkonce_odr``. This linkage type is only allowed on definitions,
214 Globals with "``linkonce``" linkage are merged with other globals of
215 the same name when linkage occurs. This can be used to implement
216 some forms of inline functions, templates, or other code which must
217 be generated in each translation unit that uses it, but where the
218 body may be overridden with a more definitive definition later.
219 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
220 that ``linkonce`` linkage does not actually allow the optimizer to
221 inline the body of this function into callers because it doesn't
222 know if this definition of the function is the definitive definition
223 within the program or whether it will be overridden by a stronger
224 definition. To enable inlining and other optimizations, use
225 "``linkonce_odr``" linkage.
227 "``weak``" linkage has the same merging semantics as ``linkonce``
228 linkage, except that unreferenced globals with ``weak`` linkage may
229 not be discarded. This is used for globals that are declared "weak"
232 "``common``" linkage is most similar to "``weak``" linkage, but they
233 are used for tentative definitions in C, such as "``int X;``" at
234 global scope. Symbols with "``common``" linkage are merged in the
235 same way as ``weak symbols``, and they may not be deleted if
236 unreferenced. ``common`` symbols may not have an explicit section,
237 must have a zero initializer, and may not be marked
238 ':ref:`constant <globalvars>`'. Functions and aliases may not have
241 .. _linkage_appending:
244 "``appending``" linkage may only be applied to global variables of
245 pointer to array type. When two global variables with appending
246 linkage are linked together, the two global arrays are appended
247 together. This is the LLVM, typesafe, equivalent of having the
248 system linker append together "sections" with identical names when
251 The semantics of this linkage follow the ELF object file model: the
252 symbol is weak until linked, if not linked, the symbol becomes null
253 instead of being an undefined reference.
254 ``linkonce_odr``, ``weak_odr``
255 Some languages allow differing globals to be merged, such as two
256 functions with different semantics. Other languages, such as
257 ``C++``, ensure that only equivalent globals are ever merged (the
258 "one definition rule" --- "ODR"). Such languages can use the
259 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
260 global will only be merged with equivalent globals. These linkage
261 types are otherwise the same as their non-``odr`` versions.
263 If none of the above identifiers are used, the global is externally
264 visible, meaning that it participates in linkage and can be used to
265 resolve external symbol references.
267 It is illegal for a function *declaration* to have any linkage type
268 other than ``external`` or ``extern_weak``.
275 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
276 :ref:`invokes <i_invoke>` can all have an optional calling convention
277 specified for the call. The calling convention of any pair of dynamic
278 caller/callee must match, or the behavior of the program is undefined.
279 The following calling conventions are supported by LLVM, and more may be
282 "``ccc``" - The C calling convention
283 This calling convention (the default if no other calling convention
284 is specified) matches the target C calling conventions. This calling
285 convention supports varargs function calls and tolerates some
286 mismatch in the declared prototype and implemented declaration of
287 the function (as does normal C).
288 "``fastcc``" - The fast calling convention
289 This calling convention attempts to make calls as fast as possible
290 (e.g. by passing things in registers). This calling convention
291 allows the target to use whatever tricks it wants to produce fast
292 code for the target, without having to conform to an externally
293 specified ABI (Application Binary Interface). `Tail calls can only
294 be optimized when this, the GHC or the HiPE convention is
295 used. <CodeGenerator.html#id80>`_ This calling convention does not
296 support varargs and requires the prototype of all callees to exactly
297 match the prototype of the function definition.
298 "``coldcc``" - The cold calling convention
299 This calling convention attempts to make code in the caller as
300 efficient as possible under the assumption that the call is not
301 commonly executed. As such, these calls often preserve all registers
302 so that the call does not break any live ranges in the caller side.
303 This calling convention does not support varargs and requires the
304 prototype of all callees to exactly match the prototype of the
305 function definition. Furthermore the inliner doesn't consider such function
307 "``cc 10``" - GHC convention
308 This calling convention has been implemented specifically for use by
309 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
310 It passes everything in registers, going to extremes to achieve this
311 by disabling callee save registers. This calling convention should
312 not be used lightly but only for specific situations such as an
313 alternative to the *register pinning* performance technique often
314 used when implementing functional programming languages. At the
315 moment only X86 supports this convention and it has the following
318 - On *X86-32* only supports up to 4 bit type parameters. No
319 floating point types are supported.
320 - On *X86-64* only supports up to 10 bit type parameters and 6
321 floating point parameters.
323 This calling convention supports `tail call
324 optimization <CodeGenerator.html#id80>`_ but requires both the
325 caller and callee are using it.
326 "``cc 11``" - The HiPE calling convention
327 This calling convention has been implemented specifically for use by
328 the `High-Performance Erlang
329 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
330 native code compiler of the `Ericsson's Open Source Erlang/OTP
331 system <http://www.erlang.org/download.shtml>`_. It uses more
332 registers for argument passing than the ordinary C calling
333 convention and defines no callee-saved registers. The calling
334 convention properly supports `tail call
335 optimization <CodeGenerator.html#id80>`_ but requires that both the
336 caller and the callee use it. It uses a *register pinning*
337 mechanism, similar to GHC's convention, for keeping frequently
338 accessed runtime components pinned to specific hardware registers.
339 At the moment only X86 supports this convention (both 32 and 64
341 "``webkit_jscc``" - WebKit's JavaScript calling convention
342 This calling convention has been implemented for `WebKit FTL JIT
343 <https://trac.webkit.org/wiki/FTLJIT>`_. It passes arguments on the
344 stack right to left (as cdecl does), and returns a value in the
345 platform's customary return register.
346 "``anyregcc``" - Dynamic calling convention for code patching
347 This is a special convention that supports patching an arbitrary code
348 sequence in place of a call site. This convention forces the call
349 arguments into registers but allows them to be dynamcially
350 allocated. This can currently only be used with calls to
351 llvm.experimental.patchpoint because only this intrinsic records
352 the location of its arguments in a side table. See :doc:`StackMaps`.
353 "``preserve_mostcc``" - The `PreserveMost` calling convention
354 This calling convention attempts to make the code in the caller as little
355 intrusive as possible. This calling convention behaves identical to the `C`
356 calling convention on how arguments and return values are passed, but it
357 uses a different set of caller/callee-saved registers. This alleviates the
358 burden of saving and recovering a large register set before and after the
359 call in the caller. If the arguments are passed in callee-saved registers,
360 then they will be preserved by the callee across the call. This doesn't
361 apply for values returned in callee-saved registers.
363 - On X86-64 the callee preserves all general purpose registers, except for
364 R11. R11 can be used as a scratch register. Floating-point registers
365 (XMMs/YMMs) are not preserved and need to be saved by the caller.
367 The idea behind this convention is to support calls to runtime functions
368 that have a hot path and a cold path. The hot path is usually a small piece
369 of code that doesn't many registers. The cold path might need to call out to
370 another function and therefore only needs to preserve the caller-saved
371 registers, which haven't already been saved by the caller. The
372 `PreserveMost` calling convention is very similar to the `cold` calling
373 convention in terms of caller/callee-saved registers, but they are used for
374 different types of function calls. `coldcc` is for function calls that are
375 rarely executed, whereas `preserve_mostcc` function calls are intended to be
376 on the hot path and definitely executed a lot. Furthermore `preserve_mostcc`
377 doesn't prevent the inliner from inlining the function call.
379 This calling convention will be used by a future version of the ObjectiveC
380 runtime and should therefore still be considered experimental at this time.
381 Although this convention was created to optimize certain runtime calls to
382 the ObjectiveC runtime, it is not limited to this runtime and might be used
383 by other runtimes in the future too. The current implementation only
384 supports X86-64, but the intention is to support more architectures in the
386 "``preserve_allcc``" - The `PreserveAll` calling convention
387 This calling convention attempts to make the code in the caller even less
388 intrusive than the `PreserveMost` calling convention. This calling
389 convention also behaves identical to the `C` calling convention on how
390 arguments and return values are passed, but it uses a different set of
391 caller/callee-saved registers. This removes the burden of saving and
392 recovering a large register set before and after the call in the caller. If
393 the arguments are passed in callee-saved registers, then they will be
394 preserved by the callee across the call. This doesn't apply for values
395 returned in callee-saved registers.
397 - On X86-64 the callee preserves all general purpose registers, except for
398 R11. R11 can be used as a scratch register. Furthermore it also preserves
399 all floating-point registers (XMMs/YMMs).
401 The idea behind this convention is to support calls to runtime functions
402 that don't need to call out to any other functions.
404 This calling convention, like the `PreserveMost` calling convention, will be
405 used by a future version of the ObjectiveC runtime and should be considered
406 experimental at this time.
407 "``cc <n>``" - Numbered convention
408 Any calling convention may be specified by number, allowing
409 target-specific calling conventions to be used. Target specific
410 calling conventions start at 64.
412 More calling conventions can be added/defined on an as-needed basis, to
413 support Pascal conventions or any other well-known target-independent
416 .. _visibilitystyles:
421 All Global Variables and Functions have one of the following visibility
424 "``default``" - Default style
425 On targets that use the ELF object file format, default visibility
426 means that the declaration is visible to other modules and, in
427 shared libraries, means that the declared entity may be overridden.
428 On Darwin, default visibility means that the declaration is visible
429 to other modules. Default visibility corresponds to "external
430 linkage" in the language.
431 "``hidden``" - Hidden style
432 Two declarations of an object with hidden visibility refer to the
433 same object if they are in the same shared object. Usually, hidden
434 visibility indicates that the symbol will not be placed into the
435 dynamic symbol table, so no other module (executable or shared
436 library) can reference it directly.
437 "``protected``" - Protected style
438 On ELF, protected visibility indicates that the symbol will be
439 placed in the dynamic symbol table, but that references within the
440 defining module will bind to the local symbol. That is, the symbol
441 cannot be overridden by another module.
443 A symbol with ``internal`` or ``private`` linkage must have ``default``
451 All Global Variables, Functions and Aliases can have one of the following
455 "``dllimport``" causes the compiler to reference a function or variable via
456 a global pointer to a pointer that is set up by the DLL exporting the
457 symbol. On Microsoft Windows targets, the pointer name is formed by
458 combining ``__imp_`` and the function or variable name.
460 "``dllexport``" causes the compiler to provide a global pointer to a pointer
461 in a DLL, so that it can be referenced with the ``dllimport`` attribute. On
462 Microsoft Windows targets, the pointer name is formed by combining
463 ``__imp_`` and the function or variable name. Since this storage class
464 exists for defining a dll interface, the compiler, assembler and linker know
465 it is externally referenced and must refrain from deleting the symbol.
469 Thread Local Storage Models
470 ---------------------------
472 A variable may be defined as ``thread_local``, which means that it will
473 not be shared by threads (each thread will have a separated copy of the
474 variable). Not all targets support thread-local variables. Optionally, a
475 TLS model may be specified:
478 For variables that are only used within the current shared library.
480 For variables in modules that will not be loaded dynamically.
482 For variables defined in the executable and only used within it.
484 If no explicit model is given, the "general dynamic" model is used.
486 The models correspond to the ELF TLS models; see `ELF Handling For
487 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
488 more information on under which circumstances the different models may
489 be used. The target may choose a different TLS model if the specified
490 model is not supported, or if a better choice of model can be made.
492 A model can also be specified in a alias, but then it only governs how
493 the alias is accessed. It will not have any effect in the aliasee.
500 LLVM IR allows you to specify both "identified" and "literal" :ref:`structure
501 types <t_struct>`. Literal types are uniqued structurally, but identified types
502 are never uniqued. An :ref:`opaque structural type <t_opaque>` can also be used
503 to forward declare a type which is not yet available.
505 An example of a identified structure specification is:
509 %mytype = type { %mytype*, i32 }
511 Prior to the LLVM 3.0 release, identified types were structurally uniqued. Only
512 literal types are uniqued in recent versions of LLVM.
519 Global variables define regions of memory allocated at compilation time
522 Global variables definitions must be initialized.
524 Global variables in other translation units can also be declared, in which
525 case they don't have an initializer.
527 Either global variable definitions or declarations may have an explicit section
528 to be placed in and may have an optional explicit alignment specified.
530 A variable may be defined as a global ``constant``, which indicates that
531 the contents of the variable will **never** be modified (enabling better
532 optimization, allowing the global data to be placed in the read-only
533 section of an executable, etc). Note that variables that need runtime
534 initialization cannot be marked ``constant`` as there is a store to the
537 LLVM explicitly allows *declarations* of global variables to be marked
538 constant, even if the final definition of the global is not. This
539 capability can be used to enable slightly better optimization of the
540 program, but requires the language definition to guarantee that
541 optimizations based on the 'constantness' are valid for the translation
542 units that do not include the definition.
544 As SSA values, global variables define pointer values that are in scope
545 (i.e. they dominate) all basic blocks in the program. Global variables
546 always define a pointer to their "content" type because they describe a
547 region of memory, and all memory objects in LLVM are accessed through
550 Global variables can be marked with ``unnamed_addr`` which indicates
551 that the address is not significant, only the content. Constants marked
552 like this can be merged with other constants if they have the same
553 initializer. Note that a constant with significant address *can* be
554 merged with a ``unnamed_addr`` constant, the result being a constant
555 whose address is significant.
557 A global variable may be declared to reside in a target-specific
558 numbered address space. For targets that support them, address spaces
559 may affect how optimizations are performed and/or what target
560 instructions are used to access the variable. The default address space
561 is zero. The address space qualifier must precede any other attributes.
563 LLVM allows an explicit section to be specified for globals. If the
564 target supports it, it will emit globals to the section specified.
565 Additionally, the global can placed in a comdat if the target has the necessary
568 By default, global initializers are optimized by assuming that global
569 variables defined within the module are not modified from their
570 initial values before the start of the global initializer. This is
571 true even for variables potentially accessible from outside the
572 module, including those with external linkage or appearing in
573 ``@llvm.used`` or dllexported variables. This assumption may be suppressed
574 by marking the variable with ``externally_initialized``.
576 An explicit alignment may be specified for a global, which must be a
577 power of 2. If not present, or if the alignment is set to zero, the
578 alignment of the global is set by the target to whatever it feels
579 convenient. If an explicit alignment is specified, the global is forced
580 to have exactly that alignment. Targets and optimizers are not allowed
581 to over-align the global if the global has an assigned section. In this
582 case, the extra alignment could be observable: for example, code could
583 assume that the globals are densely packed in their section and try to
584 iterate over them as an array, alignment padding would break this
585 iteration. The maximum alignment is ``1 << 29``.
587 Globals can also have a :ref:`DLL storage class <dllstorageclass>`.
589 Variables and aliasaes can have a
590 :ref:`Thread Local Storage Model <tls_model>`.
594 [@<GlobalVarName> =] [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal]
595 [unnamed_addr] [AddrSpace] [ExternallyInitialized]
596 <global | constant> <Type> [<InitializerConstant>]
597 [, section "name"] [, align <Alignment>]
599 For example, the following defines a global in a numbered address space
600 with an initializer, section, and alignment:
604 @G = addrspace(5) constant float 1.0, section "foo", align 4
606 The following example just declares a global variable
610 @G = external global i32
612 The following example defines a thread-local global with the
613 ``initialexec`` TLS model:
617 @G = thread_local(initialexec) global i32 0, align 4
619 .. _functionstructure:
624 LLVM function definitions consist of the "``define``" keyword, an
625 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
626 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
627 an optional :ref:`calling convention <callingconv>`,
628 an optional ``unnamed_addr`` attribute, a return type, an optional
629 :ref:`parameter attribute <paramattrs>` for the return type, a function
630 name, a (possibly empty) argument list (each with optional :ref:`parameter
631 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
632 an optional section, an optional alignment,
633 an optional :ref:`comdat <langref_comdats>`,
634 an optional :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`, an opening
635 curly brace, a list of basic blocks, and a closing curly brace.
637 LLVM function declarations consist of the "``declare``" keyword, an
638 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
639 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
640 an optional :ref:`calling convention <callingconv>`,
641 an optional ``unnamed_addr`` attribute, a return type, an optional
642 :ref:`parameter attribute <paramattrs>` for the return type, a function
643 name, a possibly empty list of arguments, an optional alignment, an optional
644 :ref:`garbage collector name <gc>` and an optional :ref:`prefix <prefixdata>`.
646 A function definition contains a list of basic blocks, forming the CFG (Control
647 Flow Graph) for the function. Each basic block may optionally start with a label
648 (giving the basic block a symbol table entry), contains a list of instructions,
649 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
650 function return). If an explicit label is not provided, a block is assigned an
651 implicit numbered label, using the next value from the same counter as used for
652 unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
653 entry block does not have an explicit label, it will be assigned label "%0",
654 then the first unnamed temporary in that block will be "%1", etc.
656 The first basic block in a function is special in two ways: it is
657 immediately executed on entrance to the function, and it is not allowed
658 to have predecessor basic blocks (i.e. there can not be any branches to
659 the entry block of a function). Because the block can have no
660 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
662 LLVM allows an explicit section to be specified for functions. If the
663 target supports it, it will emit functions to the section specified.
664 Additionally, the function can placed in a COMDAT.
666 An explicit alignment may be specified for a function. If not present,
667 or if the alignment is set to zero, the alignment of the function is set
668 by the target to whatever it feels convenient. If an explicit alignment
669 is specified, the function is forced to have at least that much
670 alignment. All alignments must be a power of 2.
672 If the ``unnamed_addr`` attribute is given, the address is know to not
673 be significant and two identical functions can be merged.
677 define [linkage] [visibility] [DLLStorageClass]
679 <ResultType> @<FunctionName> ([argument list])
680 [unnamed_addr] [fn Attrs] [section "name"] [comdat $<ComdatName>]
681 [align N] [gc] [prefix Constant] { ... }
688 Aliases, unlike function or variables, don't create any new data. They
689 are just a new symbol and metadata for an existing position.
691 Aliases have a name and an aliasee that is either a global value or a
694 Aliases may have an optional :ref:`linkage type <linkage>`, an optional
695 :ref:`visibility style <visibility>`, an optional :ref:`DLL storage class
696 <dllstorageclass>` and an optional :ref:`tls model <tls_model>`.
700 @<Name> = [Visibility] [DLLStorageClass] [ThreadLocal] [unnamed_addr] alias [Linkage] <AliaseeTy> @<Aliasee>
702 The linkage must be one of ``private``, ``internal``, ``linkonce``, ``weak``,
703 ``linkonce_odr``, ``weak_odr``, ``external``. Note that some system linkers
704 might not correctly handle dropping a weak symbol that is aliased.
706 Alias that are not ``unnamed_addr`` are guaranteed to have the same address as
707 the aliasee expression. ``unnamed_addr`` ones are only guaranteed to point
710 Since aliases are only a second name, some restrictions apply, of which
711 some can only be checked when producing an object file:
713 * The expression defining the aliasee must be computable at assembly
714 time. Since it is just a name, no relocations can be used.
716 * No alias in the expression can be weak as the possibility of the
717 intermediate alias being overridden cannot be represented in an
720 * No global value in the expression can be a declaration, since that
721 would require a relocation, which is not possible.
728 Comdat IR provides access to COFF and ELF object file COMDAT functionality.
730 Comdats have a name which represents the COMDAT key. All global objects which
731 specify this key will only end up in the final object file if the linker chooses
732 that key over some other key. Aliases are placed in the same COMDAT that their
733 aliasee computes to, if any.
735 Comdats have a selection kind to provide input on how the linker should
736 choose between keys in two different object files.
740 $<Name> = comdat SelectionKind
742 The selection kind must be one of the following:
745 The linker may choose any COMDAT key, the choice is arbitrary.
747 The linker may choose any COMDAT key but the sections must contain the
750 The linker will choose the section containing the largest COMDAT key.
752 The linker requires that only section with this COMDAT key exist.
754 The linker may choose any COMDAT key but the sections must contain the
757 Note that the Mach-O platform doesn't support COMDATs and ELF only supports
758 ``any`` as a selection kind.
760 Here is an example of a COMDAT group where a function will only be selected if
761 the COMDAT key's section is the largest:
765 $foo = comdat largest
766 @foo = global i32 2, comdat $foo
768 define void @bar() comdat $foo {
772 In a COFF object file, this will create a COMDAT section with selection kind
773 ``IMAGE_COMDAT_SELECT_LARGEST`` containing the contents of the ``@foo`` symbol
774 and another COMDAT section with selection kind
775 ``IMAGE_COMDAT_SELECT_ASSOCIATIVE`` which is associated with the first COMDAT
776 section and contains the contents of the ``@baz`` symbol.
778 There are some restrictions on the properties of the global object.
779 It, or an alias to it, must have the same name as the COMDAT group when
781 The contents and size of this object may be used during link-time to determine
782 which COMDAT groups get selected depending on the selection kind.
783 Because the name of the object must match the name of the COMDAT group, the
784 linkage of the global object must not be local; local symbols can get renamed
785 if a collision occurs in the symbol table.
787 The combined use of COMDATS and section attributes may yield surprising results.
794 @g1 = global i32 42, section "sec", comdat $foo
795 @g2 = global i32 42, section "sec", comdat $bar
797 From the object file perspective, this requires the creation of two sections
798 with the same name. This is necessary because both globals belong to different
799 COMDAT groups and COMDATs, at the object file level, are represented by
802 Note that certain IR constructs like global variables and functions may create
803 COMDATs in the object file in addition to any which are specified using COMDAT
804 IR. This arises, for example, when a global variable has linkonce_odr linkage.
806 .. _namedmetadatastructure:
811 Named metadata is a collection of metadata. :ref:`Metadata
812 nodes <metadata>` (but not metadata strings) are the only valid
813 operands for a named metadata.
817 ; Some unnamed metadata nodes, which are referenced by the named metadata.
818 !0 = metadata !{metadata !"zero"}
819 !1 = metadata !{metadata !"one"}
820 !2 = metadata !{metadata !"two"}
822 !name = !{!0, !1, !2}
829 The return type and each parameter of a function type may have a set of
830 *parameter attributes* associated with them. Parameter attributes are
831 used to communicate additional information about the result or
832 parameters of a function. Parameter attributes are considered to be part
833 of the function, not of the function type, so functions with different
834 parameter attributes can have the same function type.
836 Parameter attributes are simple keywords that follow the type specified.
837 If multiple parameter attributes are needed, they are space separated.
842 declare i32 @printf(i8* noalias nocapture, ...)
843 declare i32 @atoi(i8 zeroext)
844 declare signext i8 @returns_signed_char()
846 Note that any attributes for the function result (``nounwind``,
847 ``readonly``) come immediately after the argument list.
849 Currently, only the following parameter attributes are defined:
852 This indicates to the code generator that the parameter or return
853 value should be zero-extended to the extent required by the target's
854 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
855 the caller (for a parameter) or the callee (for a return value).
857 This indicates to the code generator that the parameter or return
858 value should be sign-extended to the extent required by the target's
859 ABI (which is usually 32-bits) by the caller (for a parameter) or
860 the callee (for a return value).
862 This indicates that this parameter or return value should be treated
863 in a special target-dependent fashion during while emitting code for
864 a function call or return (usually, by putting it in a register as
865 opposed to memory, though some targets use it to distinguish between
866 two different kinds of registers). Use of this attribute is
869 This indicates that the pointer parameter should really be passed by
870 value to the function. The attribute implies that a hidden copy of
871 the pointee is made between the caller and the callee, so the callee
872 is unable to modify the value in the caller. This attribute is only
873 valid on LLVM pointer arguments. It is generally used to pass
874 structs and arrays by value, but is also valid on pointers to
875 scalars. The copy is considered to belong to the caller not the
876 callee (for example, ``readonly`` functions should not write to
877 ``byval`` parameters). This is not a valid attribute for return
880 The byval attribute also supports specifying an alignment with the
881 align attribute. It indicates the alignment of the stack slot to
882 form and the known alignment of the pointer specified to the call
883 site. If the alignment is not specified, then the code generator
884 makes a target-specific assumption.
890 The ``inalloca`` argument attribute allows the caller to take the
891 address of outgoing stack arguments. An ``inalloca`` argument must
892 be a pointer to stack memory produced by an ``alloca`` instruction.
893 The alloca, or argument allocation, must also be tagged with the
894 inalloca keyword. Only the last argument may have the ``inalloca``
895 attribute, and that argument is guaranteed to be passed in memory.
897 An argument allocation may be used by a call at most once because
898 the call may deallocate it. The ``inalloca`` attribute cannot be
899 used in conjunction with other attributes that affect argument
900 storage, like ``inreg``, ``nest``, ``sret``, or ``byval``. The
901 ``inalloca`` attribute also disables LLVM's implicit lowering of
902 large aggregate return values, which means that frontend authors
903 must lower them with ``sret`` pointers.
905 When the call site is reached, the argument allocation must have
906 been the most recent stack allocation that is still live, or the
907 results are undefined. It is possible to allocate additional stack
908 space after an argument allocation and before its call site, but it
909 must be cleared off with :ref:`llvm.stackrestore
912 See :doc:`InAlloca` for more information on how to use this
916 This indicates that the pointer parameter specifies the address of a
917 structure that is the return value of the function in the source
918 program. This pointer must be guaranteed by the caller to be valid:
919 loads and stores to the structure may be assumed by the callee
920 not to trap and to be properly aligned. This may only be applied to
921 the first parameter. This is not a valid attribute for return
927 This indicates that pointer values :ref:`based <pointeraliasing>` on
928 the argument or return value do not alias pointer values which are
929 not *based* on it, ignoring certain "irrelevant" dependencies. For a
930 call to the parent function, dependencies between memory references
931 from before or after the call and from those during the call are
932 "irrelevant" to the ``noalias`` keyword for the arguments and return
933 value used in that call. The caller shares the responsibility with
934 the callee for ensuring that these requirements are met. For further
935 details, please see the discussion of the NoAlias response in :ref:`alias
936 analysis <Must, May, or No>`.
938 Note that this definition of ``noalias`` is intentionally similar
939 to the definition of ``restrict`` in C99 for function arguments,
940 though it is slightly weaker.
942 For function return values, C99's ``restrict`` is not meaningful,
943 while LLVM's ``noalias`` is.
945 This indicates that the callee does not make any copies of the
946 pointer that outlive the callee itself. This is not a valid
947 attribute for return values.
952 This indicates that the pointer parameter can be excised using the
953 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
954 attribute for return values and can only be applied to one parameter.
957 This indicates that the function always returns the argument as its return
958 value. This is an optimization hint to the code generator when generating
959 the caller, allowing tail call optimization and omission of register saves
960 and restores in some cases; it is not checked or enforced when generating
961 the callee. The parameter and the function return type must be valid
962 operands for the :ref:`bitcast instruction <i_bitcast>`. This is not a
963 valid attribute for return values and can only be applied to one parameter.
966 This indicates that the parameter or return pointer is not null. This
967 attribute may only be applied to pointer typed parameters. This is not
968 checked or enforced by LLVM, the caller must ensure that the pointer
969 passed in is non-null, or the callee must ensure that the returned pointer
972 ``dereferenceable(<n>)``
973 This indicates that the parameter or return pointer is dereferenceable. This
974 attribute may only be applied to pointer typed parameters. A pointer that
975 is dereferenceable can be loaded from speculatively without a risk of
976 trapping. The number of bytes known to be dereferenceable must be provided
977 in parentheses. It is legal for the number of bytes to be less than the
978 size of the pointee type. The ``nonnull`` attribute does not imply
979 dereferenceability (consider a pointer to one element past the end of an
980 array), however ``dereferenceable(<n>)`` does imply ``nonnull`` in
981 ``addrspace(0)`` (which is the default address space).
985 Garbage Collector Names
986 -----------------------
988 Each function may specify a garbage collector name, which is simply a
993 define void @f() gc "name" { ... }
995 The compiler declares the supported values of *name*. Specifying a
996 collector which will cause the compiler to alter its output in order to
997 support the named garbage collection algorithm.
1004 Prefix data is data associated with a function which the code generator
1005 will emit immediately before the function body. The purpose of this feature
1006 is to allow frontends to associate language-specific runtime metadata with
1007 specific functions and make it available through the function pointer while
1008 still allowing the function pointer to be called. To access the data for a
1009 given function, a program may bitcast the function pointer to a pointer to
1010 the constant's type. This implies that the IR symbol points to the start
1013 To maintain the semantics of ordinary function calls, the prefix data must
1014 have a particular format. Specifically, it must begin with a sequence of
1015 bytes which decode to a sequence of machine instructions, valid for the
1016 module's target, which transfer control to the point immediately succeeding
1017 the prefix data, without performing any other visible action. This allows
1018 the inliner and other passes to reason about the semantics of the function
1019 definition without needing to reason about the prefix data. Obviously this
1020 makes the format of the prefix data highly target dependent.
1022 Prefix data is laid out as if it were an initializer for a global variable
1023 of the prefix data's type. No padding is automatically placed between the
1024 prefix data and the function body. If padding is required, it must be part
1027 A trivial example of valid prefix data for the x86 architecture is ``i8 144``,
1028 which encodes the ``nop`` instruction:
1030 .. code-block:: llvm
1032 define void @f() prefix i8 144 { ... }
1034 Generally prefix data can be formed by encoding a relative branch instruction
1035 which skips the metadata, as in this example of valid prefix data for the
1036 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
1038 .. code-block:: llvm
1040 %0 = type <{ i8, i8, i8* }>
1042 define void @f() prefix %0 <{ i8 235, i8 8, i8* @md}> { ... }
1044 A function may have prefix data but no body. This has similar semantics
1045 to the ``available_externally`` linkage in that the data may be used by the
1046 optimizers but will not be emitted in the object file.
1053 Attribute groups are groups of attributes that are referenced by objects within
1054 the IR. They are important for keeping ``.ll`` files readable, because a lot of
1055 functions will use the same set of attributes. In the degenerative case of a
1056 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
1057 group will capture the important command line flags used to build that file.
1059 An attribute group is a module-level object. To use an attribute group, an
1060 object references the attribute group's ID (e.g. ``#37``). An object may refer
1061 to more than one attribute group. In that situation, the attributes from the
1062 different groups are merged.
1064 Here is an example of attribute groups for a function that should always be
1065 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
1067 .. code-block:: llvm
1069 ; Target-independent attributes:
1070 attributes #0 = { alwaysinline alignstack=4 }
1072 ; Target-dependent attributes:
1073 attributes #1 = { "no-sse" }
1075 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
1076 define void @f() #0 #1 { ... }
1083 Function attributes are set to communicate additional information about
1084 a function. Function attributes are considered to be part of the
1085 function, not of the function type, so functions with different function
1086 attributes can have the same function type.
1088 Function attributes are simple keywords that follow the type specified.
1089 If multiple attributes are needed, they are space separated. For
1092 .. code-block:: llvm
1094 define void @f() noinline { ... }
1095 define void @f() alwaysinline { ... }
1096 define void @f() alwaysinline optsize { ... }
1097 define void @f() optsize { ... }
1100 This attribute indicates that, when emitting the prologue and
1101 epilogue, the backend should forcibly align the stack pointer.
1102 Specify the desired alignment, which must be a power of two, in
1105 This attribute indicates that the inliner should attempt to inline
1106 this function into callers whenever possible, ignoring any active
1107 inlining size threshold for this caller.
1109 This indicates that the callee function at a call site should be
1110 recognized as a built-in function, even though the function's declaration
1111 uses the ``nobuiltin`` attribute. This is only valid at call sites for
1112 direct calls to functions which are declared with the ``nobuiltin``
1115 This attribute indicates that this function is rarely called. When
1116 computing edge weights, basic blocks post-dominated by a cold
1117 function call are also considered to be cold; and, thus, given low
1120 This attribute indicates that the source code contained a hint that
1121 inlining this function is desirable (such as the "inline" keyword in
1122 C/C++). It is just a hint; it imposes no requirements on the
1125 This attribute indicates that the function should be added to a
1126 jump-instruction table at code-generation time, and that all address-taken
1127 references to this function should be replaced with a reference to the
1128 appropriate jump-instruction-table function pointer. Note that this creates
1129 a new pointer for the original function, which means that code that depends
1130 on function-pointer identity can break. So, any function annotated with
1131 ``jumptable`` must also be ``unnamed_addr``.
1133 This attribute suggests that optimization passes and code generator
1134 passes make choices that keep the code size of this function as small
1135 as possible and perform optimizations that may sacrifice runtime
1136 performance in order to minimize the size of the generated code.
1138 This attribute disables prologue / epilogue emission for the
1139 function. This can have very system-specific consequences.
1141 This indicates that the callee function at a call site is not recognized as
1142 a built-in function. LLVM will retain the original call and not replace it
1143 with equivalent code based on the semantics of the built-in function, unless
1144 the call site uses the ``builtin`` attribute. This is valid at call sites
1145 and on function declarations and definitions.
1147 This attribute indicates that calls to the function cannot be
1148 duplicated. A call to a ``noduplicate`` function may be moved
1149 within its parent function, but may not be duplicated within
1150 its parent function.
1152 A function containing a ``noduplicate`` call may still
1153 be an inlining candidate, provided that the call is not
1154 duplicated by inlining. That implies that the function has
1155 internal linkage and only has one call site, so the original
1156 call is dead after inlining.
1158 This attributes disables implicit floating point instructions.
1160 This attribute indicates that the inliner should never inline this
1161 function in any situation. This attribute may not be used together
1162 with the ``alwaysinline`` attribute.
1164 This attribute suppresses lazy symbol binding for the function. This
1165 may make calls to the function faster, at the cost of extra program
1166 startup time if the function is not called during program startup.
1168 This attribute indicates that the code generator should not use a
1169 red zone, even if the target-specific ABI normally permits it.
1171 This function attribute indicates that the function never returns
1172 normally. This produces undefined behavior at runtime if the
1173 function ever does dynamically return.
1175 This function attribute indicates that the function never returns
1176 with an unwind or exceptional control flow. If the function does
1177 unwind, its runtime behavior is undefined.
1179 This function attribute indicates that the function is not optimized
1180 by any optimization or code generator passes with the
1181 exception of interprocedural optimization passes.
1182 This attribute cannot be used together with the ``alwaysinline``
1183 attribute; this attribute is also incompatible
1184 with the ``minsize`` attribute and the ``optsize`` attribute.
1186 This attribute requires the ``noinline`` attribute to be specified on
1187 the function as well, so the function is never inlined into any caller.
1188 Only functions with the ``alwaysinline`` attribute are valid
1189 candidates for inlining into the body of this function.
1191 This attribute suggests that optimization passes and code generator
1192 passes make choices that keep the code size of this function low,
1193 and otherwise do optimizations specifically to reduce code size as
1194 long as they do not significantly impact runtime performance.
1196 On a function, this attribute indicates that the function computes its
1197 result (or decides to unwind an exception) based strictly on its arguments,
1198 without dereferencing any pointer arguments or otherwise accessing
1199 any mutable state (e.g. memory, control registers, etc) visible to
1200 caller functions. It does not write through any pointer arguments
1201 (including ``byval`` arguments) and never changes any state visible
1202 to callers. This means that it cannot unwind exceptions by calling
1203 the ``C++`` exception throwing methods.
1205 On an argument, this attribute indicates that the function does not
1206 dereference that pointer argument, even though it may read or write the
1207 memory that the pointer points to if accessed through other pointers.
1209 On a function, this attribute indicates that the function does not write
1210 through any pointer arguments (including ``byval`` arguments) or otherwise
1211 modify any state (e.g. memory, control registers, etc) visible to
1212 caller functions. It may dereference pointer arguments and read
1213 state that may be set in the caller. A readonly function always
1214 returns the same value (or unwinds an exception identically) when
1215 called with the same set of arguments and global state. It cannot
1216 unwind an exception by calling the ``C++`` exception throwing
1219 On an argument, this attribute indicates that the function does not write
1220 through this pointer argument, even though it may write to the memory that
1221 the pointer points to.
1223 This attribute indicates that this function can return twice. The C
1224 ``setjmp`` is an example of such a function. The compiler disables
1225 some optimizations (like tail calls) in the caller of these
1227 ``sanitize_address``
1228 This attribute indicates that AddressSanitizer checks
1229 (dynamic address safety analysis) are enabled for this function.
1231 This attribute indicates that MemorySanitizer checks (dynamic detection
1232 of accesses to uninitialized memory) are enabled for this function.
1234 This attribute indicates that ThreadSanitizer checks
1235 (dynamic thread safety analysis) are enabled for this function.
1237 This attribute indicates that the function should emit a stack
1238 smashing protector. It is in the form of a "canary" --- a random value
1239 placed on the stack before the local variables that's checked upon
1240 return from the function to see if it has been overwritten. A
1241 heuristic is used to determine if a function needs stack protectors
1242 or not. The heuristic used will enable protectors for functions with:
1244 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1245 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1246 - Calls to alloca() with variable sizes or constant sizes greater than
1247 ``ssp-buffer-size``.
1249 Variables that are identified as requiring a protector will be arranged
1250 on the stack such that they are adjacent to the stack protector guard.
1252 If a function that has an ``ssp`` attribute is inlined into a
1253 function that doesn't have an ``ssp`` attribute, then the resulting
1254 function will have an ``ssp`` attribute.
1256 This attribute indicates that the function should *always* emit a
1257 stack smashing protector. This overrides the ``ssp`` function
1260 Variables that are identified as requiring a protector will be arranged
1261 on the stack such that they are adjacent to the stack protector guard.
1262 The specific layout rules are:
1264 #. Large arrays and structures containing large arrays
1265 (``>= ssp-buffer-size``) are closest to the stack protector.
1266 #. Small arrays and structures containing small arrays
1267 (``< ssp-buffer-size``) are 2nd closest to the protector.
1268 #. Variables that have had their address taken are 3rd closest to the
1271 If a function that has an ``sspreq`` attribute is inlined into a
1272 function that doesn't have an ``sspreq`` attribute or which has an
1273 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1274 an ``sspreq`` attribute.
1276 This attribute indicates that the function should emit a stack smashing
1277 protector. This attribute causes a strong heuristic to be used when
1278 determining if a function needs stack protectors. The strong heuristic
1279 will enable protectors for functions with:
1281 - Arrays of any size and type
1282 - Aggregates containing an array of any size and type.
1283 - Calls to alloca().
1284 - Local variables that have had their address taken.
1286 Variables that are identified as requiring a protector will be arranged
1287 on the stack such that they are adjacent to the stack protector guard.
1288 The specific layout rules are:
1290 #. Large arrays and structures containing large arrays
1291 (``>= ssp-buffer-size``) are closest to the stack protector.
1292 #. Small arrays and structures containing small arrays
1293 (``< ssp-buffer-size``) are 2nd closest to the protector.
1294 #. Variables that have had their address taken are 3rd closest to the
1297 This overrides the ``ssp`` function attribute.
1299 If a function that has an ``sspstrong`` attribute is inlined into a
1300 function that doesn't have an ``sspstrong`` attribute, then the
1301 resulting function will have an ``sspstrong`` attribute.
1303 This attribute indicates that the ABI being targeted requires that
1304 an unwind table entry be produce for this function even if we can
1305 show that no exceptions passes by it. This is normally the case for
1306 the ELF x86-64 abi, but it can be disabled for some compilation
1311 Module-Level Inline Assembly
1312 ----------------------------
1314 Modules may contain "module-level inline asm" blocks, which corresponds
1315 to the GCC "file scope inline asm" blocks. These blocks are internally
1316 concatenated by LLVM and treated as a single unit, but may be separated
1317 in the ``.ll`` file if desired. The syntax is very simple:
1319 .. code-block:: llvm
1321 module asm "inline asm code goes here"
1322 module asm "more can go here"
1324 The strings can contain any character by escaping non-printable
1325 characters. The escape sequence used is simply "\\xx" where "xx" is the
1326 two digit hex code for the number.
1328 The inline asm code is simply printed to the machine code .s file when
1329 assembly code is generated.
1331 .. _langref_datalayout:
1336 A module may specify a target specific data layout string that specifies
1337 how data is to be laid out in memory. The syntax for the data layout is
1340 .. code-block:: llvm
1342 target datalayout = "layout specification"
1344 The *layout specification* consists of a list of specifications
1345 separated by the minus sign character ('-'). Each specification starts
1346 with a letter and may include other information after the letter to
1347 define some aspect of the data layout. The specifications accepted are
1351 Specifies that the target lays out data in big-endian form. That is,
1352 the bits with the most significance have the lowest address
1355 Specifies that the target lays out data in little-endian form. That
1356 is, the bits with the least significance have the lowest address
1359 Specifies the natural alignment of the stack in bits. Alignment
1360 promotion of stack variables is limited to the natural stack
1361 alignment to avoid dynamic stack realignment. The stack alignment
1362 must be a multiple of 8-bits. If omitted, the natural stack
1363 alignment defaults to "unspecified", which does not prevent any
1364 alignment promotions.
1365 ``p[n]:<size>:<abi>:<pref>``
1366 This specifies the *size* of a pointer and its ``<abi>`` and
1367 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1368 bits. The address space, ``n`` is optional, and if not specified,
1369 denotes the default address space 0. The value of ``n`` must be
1370 in the range [1,2^23).
1371 ``i<size>:<abi>:<pref>``
1372 This specifies the alignment for an integer type of a given bit
1373 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1374 ``v<size>:<abi>:<pref>``
1375 This specifies the alignment for a vector type of a given bit
1377 ``f<size>:<abi>:<pref>``
1378 This specifies the alignment for a floating point type of a given bit
1379 ``<size>``. Only values of ``<size>`` that are supported by the target
1380 will work. 32 (float) and 64 (double) are supported on all targets; 80
1381 or 128 (different flavors of long double) are also supported on some
1384 This specifies the alignment for an object of aggregate type.
1386 If present, specifies that llvm names are mangled in the output. The
1389 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1390 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1391 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1392 symbols get a ``_`` prefix.
1393 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1394 functions also get a suffix based on the frame size.
1395 ``n<size1>:<size2>:<size3>...``
1396 This specifies a set of native integer widths for the target CPU in
1397 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1398 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1399 this set are considered to support most general arithmetic operations
1402 On every specification that takes a ``<abi>:<pref>``, specifying the
1403 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1404 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1406 When constructing the data layout for a given target, LLVM starts with a
1407 default set of specifications which are then (possibly) overridden by
1408 the specifications in the ``datalayout`` keyword. The default
1409 specifications are given in this list:
1411 - ``E`` - big endian
1412 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1413 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1414 same as the default address space.
1415 - ``S0`` - natural stack alignment is unspecified
1416 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1417 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1418 - ``i16:16:16`` - i16 is 16-bit aligned
1419 - ``i32:32:32`` - i32 is 32-bit aligned
1420 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1421 alignment of 64-bits
1422 - ``f16:16:16`` - half is 16-bit aligned
1423 - ``f32:32:32`` - float is 32-bit aligned
1424 - ``f64:64:64`` - double is 64-bit aligned
1425 - ``f128:128:128`` - quad is 128-bit aligned
1426 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1427 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1428 - ``a:0:64`` - aggregates are 64-bit aligned
1430 When LLVM is determining the alignment for a given type, it uses the
1433 #. If the type sought is an exact match for one of the specifications,
1434 that specification is used.
1435 #. If no match is found, and the type sought is an integer type, then
1436 the smallest integer type that is larger than the bitwidth of the
1437 sought type is used. If none of the specifications are larger than
1438 the bitwidth then the largest integer type is used. For example,
1439 given the default specifications above, the i7 type will use the
1440 alignment of i8 (next largest) while both i65 and i256 will use the
1441 alignment of i64 (largest specified).
1442 #. If no match is found, and the type sought is a vector type, then the
1443 largest vector type that is smaller than the sought vector type will
1444 be used as a fall back. This happens because <128 x double> can be
1445 implemented in terms of 64 <2 x double>, for example.
1447 The function of the data layout string may not be what you expect.
1448 Notably, this is not a specification from the frontend of what alignment
1449 the code generator should use.
1451 Instead, if specified, the target data layout is required to match what
1452 the ultimate *code generator* expects. This string is used by the
1453 mid-level optimizers to improve code, and this only works if it matches
1454 what the ultimate code generator uses. If you would like to generate IR
1455 that does not embed this target-specific detail into the IR, then you
1456 don't have to specify the string. This will disable some optimizations
1457 that require precise layout information, but this also prevents those
1458 optimizations from introducing target specificity into the IR.
1465 A module may specify a target triple string that describes the target
1466 host. The syntax for the target triple is simply:
1468 .. code-block:: llvm
1470 target triple = "x86_64-apple-macosx10.7.0"
1472 The *target triple* string consists of a series of identifiers delimited
1473 by the minus sign character ('-'). The canonical forms are:
1477 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1478 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1480 This information is passed along to the backend so that it generates
1481 code for the proper architecture. It's possible to override this on the
1482 command line with the ``-mtriple`` command line option.
1484 .. _pointeraliasing:
1486 Pointer Aliasing Rules
1487 ----------------------
1489 Any memory access must be done through a pointer value associated with
1490 an address range of the memory access, otherwise the behavior is
1491 undefined. Pointer values are associated with address ranges according
1492 to the following rules:
1494 - A pointer value is associated with the addresses associated with any
1495 value it is *based* on.
1496 - An address of a global variable is associated with the address range
1497 of the variable's storage.
1498 - The result value of an allocation instruction is associated with the
1499 address range of the allocated storage.
1500 - A null pointer in the default address-space is associated with no
1502 - An integer constant other than zero or a pointer value returned from
1503 a function not defined within LLVM may be associated with address
1504 ranges allocated through mechanisms other than those provided by
1505 LLVM. Such ranges shall not overlap with any ranges of addresses
1506 allocated by mechanisms provided by LLVM.
1508 A pointer value is *based* on another pointer value according to the
1511 - A pointer value formed from a ``getelementptr`` operation is *based*
1512 on the first operand of the ``getelementptr``.
1513 - The result value of a ``bitcast`` is *based* on the operand of the
1515 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1516 values that contribute (directly or indirectly) to the computation of
1517 the pointer's value.
1518 - The "*based* on" relationship is transitive.
1520 Note that this definition of *"based"* is intentionally similar to the
1521 definition of *"based"* in C99, though it is slightly weaker.
1523 LLVM IR does not associate types with memory. The result type of a
1524 ``load`` merely indicates the size and alignment of the memory from
1525 which to load, as well as the interpretation of the value. The first
1526 operand type of a ``store`` similarly only indicates the size and
1527 alignment of the store.
1529 Consequently, type-based alias analysis, aka TBAA, aka
1530 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1531 :ref:`Metadata <metadata>` may be used to encode additional information
1532 which specialized optimization passes may use to implement type-based
1537 Volatile Memory Accesses
1538 ------------------------
1540 Certain memory accesses, such as :ref:`load <i_load>`'s,
1541 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1542 marked ``volatile``. The optimizers must not change the number of
1543 volatile operations or change their order of execution relative to other
1544 volatile operations. The optimizers *may* change the order of volatile
1545 operations relative to non-volatile operations. This is not Java's
1546 "volatile" and has no cross-thread synchronization behavior.
1548 IR-level volatile loads and stores cannot safely be optimized into
1549 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1550 flagged volatile. Likewise, the backend should never split or merge
1551 target-legal volatile load/store instructions.
1553 .. admonition:: Rationale
1555 Platforms may rely on volatile loads and stores of natively supported
1556 data width to be executed as single instruction. For example, in C
1557 this holds for an l-value of volatile primitive type with native
1558 hardware support, but not necessarily for aggregate types. The
1559 frontend upholds these expectations, which are intentionally
1560 unspecified in the IR. The rules above ensure that IR transformation
1561 do not violate the frontend's contract with the language.
1565 Memory Model for Concurrent Operations
1566 --------------------------------------
1568 The LLVM IR does not define any way to start parallel threads of
1569 execution or to register signal handlers. Nonetheless, there are
1570 platform-specific ways to create them, and we define LLVM IR's behavior
1571 in their presence. This model is inspired by the C++0x memory model.
1573 For a more informal introduction to this model, see the :doc:`Atomics`.
1575 We define a *happens-before* partial order as the least partial order
1578 - Is a superset of single-thread program order, and
1579 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1580 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1581 techniques, like pthread locks, thread creation, thread joining,
1582 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1583 Constraints <ordering>`).
1585 Note that program order does not introduce *happens-before* edges
1586 between a thread and signals executing inside that thread.
1588 Every (defined) read operation (load instructions, memcpy, atomic
1589 loads/read-modify-writes, etc.) R reads a series of bytes written by
1590 (defined) write operations (store instructions, atomic
1591 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1592 section, initialized globals are considered to have a write of the
1593 initializer which is atomic and happens before any other read or write
1594 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1595 may see any write to the same byte, except:
1597 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1598 write\ :sub:`2` happens before R\ :sub:`byte`, then
1599 R\ :sub:`byte` does not see write\ :sub:`1`.
1600 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1601 R\ :sub:`byte` does not see write\ :sub:`3`.
1603 Given that definition, R\ :sub:`byte` is defined as follows:
1605 - If R is volatile, the result is target-dependent. (Volatile is
1606 supposed to give guarantees which can support ``sig_atomic_t`` in
1607 C/C++, and may be used for accesses to addresses which do not behave
1608 like normal memory. It does not generally provide cross-thread
1610 - Otherwise, if there is no write to the same byte that happens before
1611 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1612 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1613 R\ :sub:`byte` returns the value written by that write.
1614 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1615 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1616 Memory Ordering Constraints <ordering>` section for additional
1617 constraints on how the choice is made.
1618 - Otherwise R\ :sub:`byte` returns ``undef``.
1620 R returns the value composed of the series of bytes it read. This
1621 implies that some bytes within the value may be ``undef`` **without**
1622 the entire value being ``undef``. Note that this only defines the
1623 semantics of the operation; it doesn't mean that targets will emit more
1624 than one instruction to read the series of bytes.
1626 Note that in cases where none of the atomic intrinsics are used, this
1627 model places only one restriction on IR transformations on top of what
1628 is required for single-threaded execution: introducing a store to a byte
1629 which might not otherwise be stored is not allowed in general.
1630 (Specifically, in the case where another thread might write to and read
1631 from an address, introducing a store can change a load that may see
1632 exactly one write into a load that may see multiple writes.)
1636 Atomic Memory Ordering Constraints
1637 ----------------------------------
1639 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1640 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1641 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1642 ordering parameters that determine which other atomic instructions on
1643 the same address they *synchronize with*. These semantics are borrowed
1644 from Java and C++0x, but are somewhat more colloquial. If these
1645 descriptions aren't precise enough, check those specs (see spec
1646 references in the :doc:`atomics guide <Atomics>`).
1647 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1648 differently since they don't take an address. See that instruction's
1649 documentation for details.
1651 For a simpler introduction to the ordering constraints, see the
1655 The set of values that can be read is governed by the happens-before
1656 partial order. A value cannot be read unless some operation wrote
1657 it. This is intended to provide a guarantee strong enough to model
1658 Java's non-volatile shared variables. This ordering cannot be
1659 specified for read-modify-write operations; it is not strong enough
1660 to make them atomic in any interesting way.
1662 In addition to the guarantees of ``unordered``, there is a single
1663 total order for modifications by ``monotonic`` operations on each
1664 address. All modification orders must be compatible with the
1665 happens-before order. There is no guarantee that the modification
1666 orders can be combined to a global total order for the whole program
1667 (and this often will not be possible). The read in an atomic
1668 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1669 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1670 order immediately before the value it writes. If one atomic read
1671 happens before another atomic read of the same address, the later
1672 read must see the same value or a later value in the address's
1673 modification order. This disallows reordering of ``monotonic`` (or
1674 stronger) operations on the same address. If an address is written
1675 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1676 read that address repeatedly, the other threads must eventually see
1677 the write. This corresponds to the C++0x/C1x
1678 ``memory_order_relaxed``.
1680 In addition to the guarantees of ``monotonic``, a
1681 *synchronizes-with* edge may be formed with a ``release`` operation.
1682 This is intended to model C++'s ``memory_order_acquire``.
1684 In addition to the guarantees of ``monotonic``, if this operation
1685 writes a value which is subsequently read by an ``acquire``
1686 operation, it *synchronizes-with* that operation. (This isn't a
1687 complete description; see the C++0x definition of a release
1688 sequence.) This corresponds to the C++0x/C1x
1689 ``memory_order_release``.
1690 ``acq_rel`` (acquire+release)
1691 Acts as both an ``acquire`` and ``release`` operation on its
1692 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1693 ``seq_cst`` (sequentially consistent)
1694 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1695 operation which only reads, ``release`` for an operation which only
1696 writes), there is a global total order on all
1697 sequentially-consistent operations on all addresses, which is
1698 consistent with the *happens-before* partial order and with the
1699 modification orders of all the affected addresses. Each
1700 sequentially-consistent read sees the last preceding write to the
1701 same address in this global order. This corresponds to the C++0x/C1x
1702 ``memory_order_seq_cst`` and Java volatile.
1706 If an atomic operation is marked ``singlethread``, it only *synchronizes
1707 with* or participates in modification and seq\_cst total orderings with
1708 other operations running in the same thread (for example, in signal
1716 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1717 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1718 :ref:`frem <i_frem>`) have the following flags that can set to enable
1719 otherwise unsafe floating point operations
1722 No NaNs - Allow optimizations to assume the arguments and result are not
1723 NaN. Such optimizations are required to retain defined behavior over
1724 NaNs, but the value of the result is undefined.
1727 No Infs - Allow optimizations to assume the arguments and result are not
1728 +/-Inf. Such optimizations are required to retain defined behavior over
1729 +/-Inf, but the value of the result is undefined.
1732 No Signed Zeros - Allow optimizations to treat the sign of a zero
1733 argument or result as insignificant.
1736 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1737 argument rather than perform division.
1740 Fast - Allow algebraically equivalent transformations that may
1741 dramatically change results in floating point (e.g. reassociate). This
1742 flag implies all the others.
1749 The LLVM type system is one of the most important features of the
1750 intermediate representation. Being typed enables a number of
1751 optimizations to be performed on the intermediate representation
1752 directly, without having to do extra analyses on the side before the
1753 transformation. A strong type system makes it easier to read the
1754 generated code and enables novel analyses and transformations that are
1755 not feasible to perform on normal three address code representations.
1765 The void type does not represent any value and has no size.
1783 The function type can be thought of as a function signature. It consists of a
1784 return type and a list of formal parameter types. The return type of a function
1785 type is a void type or first class type --- except for :ref:`label <t_label>`
1786 and :ref:`metadata <t_metadata>` types.
1792 <returntype> (<parameter list>)
1794 ...where '``<parameter list>``' is a comma-separated list of type
1795 specifiers. Optionally, the parameter list may include a type ``...``, which
1796 indicates that the function takes a variable number of arguments. Variable
1797 argument functions can access their arguments with the :ref:`variable argument
1798 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
1799 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
1803 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1804 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1805 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1806 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1807 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1808 | ``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. |
1809 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1810 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1811 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1818 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1819 Values of these types are the only ones which can be produced by
1827 These are the types that are valid in registers from CodeGen's perspective.
1836 The integer type is a very simple type that simply specifies an
1837 arbitrary bit width for the integer type desired. Any bit width from 1
1838 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1846 The number of bits the integer will occupy is specified by the ``N``
1852 +----------------+------------------------------------------------+
1853 | ``i1`` | a single-bit integer. |
1854 +----------------+------------------------------------------------+
1855 | ``i32`` | a 32-bit integer. |
1856 +----------------+------------------------------------------------+
1857 | ``i1942652`` | a really big integer of over 1 million bits. |
1858 +----------------+------------------------------------------------+
1862 Floating Point Types
1863 """"""""""""""""""""
1872 - 16-bit floating point value
1875 - 32-bit floating point value
1878 - 64-bit floating point value
1881 - 128-bit floating point value (112-bit mantissa)
1884 - 80-bit floating point value (X87)
1887 - 128-bit floating point value (two 64-bits)
1894 The x86_mmx type represents a value held in an MMX register on an x86
1895 machine. The operations allowed on it are quite limited: parameters and
1896 return values, load and store, and bitcast. User-specified MMX
1897 instructions are represented as intrinsic or asm calls with arguments
1898 and/or results of this type. There are no arrays, vectors or constants
1915 The pointer type is used to specify memory locations. Pointers are
1916 commonly used to reference objects in memory.
1918 Pointer types may have an optional address space attribute defining the
1919 numbered address space where the pointed-to object resides. The default
1920 address space is number zero. The semantics of non-zero address spaces
1921 are target-specific.
1923 Note that LLVM does not permit pointers to void (``void*``) nor does it
1924 permit pointers to labels (``label*``). Use ``i8*`` instead.
1934 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1935 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
1936 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1937 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
1938 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1939 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
1940 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1949 A vector type is a simple derived type that represents a vector of
1950 elements. Vector types are used when multiple primitive data are
1951 operated in parallel using a single instruction (SIMD). A vector type
1952 requires a size (number of elements) and an underlying primitive data
1953 type. Vector types are considered :ref:`first class <t_firstclass>`.
1959 < <# elements> x <elementtype> >
1961 The number of elements is a constant integer value larger than 0;
1962 elementtype may be any integer or floating point type, or a pointer to
1963 these types. Vectors of size zero are not allowed.
1967 +-------------------+--------------------------------------------------+
1968 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
1969 +-------------------+--------------------------------------------------+
1970 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
1971 +-------------------+--------------------------------------------------+
1972 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
1973 +-------------------+--------------------------------------------------+
1974 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
1975 +-------------------+--------------------------------------------------+
1984 The label type represents code labels.
1999 The metadata type represents embedded metadata. No derived types may be
2000 created from metadata except for :ref:`function <t_function>` arguments.
2013 Aggregate Types are a subset of derived types that can contain multiple
2014 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
2015 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
2025 The array type is a very simple derived type that arranges elements
2026 sequentially in memory. The array type requires a size (number of
2027 elements) and an underlying data type.
2033 [<# elements> x <elementtype>]
2035 The number of elements is a constant integer value; ``elementtype`` may
2036 be any type with a size.
2040 +------------------+--------------------------------------+
2041 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
2042 +------------------+--------------------------------------+
2043 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
2044 +------------------+--------------------------------------+
2045 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
2046 +------------------+--------------------------------------+
2048 Here are some examples of multidimensional arrays:
2050 +-----------------------------+----------------------------------------------------------+
2051 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
2052 +-----------------------------+----------------------------------------------------------+
2053 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
2054 +-----------------------------+----------------------------------------------------------+
2055 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
2056 +-----------------------------+----------------------------------------------------------+
2058 There is no restriction on indexing beyond the end of the array implied
2059 by a static type (though there are restrictions on indexing beyond the
2060 bounds of an allocated object in some cases). This means that
2061 single-dimension 'variable sized array' addressing can be implemented in
2062 LLVM with a zero length array type. An implementation of 'pascal style
2063 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
2073 The structure type is used to represent a collection of data members
2074 together in memory. The elements of a structure may be any type that has
2077 Structures in memory are accessed using '``load``' and '``store``' by
2078 getting a pointer to a field with the '``getelementptr``' instruction.
2079 Structures in registers are accessed using the '``extractvalue``' and
2080 '``insertvalue``' instructions.
2082 Structures may optionally be "packed" structures, which indicate that
2083 the alignment of the struct is one byte, and that there is no padding
2084 between the elements. In non-packed structs, padding between field types
2085 is inserted as defined by the DataLayout string in the module, which is
2086 required to match what the underlying code generator expects.
2088 Structures can either be "literal" or "identified". A literal structure
2089 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
2090 identified types are always defined at the top level with a name.
2091 Literal types are uniqued by their contents and can never be recursive
2092 or opaque since there is no way to write one. Identified types can be
2093 recursive, can be opaqued, and are never uniqued.
2099 %T1 = type { <type list> } ; Identified normal struct type
2100 %T2 = type <{ <type list> }> ; Identified packed struct type
2104 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2105 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
2106 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2107 | ``{ 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``. |
2108 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2109 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
2110 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2114 Opaque Structure Types
2115 """"""""""""""""""""""
2119 Opaque structure types are used to represent named structure types that
2120 do not have a body specified. This corresponds (for example) to the C
2121 notion of a forward declared structure.
2132 +--------------+-------------------+
2133 | ``opaque`` | An opaque type. |
2134 +--------------+-------------------+
2141 LLVM has several different basic types of constants. This section
2142 describes them all and their syntax.
2147 **Boolean constants**
2148 The two strings '``true``' and '``false``' are both valid constants
2150 **Integer constants**
2151 Standard integers (such as '4') are constants of the
2152 :ref:`integer <t_integer>` type. Negative numbers may be used with
2154 **Floating point constants**
2155 Floating point constants use standard decimal notation (e.g.
2156 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2157 hexadecimal notation (see below). The assembler requires the exact
2158 decimal value of a floating-point constant. For example, the
2159 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2160 decimal in binary. Floating point constants must have a :ref:`floating
2161 point <t_floating>` type.
2162 **Null pointer constants**
2163 The identifier '``null``' is recognized as a null pointer constant
2164 and must be of :ref:`pointer type <t_pointer>`.
2166 The one non-intuitive notation for constants is the hexadecimal form of
2167 floating point constants. For example, the form
2168 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2169 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2170 constants are required (and the only time that they are generated by the
2171 disassembler) is when a floating point constant must be emitted but it
2172 cannot be represented as a decimal floating point number in a reasonable
2173 number of digits. For example, NaN's, infinities, and other special
2174 values are represented in their IEEE hexadecimal format so that assembly
2175 and disassembly do not cause any bits to change in the constants.
2177 When using the hexadecimal form, constants of types half, float, and
2178 double are represented using the 16-digit form shown above (which
2179 matches the IEEE754 representation for double); half and float values
2180 must, however, be exactly representable as IEEE 754 half and single
2181 precision, respectively. Hexadecimal format is always used for long
2182 double, and there are three forms of long double. The 80-bit format used
2183 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2184 128-bit format used by PowerPC (two adjacent doubles) is represented by
2185 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2186 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2187 will only work if they match the long double format on your target.
2188 The IEEE 16-bit format (half precision) is represented by ``0xH``
2189 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2190 (sign bit at the left).
2192 There are no constants of type x86_mmx.
2194 .. _complexconstants:
2199 Complex constants are a (potentially recursive) combination of simple
2200 constants and smaller complex constants.
2202 **Structure constants**
2203 Structure constants are represented with notation similar to
2204 structure type definitions (a comma separated list of elements,
2205 surrounded by braces (``{}``)). For example:
2206 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2207 "``@G = external global i32``". Structure constants must have
2208 :ref:`structure type <t_struct>`, and the number and types of elements
2209 must match those specified by the type.
2211 Array constants are represented with notation similar to array type
2212 definitions (a comma separated list of elements, surrounded by
2213 square brackets (``[]``)). For example:
2214 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2215 :ref:`array type <t_array>`, and the number and types of elements must
2216 match those specified by the type.
2217 **Vector constants**
2218 Vector constants are represented with notation similar to vector
2219 type definitions (a comma separated list of elements, surrounded by
2220 less-than/greater-than's (``<>``)). For example:
2221 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2222 must have :ref:`vector type <t_vector>`, and the number and types of
2223 elements must match those specified by the type.
2224 **Zero initialization**
2225 The string '``zeroinitializer``' can be used to zero initialize a
2226 value to zero of *any* type, including scalar and
2227 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2228 having to print large zero initializers (e.g. for large arrays) and
2229 is always exactly equivalent to using explicit zero initializers.
2231 A metadata node is a structure-like constant with :ref:`metadata
2232 type <t_metadata>`. For example:
2233 "``metadata !{ i32 0, metadata !"test" }``". Unlike other
2234 constants that are meant to be interpreted as part of the
2235 instruction stream, metadata is a place to attach additional
2236 information such as debug info.
2238 Global Variable and Function Addresses
2239 --------------------------------------
2241 The addresses of :ref:`global variables <globalvars>` and
2242 :ref:`functions <functionstructure>` are always implicitly valid
2243 (link-time) constants. These constants are explicitly referenced when
2244 the :ref:`identifier for the global <identifiers>` is used and always have
2245 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2248 .. code-block:: llvm
2252 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2259 The string '``undef``' can be used anywhere a constant is expected, and
2260 indicates that the user of the value may receive an unspecified
2261 bit-pattern. Undefined values may be of any type (other than '``label``'
2262 or '``void``') and be used anywhere a constant is permitted.
2264 Undefined values are useful because they indicate to the compiler that
2265 the program is well defined no matter what value is used. This gives the
2266 compiler more freedom to optimize. Here are some examples of
2267 (potentially surprising) transformations that are valid (in pseudo IR):
2269 .. code-block:: llvm
2279 This is safe because all of the output bits are affected by the undef
2280 bits. Any output bit can have a zero or one depending on the input bits.
2282 .. code-block:: llvm
2293 These logical operations have bits that are not always affected by the
2294 input. For example, if ``%X`` has a zero bit, then the output of the
2295 '``and``' operation will always be a zero for that bit, no matter what
2296 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2297 optimize or assume that the result of the '``and``' is '``undef``'.
2298 However, it is safe to assume that all bits of the '``undef``' could be
2299 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2300 all the bits of the '``undef``' operand to the '``or``' could be set,
2301 allowing the '``or``' to be folded to -1.
2303 .. code-block:: llvm
2305 %A = select undef, %X, %Y
2306 %B = select undef, 42, %Y
2307 %C = select %X, %Y, undef
2317 This set of examples shows that undefined '``select``' (and conditional
2318 branch) conditions can go *either way*, but they have to come from one
2319 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2320 both known to have a clear low bit, then ``%A`` would have to have a
2321 cleared low bit. However, in the ``%C`` example, the optimizer is
2322 allowed to assume that the '``undef``' operand could be the same as
2323 ``%Y``, allowing the whole '``select``' to be eliminated.
2325 .. code-block:: llvm
2327 %A = xor undef, undef
2344 This example points out that two '``undef``' operands are not
2345 necessarily the same. This can be surprising to people (and also matches
2346 C semantics) where they assume that "``X^X``" is always zero, even if
2347 ``X`` is undefined. This isn't true for a number of reasons, but the
2348 short answer is that an '``undef``' "variable" can arbitrarily change
2349 its value over its "live range". This is true because the variable
2350 doesn't actually *have a live range*. Instead, the value is logically
2351 read from arbitrary registers that happen to be around when needed, so
2352 the value is not necessarily consistent over time. In fact, ``%A`` and
2353 ``%C`` need to have the same semantics or the core LLVM "replace all
2354 uses with" concept would not hold.
2356 .. code-block:: llvm
2364 These examples show the crucial difference between an *undefined value*
2365 and *undefined behavior*. An undefined value (like '``undef``') is
2366 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2367 operation can be constant folded to '``undef``', because the '``undef``'
2368 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2369 However, in the second example, we can make a more aggressive
2370 assumption: because the ``undef`` is allowed to be an arbitrary value,
2371 we are allowed to assume that it could be zero. Since a divide by zero
2372 has *undefined behavior*, we are allowed to assume that the operation
2373 does not execute at all. This allows us to delete the divide and all
2374 code after it. Because the undefined operation "can't happen", the
2375 optimizer can assume that it occurs in dead code.
2377 .. code-block:: llvm
2379 a: store undef -> %X
2380 b: store %X -> undef
2385 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2386 value can be assumed to not have any effect; we can assume that the
2387 value is overwritten with bits that happen to match what was already
2388 there. However, a store *to* an undefined location could clobber
2389 arbitrary memory, therefore, it has undefined behavior.
2396 Poison values are similar to :ref:`undef values <undefvalues>`, however
2397 they also represent the fact that an instruction or constant expression
2398 which cannot evoke side effects has nevertheless detected a condition
2399 which results in undefined behavior.
2401 There is currently no way of representing a poison value in the IR; they
2402 only exist when produced by operations such as :ref:`add <i_add>` with
2405 Poison value behavior is defined in terms of value *dependence*:
2407 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2408 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2409 their dynamic predecessor basic block.
2410 - Function arguments depend on the corresponding actual argument values
2411 in the dynamic callers of their functions.
2412 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2413 instructions that dynamically transfer control back to them.
2414 - :ref:`Invoke <i_invoke>` instructions depend on the
2415 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2416 call instructions that dynamically transfer control back to them.
2417 - Non-volatile loads and stores depend on the most recent stores to all
2418 of the referenced memory addresses, following the order in the IR
2419 (including loads and stores implied by intrinsics such as
2420 :ref:`@llvm.memcpy <int_memcpy>`.)
2421 - An instruction with externally visible side effects depends on the
2422 most recent preceding instruction with externally visible side
2423 effects, following the order in the IR. (This includes :ref:`volatile
2424 operations <volatile>`.)
2425 - An instruction *control-depends* on a :ref:`terminator
2426 instruction <terminators>` if the terminator instruction has
2427 multiple successors and the instruction is always executed when
2428 control transfers to one of the successors, and may not be executed
2429 when control is transferred to another.
2430 - Additionally, an instruction also *control-depends* on a terminator
2431 instruction if the set of instructions it otherwise depends on would
2432 be different if the terminator had transferred control to a different
2434 - Dependence is transitive.
2436 Poison Values have the same behavior as :ref:`undef values <undefvalues>`,
2437 with the additional affect that any instruction which has a *dependence*
2438 on a poison value has undefined behavior.
2440 Here are some examples:
2442 .. code-block:: llvm
2445 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2446 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2447 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
2448 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2450 store i32 %poison, i32* @g ; Poison value stored to memory.
2451 %poison2 = load i32* @g ; Poison value loaded back from memory.
2453 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2455 %narrowaddr = bitcast i32* @g to i16*
2456 %wideaddr = bitcast i32* @g to i64*
2457 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2458 %poison4 = load i64* %wideaddr ; Returns a poison value.
2460 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2461 br i1 %cmp, label %true, label %end ; Branch to either destination.
2464 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2465 ; it has undefined behavior.
2469 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2470 ; Both edges into this PHI are
2471 ; control-dependent on %cmp, so this
2472 ; always results in a poison value.
2474 store volatile i32 0, i32* @g ; This would depend on the store in %true
2475 ; if %cmp is true, or the store in %entry
2476 ; otherwise, so this is undefined behavior.
2478 br i1 %cmp, label %second_true, label %second_end
2479 ; The same branch again, but this time the
2480 ; true block doesn't have side effects.
2487 store volatile i32 0, i32* @g ; This time, the instruction always depends
2488 ; on the store in %end. Also, it is
2489 ; control-equivalent to %end, so this is
2490 ; well-defined (ignoring earlier undefined
2491 ; behavior in this example).
2495 Addresses of Basic Blocks
2496 -------------------------
2498 ``blockaddress(@function, %block)``
2500 The '``blockaddress``' constant computes the address of the specified
2501 basic block in the specified function, and always has an ``i8*`` type.
2502 Taking the address of the entry block is illegal.
2504 This value only has defined behavior when used as an operand to the
2505 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2506 against null. Pointer equality tests between labels addresses results in
2507 undefined behavior --- though, again, comparison against null is ok, and
2508 no label is equal to the null pointer. This may be passed around as an
2509 opaque pointer sized value as long as the bits are not inspected. This
2510 allows ``ptrtoint`` and arithmetic to be performed on these values so
2511 long as the original value is reconstituted before the ``indirectbr``
2514 Finally, some targets may provide defined semantics when using the value
2515 as the operand to an inline assembly, but that is target specific.
2519 Constant Expressions
2520 --------------------
2522 Constant expressions are used to allow expressions involving other
2523 constants to be used as constants. Constant expressions may be of any
2524 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2525 that does not have side effects (e.g. load and call are not supported).
2526 The following is the syntax for constant expressions:
2528 ``trunc (CST to TYPE)``
2529 Truncate a constant to another type. The bit size of CST must be
2530 larger than the bit size of TYPE. Both types must be integers.
2531 ``zext (CST to TYPE)``
2532 Zero extend a constant to another type. The bit size of CST must be
2533 smaller than the bit size of TYPE. Both types must be integers.
2534 ``sext (CST to TYPE)``
2535 Sign extend a constant to another type. The bit size of CST must be
2536 smaller than the bit size of TYPE. Both types must be integers.
2537 ``fptrunc (CST to TYPE)``
2538 Truncate a floating point constant to another floating point type.
2539 The size of CST must be larger than the size of TYPE. Both types
2540 must be floating point.
2541 ``fpext (CST to TYPE)``
2542 Floating point extend a constant to another type. The size of CST
2543 must be smaller or equal to the size of TYPE. Both types must be
2545 ``fptoui (CST to TYPE)``
2546 Convert a floating point constant to the corresponding unsigned
2547 integer constant. TYPE must be a scalar or vector integer type. CST
2548 must be of scalar or vector floating point type. Both CST and TYPE
2549 must be scalars, or vectors of the same number of elements. If the
2550 value won't fit in the integer type, the results are undefined.
2551 ``fptosi (CST to TYPE)``
2552 Convert a floating point constant to the corresponding signed
2553 integer constant. TYPE must be a scalar or vector integer type. CST
2554 must be of scalar or vector floating point type. Both CST and TYPE
2555 must be scalars, or vectors of the same number of elements. If the
2556 value won't fit in the integer type, the results are undefined.
2557 ``uitofp (CST to TYPE)``
2558 Convert an unsigned integer constant to the corresponding floating
2559 point constant. TYPE must be a scalar or vector floating point type.
2560 CST must be of scalar or vector integer type. Both CST and TYPE must
2561 be scalars, or vectors of the same number of elements. If the value
2562 won't fit in the floating point type, the results are undefined.
2563 ``sitofp (CST to TYPE)``
2564 Convert a signed integer constant to the corresponding floating
2565 point constant. TYPE must be a scalar or vector floating point type.
2566 CST must be of scalar or vector integer type. Both CST and TYPE must
2567 be scalars, or vectors of the same number of elements. If the value
2568 won't fit in the floating point type, the results are undefined.
2569 ``ptrtoint (CST to TYPE)``
2570 Convert a pointer typed constant to the corresponding integer
2571 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2572 pointer type. The ``CST`` value is zero extended, truncated, or
2573 unchanged to make it fit in ``TYPE``.
2574 ``inttoptr (CST to TYPE)``
2575 Convert an integer constant to a pointer constant. TYPE must be a
2576 pointer type. CST must be of integer type. The CST value is zero
2577 extended, truncated, or unchanged to make it fit in a pointer size.
2578 This one is *really* dangerous!
2579 ``bitcast (CST to TYPE)``
2580 Convert a constant, CST, to another TYPE. The constraints of the
2581 operands are the same as those for the :ref:`bitcast
2582 instruction <i_bitcast>`.
2583 ``addrspacecast (CST to TYPE)``
2584 Convert a constant pointer or constant vector of pointer, CST, to another
2585 TYPE in a different address space. The constraints of the operands are the
2586 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2587 ``getelementptr (CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)``
2588 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2589 constants. As with the :ref:`getelementptr <i_getelementptr>`
2590 instruction, the index list may have zero or more indexes, which are
2591 required to make sense for the type of "CSTPTR".
2592 ``select (COND, VAL1, VAL2)``
2593 Perform the :ref:`select operation <i_select>` on constants.
2594 ``icmp COND (VAL1, VAL2)``
2595 Performs the :ref:`icmp operation <i_icmp>` on constants.
2596 ``fcmp COND (VAL1, VAL2)``
2597 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2598 ``extractelement (VAL, IDX)``
2599 Perform the :ref:`extractelement operation <i_extractelement>` on
2601 ``insertelement (VAL, ELT, IDX)``
2602 Perform the :ref:`insertelement operation <i_insertelement>` on
2604 ``shufflevector (VEC1, VEC2, IDXMASK)``
2605 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2607 ``extractvalue (VAL, IDX0, IDX1, ...)``
2608 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2609 constants. The index list is interpreted in a similar manner as
2610 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2611 least one index value must be specified.
2612 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2613 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2614 The index list is interpreted in a similar manner as indices in a
2615 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2616 value must be specified.
2617 ``OPCODE (LHS, RHS)``
2618 Perform the specified operation of the LHS and RHS constants. OPCODE
2619 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2620 binary <bitwiseops>` operations. The constraints on operands are
2621 the same as those for the corresponding instruction (e.g. no bitwise
2622 operations on floating point values are allowed).
2629 Inline Assembler Expressions
2630 ----------------------------
2632 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2633 Inline Assembly <moduleasm>`) through the use of a special value. This
2634 value represents the inline assembler as a string (containing the
2635 instructions to emit), a list of operand constraints (stored as a
2636 string), a flag that indicates whether or not the inline asm expression
2637 has side effects, and a flag indicating whether the function containing
2638 the asm needs to align its stack conservatively. An example inline
2639 assembler expression is:
2641 .. code-block:: llvm
2643 i32 (i32) asm "bswap $0", "=r,r"
2645 Inline assembler expressions may **only** be used as the callee operand
2646 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2647 Thus, typically we have:
2649 .. code-block:: llvm
2651 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2653 Inline asms with side effects not visible in the constraint list must be
2654 marked as having side effects. This is done through the use of the
2655 '``sideeffect``' keyword, like so:
2657 .. code-block:: llvm
2659 call void asm sideeffect "eieio", ""()
2661 In some cases inline asms will contain code that will not work unless
2662 the stack is aligned in some way, such as calls or SSE instructions on
2663 x86, yet will not contain code that does that alignment within the asm.
2664 The compiler should make conservative assumptions about what the asm
2665 might contain and should generate its usual stack alignment code in the
2666 prologue if the '``alignstack``' keyword is present:
2668 .. code-block:: llvm
2670 call void asm alignstack "eieio", ""()
2672 Inline asms also support using non-standard assembly dialects. The
2673 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2674 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2675 the only supported dialects. An example is:
2677 .. code-block:: llvm
2679 call void asm inteldialect "eieio", ""()
2681 If multiple keywords appear the '``sideeffect``' keyword must come
2682 first, the '``alignstack``' keyword second and the '``inteldialect``'
2688 The call instructions that wrap inline asm nodes may have a
2689 "``!srcloc``" MDNode attached to it that contains a list of constant
2690 integers. If present, the code generator will use the integer as the
2691 location cookie value when report errors through the ``LLVMContext``
2692 error reporting mechanisms. This allows a front-end to correlate backend
2693 errors that occur with inline asm back to the source code that produced
2696 .. code-block:: llvm
2698 call void asm sideeffect "something bad", ""(), !srcloc !42
2700 !42 = !{ i32 1234567 }
2702 It is up to the front-end to make sense of the magic numbers it places
2703 in the IR. If the MDNode contains multiple constants, the code generator
2704 will use the one that corresponds to the line of the asm that the error
2709 Metadata Nodes and Metadata Strings
2710 -----------------------------------
2712 LLVM IR allows metadata to be attached to instructions in the program
2713 that can convey extra information about the code to the optimizers and
2714 code generator. One example application of metadata is source-level
2715 debug information. There are two metadata primitives: strings and nodes.
2716 All metadata has the ``metadata`` type and is identified in syntax by a
2717 preceding exclamation point ('``!``').
2719 A metadata string is a string surrounded by double quotes. It can
2720 contain any character by escaping non-printable characters with
2721 "``\xx``" where "``xx``" is the two digit hex code. For example:
2724 Metadata nodes are represented with notation similar to structure
2725 constants (a comma separated list of elements, surrounded by braces and
2726 preceded by an exclamation point). Metadata nodes can have any values as
2727 their operand. For example:
2729 .. code-block:: llvm
2731 !{ metadata !"test\00", i32 10}
2733 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2734 metadata nodes, which can be looked up in the module symbol table. For
2737 .. code-block:: llvm
2739 !foo = metadata !{!4, !3}
2741 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2742 function is using two metadata arguments:
2744 .. code-block:: llvm
2746 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2748 Metadata can be attached with an instruction. Here metadata ``!21`` is
2749 attached to the ``add`` instruction using the ``!dbg`` identifier:
2751 .. code-block:: llvm
2753 %indvar.next = add i64 %indvar, 1, !dbg !21
2755 More information about specific metadata nodes recognized by the
2756 optimizers and code generator is found below.
2761 In LLVM IR, memory does not have types, so LLVM's own type system is not
2762 suitable for doing TBAA. Instead, metadata is added to the IR to
2763 describe a type system of a higher level language. This can be used to
2764 implement typical C/C++ TBAA, but it can also be used to implement
2765 custom alias analysis behavior for other languages.
2767 The current metadata format is very simple. TBAA metadata nodes have up
2768 to three fields, e.g.:
2770 .. code-block:: llvm
2772 !0 = metadata !{ metadata !"an example type tree" }
2773 !1 = metadata !{ metadata !"int", metadata !0 }
2774 !2 = metadata !{ metadata !"float", metadata !0 }
2775 !3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2777 The first field is an identity field. It can be any value, usually a
2778 metadata string, which uniquely identifies the type. The most important
2779 name in the tree is the name of the root node. Two trees with different
2780 root node names are entirely disjoint, even if they have leaves with
2783 The second field identifies the type's parent node in the tree, or is
2784 null or omitted for a root node. A type is considered to alias all of
2785 its descendants and all of its ancestors in the tree. Also, a type is
2786 considered to alias all types in other trees, so that bitcode produced
2787 from multiple front-ends is handled conservatively.
2789 If the third field is present, it's an integer which if equal to 1
2790 indicates that the type is "constant" (meaning
2791 ``pointsToConstantMemory`` should return true; see `other useful
2792 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
2794 '``tbaa.struct``' Metadata
2795 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2797 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
2798 aggregate assignment operations in C and similar languages, however it
2799 is defined to copy a contiguous region of memory, which is more than
2800 strictly necessary for aggregate types which contain holes due to
2801 padding. Also, it doesn't contain any TBAA information about the fields
2804 ``!tbaa.struct`` metadata can describe which memory subregions in a
2805 memcpy are padding and what the TBAA tags of the struct are.
2807 The current metadata format is very simple. ``!tbaa.struct`` metadata
2808 nodes are a list of operands which are in conceptual groups of three.
2809 For each group of three, the first operand gives the byte offset of a
2810 field in bytes, the second gives its size in bytes, and the third gives
2813 .. code-block:: llvm
2815 !4 = metadata !{ i64 0, i64 4, metadata !1, i64 8, i64 4, metadata !2 }
2817 This describes a struct with two fields. The first is at offset 0 bytes
2818 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
2819 and has size 4 bytes and has tbaa tag !2.
2821 Note that the fields need not be contiguous. In this example, there is a
2822 4 byte gap between the two fields. This gap represents padding which
2823 does not carry useful data and need not be preserved.
2825 '``fpmath``' Metadata
2826 ^^^^^^^^^^^^^^^^^^^^^
2828 ``fpmath`` metadata may be attached to any instruction of floating point
2829 type. It can be used to express the maximum acceptable error in the
2830 result of that instruction, in ULPs, thus potentially allowing the
2831 compiler to use a more efficient but less accurate method of computing
2832 it. ULP is defined as follows:
2834 If ``x`` is a real number that lies between two finite consecutive
2835 floating-point numbers ``a`` and ``b``, without being equal to one
2836 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
2837 distance between the two non-equal finite floating-point numbers
2838 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
2840 The metadata node shall consist of a single positive floating point
2841 number representing the maximum relative error, for example:
2843 .. code-block:: llvm
2845 !0 = metadata !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
2847 '``range``' Metadata
2848 ^^^^^^^^^^^^^^^^^^^^
2850 ``range`` metadata may be attached only to ``load``, ``call`` and ``invoke`` of
2851 integer types. It expresses the possible ranges the loaded value or the value
2852 returned by the called function at this call site is in. The ranges are
2853 represented with a flattened list of integers. The loaded value or the value
2854 returned is known to be in the union of the ranges defined by each consecutive
2855 pair. Each pair has the following properties:
2857 - The type must match the type loaded by the instruction.
2858 - The pair ``a,b`` represents the range ``[a,b)``.
2859 - Both ``a`` and ``b`` are constants.
2860 - The range is allowed to wrap.
2861 - The range should not represent the full or empty set. That is,
2864 In addition, the pairs must be in signed order of the lower bound and
2865 they must be non-contiguous.
2869 .. code-block:: llvm
2871 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
2872 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
2873 %c = call i8 @foo(), !range !2 ; Can only be 0, 1, 3, 4 or 5
2874 %d = invoke i8 @bar() to label %cont
2875 unwind label %lpad, !range !3 ; Can only be -2, -1, 3, 4 or 5
2877 !0 = metadata !{ i8 0, i8 2 }
2878 !1 = metadata !{ i8 255, i8 2 }
2879 !2 = metadata !{ i8 0, i8 2, i8 3, i8 6 }
2880 !3 = metadata !{ i8 -2, i8 0, i8 3, i8 6 }
2885 It is sometimes useful to attach information to loop constructs. Currently,
2886 loop metadata is implemented as metadata attached to the branch instruction
2887 in the loop latch block. This type of metadata refer to a metadata node that is
2888 guaranteed to be separate for each loop. The loop identifier metadata is
2889 specified with the name ``llvm.loop``.
2891 The loop identifier metadata is implemented using a metadata that refers to
2892 itself to avoid merging it with any other identifier metadata, e.g.,
2893 during module linkage or function inlining. That is, each loop should refer
2894 to their own identification metadata even if they reside in separate functions.
2895 The following example contains loop identifier metadata for two separate loop
2898 .. code-block:: llvm
2900 !0 = metadata !{ metadata !0 }
2901 !1 = metadata !{ metadata !1 }
2903 The loop identifier metadata can be used to specify additional
2904 per-loop metadata. Any operands after the first operand can be treated
2905 as user-defined metadata. For example the ``llvm.loop.unroll.count``
2906 suggests an unroll factor to the loop unroller:
2908 .. code-block:: llvm
2910 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
2912 !0 = metadata !{ metadata !0, metadata !1 }
2913 !1 = metadata !{ metadata !"llvm.loop.vectorize.width", i32 4 }
2915 '``llvm.loop.vectorize``'
2916 ^^^^^^^^^^^^^^^^^^^^^^^^^
2918 Metadata prefixed with ``llvm.loop.vectorize`` is used to control
2919 per-loop vectorization parameters such as vectorization width and
2920 interleave count. ``llvm.loop.vectorize`` metadata should be used in
2921 conjunction with ``llvm.loop`` loop identification metadata. The
2922 ``llvm.loop.vectorize`` metadata are only optimization hints and the
2923 vectorizer will only vectorize loops if it believes it is safe to do
2924 so. The ``llvm.mem.parallel_loop_access`` metadata which contains
2925 information about loop-carried memory dependencies can be helpful in
2926 determining the safety of loop vectorization.
2928 '``llvm.loop.vectorize.unroll``' Metadata
2929 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2931 This metadata suggests an interleave count to the loop vectorizer.
2932 The first operand is the string ``llvm.loop.vectorize.unroll`` and the
2933 second operand is an integer specifying the interleave count. For
2936 .. code-block:: llvm
2938 !0 = metadata !{ metadata !"llvm.loop.vectorize.unroll", i32 4 }
2940 Note that setting ``llvm.loop.vectorize.unroll`` to 1 disables
2941 interleaving multiple iterations of the loop. If
2942 ``llvm.loop.vectorize.unroll`` is set to 0 then the interleave count
2943 will be determined automatically.
2945 '``llvm.loop.vectorize.width``' Metadata
2946 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2948 This metadata sets the target width of the vectorizer. The first
2949 operand is the string ``llvm.loop.vectorize.width`` and the second
2950 operand is an integer specifying the width. For example:
2952 .. code-block:: llvm
2954 !0 = metadata !{ metadata !"llvm.loop.vectorize.width", i32 4 }
2956 Note that setting ``llvm.loop.vectorize.width`` to 1 disables
2957 vectorization of the loop. If ``llvm.loop.vectorize.width`` is set to
2958 0 or if the loop does not have this metadata the width will be
2959 determined automatically.
2961 '``llvm.loop.unroll``'
2962 ^^^^^^^^^^^^^^^^^^^^^^
2964 Metadata prefixed with ``llvm.loop.unroll`` are loop unrolling
2965 optimization hints such as the unroll factor. ``llvm.loop.unroll``
2966 metadata should be used in conjunction with ``llvm.loop`` loop
2967 identification metadata. The ``llvm.loop.unroll`` metadata are only
2968 optimization hints and the unrolling will only be performed if the
2969 optimizer believes it is safe to do so.
2971 '``llvm.loop.unroll.enable``' Metadata
2972 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2974 This metadata either disables loop unrolling or suggests that the loop
2975 be unrolled fully. The first operand is the string
2976 ``llvm.loop.unroll.enable`` and the second operand is a bit. If the
2977 bit operand value is 0 loop unrolling is disabled. A value of 1
2978 indicates that the loop should be fully unrolled. For example:
2980 .. code-block:: llvm
2982 !0 = metadata !{ metadata !"llvm.loop.unroll.enable", i1 0 }
2983 !1 = metadata !{ metadata !"llvm.loop.unroll.enable", i1 1 }
2985 '``llvm.loop.unroll.count``' Metadata
2986 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2988 This metadata suggests an unroll factor to the loop unroller. The
2989 first operand is the string ``llvm.loop.unroll.count`` and the second
2990 operand is a positive integer specifying the unroll factor. For
2993 .. code-block:: llvm
2995 !0 = metadata !{ metadata !"llvm.loop.unroll.count", i32 4 }
2997 If the trip count of the loop is less than the unroll count the loop
2998 will be partially unrolled.
3000 If a loop has both a ``llvm.loop.unroll.enable`` metadata and
3001 ``llvm.loop.unroll.count`` metadata the behavior depends upon the
3002 value of the ``llvm.loop.unroll.enable`` operand. If the value is 0,
3003 the loop will not be unrolled. If the value is 1, the loop will be
3004 unrolled with a factor determined by the ``llvm.loop.unroll.count``
3005 operand effectively ignoring the ``llvm.loop.unroll.enable`` metadata.
3010 Metadata types used to annotate memory accesses with information helpful
3011 for optimizations are prefixed with ``llvm.mem``.
3013 '``llvm.mem.parallel_loop_access``' Metadata
3014 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3016 The ``llvm.mem.parallel_loop_access`` metadata refers to a loop identifier,
3017 or metadata containing a list of loop identifiers for nested loops.
3018 The metadata is attached to memory accessing instructions and denotes that
3019 no loop carried memory dependence exist between it and other instructions denoted
3020 with the same loop identifier.
3022 Precisely, given two instructions ``m1`` and ``m2`` that both have the
3023 ``llvm.mem.parallel_loop_access`` metadata, with ``L1`` and ``L2`` being the
3024 set of loops associated with that metadata, respectively, then there is no loop
3025 carried dependence between ``m1`` and ``m2`` for loops in both ``L1`` and
3028 As a special case, if all memory accessing instructions in a loop have
3029 ``llvm.mem.parallel_loop_access`` metadata that refers to that loop, then the
3030 loop has no loop carried memory dependences and is considered to be a parallel
3033 Note that if not all memory access instructions have such metadata referring to
3034 the loop, then the loop is considered not being trivially parallel. Additional
3035 memory dependence analysis is required to make that determination. As a fail
3036 safe mechanism, this causes loops that were originally parallel to be considered
3037 sequential (if optimization passes that are unaware of the parallel semantics
3038 insert new memory instructions into the loop body).
3040 Example of a loop that is considered parallel due to its correct use of
3041 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
3042 metadata types that refer to the same loop identifier metadata.
3044 .. code-block:: llvm
3048 %val0 = load i32* %arrayidx, !llvm.mem.parallel_loop_access !0
3050 store i32 %val0, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3052 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
3056 !0 = metadata !{ metadata !0 }
3058 It is also possible to have nested parallel loops. In that case the
3059 memory accesses refer to a list of loop identifier metadata nodes instead of
3060 the loop identifier metadata node directly:
3062 .. code-block:: llvm
3066 %val1 = load i32* %arrayidx3, !llvm.mem.parallel_loop_access !2
3068 br label %inner.for.body
3072 %val0 = load i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3074 store i32 %val0, i32* %arrayidx2, !llvm.mem.parallel_loop_access !0
3076 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
3080 store i32 %val1, i32* %arrayidx4, !llvm.mem.parallel_loop_access !2
3082 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
3084 outer.for.end: ; preds = %for.body
3086 !0 = metadata !{ metadata !1, metadata !2 } ; a list of loop identifiers
3087 !1 = metadata !{ metadata !1 } ; an identifier for the inner loop
3088 !2 = metadata !{ metadata !2 } ; an identifier for the outer loop
3090 Module Flags Metadata
3091 =====================
3093 Information about the module as a whole is difficult to convey to LLVM's
3094 subsystems. The LLVM IR isn't sufficient to transmit this information.
3095 The ``llvm.module.flags`` named metadata exists in order to facilitate
3096 this. These flags are in the form of key / value pairs --- much like a
3097 dictionary --- making it easy for any subsystem who cares about a flag to
3100 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
3101 Each triplet has the following form:
3103 - The first element is a *behavior* flag, which specifies the behavior
3104 when two (or more) modules are merged together, and it encounters two
3105 (or more) metadata with the same ID. The supported behaviors are
3107 - The second element is a metadata string that is a unique ID for the
3108 metadata. Each module may only have one flag entry for each unique ID (not
3109 including entries with the **Require** behavior).
3110 - The third element is the value of the flag.
3112 When two (or more) modules are merged together, the resulting
3113 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
3114 each unique metadata ID string, there will be exactly one entry in the merged
3115 modules ``llvm.module.flags`` metadata table, and the value for that entry will
3116 be determined by the merge behavior flag, as described below. The only exception
3117 is that entries with the *Require* behavior are always preserved.
3119 The following behaviors are supported:
3130 Emits an error if two values disagree, otherwise the resulting value
3131 is that of the operands.
3135 Emits a warning if two values disagree. The result value will be the
3136 operand for the flag from the first module being linked.
3140 Adds a requirement that another module flag be present and have a
3141 specified value after linking is performed. The value must be a
3142 metadata pair, where the first element of the pair is the ID of the
3143 module flag to be restricted, and the second element of the pair is
3144 the value the module flag should be restricted to. This behavior can
3145 be used to restrict the allowable results (via triggering of an
3146 error) of linking IDs with the **Override** behavior.
3150 Uses the specified value, regardless of the behavior or value of the
3151 other module. If both modules specify **Override**, but the values
3152 differ, an error will be emitted.
3156 Appends the two values, which are required to be metadata nodes.
3160 Appends the two values, which are required to be metadata
3161 nodes. However, duplicate entries in the second list are dropped
3162 during the append operation.
3164 It is an error for a particular unique flag ID to have multiple behaviors,
3165 except in the case of **Require** (which adds restrictions on another metadata
3166 value) or **Override**.
3168 An example of module flags:
3170 .. code-block:: llvm
3172 !0 = metadata !{ i32 1, metadata !"foo", i32 1 }
3173 !1 = metadata !{ i32 4, metadata !"bar", i32 37 }
3174 !2 = metadata !{ i32 2, metadata !"qux", i32 42 }
3175 !3 = metadata !{ i32 3, metadata !"qux",
3177 metadata !"foo", i32 1
3180 !llvm.module.flags = !{ !0, !1, !2, !3 }
3182 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
3183 if two or more ``!"foo"`` flags are seen is to emit an error if their
3184 values are not equal.
3186 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
3187 behavior if two or more ``!"bar"`` flags are seen is to use the value
3190 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
3191 behavior if two or more ``!"qux"`` flags are seen is to emit a
3192 warning if their values are not equal.
3194 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
3198 metadata !{ metadata !"foo", i32 1 }
3200 The behavior is to emit an error if the ``llvm.module.flags`` does not
3201 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
3204 Objective-C Garbage Collection Module Flags Metadata
3205 ----------------------------------------------------
3207 On the Mach-O platform, Objective-C stores metadata about garbage
3208 collection in a special section called "image info". The metadata
3209 consists of a version number and a bitmask specifying what types of
3210 garbage collection are supported (if any) by the file. If two or more
3211 modules are linked together their garbage collection metadata needs to
3212 be merged rather than appended together.
3214 The Objective-C garbage collection module flags metadata consists of the
3215 following key-value pairs:
3224 * - ``Objective-C Version``
3225 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
3227 * - ``Objective-C Image Info Version``
3228 - **[Required]** --- The version of the image info section. Currently
3231 * - ``Objective-C Image Info Section``
3232 - **[Required]** --- The section to place the metadata. Valid values are
3233 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
3234 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
3235 Objective-C ABI version 2.
3237 * - ``Objective-C Garbage Collection``
3238 - **[Required]** --- Specifies whether garbage collection is supported or
3239 not. Valid values are 0, for no garbage collection, and 2, for garbage
3240 collection supported.
3242 * - ``Objective-C GC Only``
3243 - **[Optional]** --- Specifies that only garbage collection is supported.
3244 If present, its value must be 6. This flag requires that the
3245 ``Objective-C Garbage Collection`` flag have the value 2.
3247 Some important flag interactions:
3249 - If a module with ``Objective-C Garbage Collection`` set to 0 is
3250 merged with a module with ``Objective-C Garbage Collection`` set to
3251 2, then the resulting module has the
3252 ``Objective-C Garbage Collection`` flag set to 0.
3253 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
3254 merged with a module with ``Objective-C GC Only`` set to 6.
3256 Automatic Linker Flags Module Flags Metadata
3257 --------------------------------------------
3259 Some targets support embedding flags to the linker inside individual object
3260 files. Typically this is used in conjunction with language extensions which
3261 allow source files to explicitly declare the libraries they depend on, and have
3262 these automatically be transmitted to the linker via object files.
3264 These flags are encoded in the IR using metadata in the module flags section,
3265 using the ``Linker Options`` key. The merge behavior for this flag is required
3266 to be ``AppendUnique``, and the value for the key is expected to be a metadata
3267 node which should be a list of other metadata nodes, each of which should be a
3268 list of metadata strings defining linker options.
3270 For example, the following metadata section specifies two separate sets of
3271 linker options, presumably to link against ``libz`` and the ``Cocoa``
3274 !0 = metadata !{ i32 6, metadata !"Linker Options",
3276 metadata !{ metadata !"-lz" },
3277 metadata !{ metadata !"-framework", metadata !"Cocoa" } } }
3278 !llvm.module.flags = !{ !0 }
3280 The metadata encoding as lists of lists of options, as opposed to a collapsed
3281 list of options, is chosen so that the IR encoding can use multiple option
3282 strings to specify e.g., a single library, while still having that specifier be
3283 preserved as an atomic element that can be recognized by a target specific
3284 assembly writer or object file emitter.
3286 Each individual option is required to be either a valid option for the target's
3287 linker, or an option that is reserved by the target specific assembly writer or
3288 object file emitter. No other aspect of these options is defined by the IR.
3290 C type width Module Flags Metadata
3291 ----------------------------------
3293 The ARM backend emits a section into each generated object file describing the
3294 options that it was compiled with (in a compiler-independent way) to prevent
3295 linking incompatible objects, and to allow automatic library selection. Some
3296 of these options are not visible at the IR level, namely wchar_t width and enum
3299 To pass this information to the backend, these options are encoded in module
3300 flags metadata, using the following key-value pairs:
3310 - * 0 --- sizeof(wchar_t) == 4
3311 * 1 --- sizeof(wchar_t) == 2
3314 - * 0 --- Enums are at least as large as an ``int``.
3315 * 1 --- Enums are stored in the smallest integer type which can
3316 represent all of its values.
3318 For example, the following metadata section specifies that the module was
3319 compiled with a ``wchar_t`` width of 4 bytes, and the underlying type of an
3320 enum is the smallest type which can represent all of its values::
3322 !llvm.module.flags = !{!0, !1}
3323 !0 = metadata !{i32 1, metadata !"short_wchar", i32 1}
3324 !1 = metadata !{i32 1, metadata !"short_enum", i32 0}
3326 .. _intrinsicglobalvariables:
3328 Intrinsic Global Variables
3329 ==========================
3331 LLVM has a number of "magic" global variables that contain data that
3332 affect code generation or other IR semantics. These are documented here.
3333 All globals of this sort should have a section specified as
3334 "``llvm.metadata``". This section and all globals that start with
3335 "``llvm.``" are reserved for use by LLVM.
3339 The '``llvm.used``' Global Variable
3340 -----------------------------------
3342 The ``@llvm.used`` global is an array which has
3343 :ref:`appending linkage <linkage_appending>`. This array contains a list of
3344 pointers to named global variables, functions and aliases which may optionally
3345 have a pointer cast formed of bitcast or getelementptr. For example, a legal
3348 .. code-block:: llvm
3353 @llvm.used = appending global [2 x i8*] [
3355 i8* bitcast (i32* @Y to i8*)
3356 ], section "llvm.metadata"
3358 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
3359 and linker are required to treat the symbol as if there is a reference to the
3360 symbol that it cannot see (which is why they have to be named). For example, if
3361 a variable has internal linkage and no references other than that from the
3362 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
3363 references from inline asms and other things the compiler cannot "see", and
3364 corresponds to "``attribute((used))``" in GNU C.
3366 On some targets, the code generator must emit a directive to the
3367 assembler or object file to prevent the assembler and linker from
3368 molesting the symbol.
3370 .. _gv_llvmcompilerused:
3372 The '``llvm.compiler.used``' Global Variable
3373 --------------------------------------------
3375 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
3376 directive, except that it only prevents the compiler from touching the
3377 symbol. On targets that support it, this allows an intelligent linker to
3378 optimize references to the symbol without being impeded as it would be
3381 This is a rare construct that should only be used in rare circumstances,
3382 and should not be exposed to source languages.
3384 .. _gv_llvmglobalctors:
3386 The '``llvm.global_ctors``' Global Variable
3387 -------------------------------------------
3389 .. code-block:: llvm
3391 %0 = type { i32, void ()*, i8* }
3392 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor, i8* @data }]
3394 The ``@llvm.global_ctors`` array contains a list of constructor
3395 functions, priorities, and an optional associated global or function.
3396 The functions referenced by this array will be called in ascending order
3397 of priority (i.e. lowest first) when the module is loaded. The order of
3398 functions with the same priority is not defined.
3400 If the third field is present, non-null, and points to a global variable
3401 or function, the initializer function will only run if the associated
3402 data from the current module is not discarded.
3404 .. _llvmglobaldtors:
3406 The '``llvm.global_dtors``' Global Variable
3407 -------------------------------------------
3409 .. code-block:: llvm
3411 %0 = type { i32, void ()*, i8* }
3412 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor, i8* @data }]
3414 The ``@llvm.global_dtors`` array contains a list of destructor
3415 functions, priorities, and an optional associated global or function.
3416 The functions referenced by this array will be called in descending
3417 order of priority (i.e. highest first) when the module is unloaded. The
3418 order of functions with the same priority is not defined.
3420 If the third field is present, non-null, and points to a global variable
3421 or function, the destructor function will only run if the associated
3422 data from the current module is not discarded.
3424 Instruction Reference
3425 =====================
3427 The LLVM instruction set consists of several different classifications
3428 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
3429 instructions <binaryops>`, :ref:`bitwise binary
3430 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
3431 :ref:`other instructions <otherops>`.
3435 Terminator Instructions
3436 -----------------------
3438 As mentioned :ref:`previously <functionstructure>`, every basic block in a
3439 program ends with a "Terminator" instruction, which indicates which
3440 block should be executed after the current block is finished. These
3441 terminator instructions typically yield a '``void``' value: they produce
3442 control flow, not values (the one exception being the
3443 ':ref:`invoke <i_invoke>`' instruction).
3445 The terminator instructions are: ':ref:`ret <i_ret>`',
3446 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
3447 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
3448 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
3452 '``ret``' Instruction
3453 ^^^^^^^^^^^^^^^^^^^^^
3460 ret <type> <value> ; Return a value from a non-void function
3461 ret void ; Return from void function
3466 The '``ret``' instruction is used to return control flow (and optionally
3467 a value) from a function back to the caller.
3469 There are two forms of the '``ret``' instruction: one that returns a
3470 value and then causes control flow, and one that just causes control
3476 The '``ret``' instruction optionally accepts a single argument, the
3477 return value. The type of the return value must be a ':ref:`first
3478 class <t_firstclass>`' type.
3480 A function is not :ref:`well formed <wellformed>` if it it has a non-void
3481 return type and contains a '``ret``' instruction with no return value or
3482 a return value with a type that does not match its type, or if it has a
3483 void return type and contains a '``ret``' instruction with a return
3489 When the '``ret``' instruction is executed, control flow returns back to
3490 the calling function's context. If the caller is a
3491 ":ref:`call <i_call>`" instruction, execution continues at the
3492 instruction after the call. If the caller was an
3493 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
3494 beginning of the "normal" destination block. If the instruction returns
3495 a value, that value shall set the call or invoke instruction's return
3501 .. code-block:: llvm
3503 ret i32 5 ; Return an integer value of 5
3504 ret void ; Return from a void function
3505 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
3509 '``br``' Instruction
3510 ^^^^^^^^^^^^^^^^^^^^
3517 br i1 <cond>, label <iftrue>, label <iffalse>
3518 br label <dest> ; Unconditional branch
3523 The '``br``' instruction is used to cause control flow to transfer to a
3524 different basic block in the current function. There are two forms of
3525 this instruction, corresponding to a conditional branch and an
3526 unconditional branch.
3531 The conditional branch form of the '``br``' instruction takes a single
3532 '``i1``' value and two '``label``' values. The unconditional form of the
3533 '``br``' instruction takes a single '``label``' value as a target.
3538 Upon execution of a conditional '``br``' instruction, the '``i1``'
3539 argument is evaluated. If the value is ``true``, control flows to the
3540 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
3541 to the '``iffalse``' ``label`` argument.
3546 .. code-block:: llvm
3549 %cond = icmp eq i32 %a, %b
3550 br i1 %cond, label %IfEqual, label %IfUnequal
3558 '``switch``' Instruction
3559 ^^^^^^^^^^^^^^^^^^^^^^^^
3566 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
3571 The '``switch``' instruction is used to transfer control flow to one of
3572 several different places. It is a generalization of the '``br``'
3573 instruction, allowing a branch to occur to one of many possible
3579 The '``switch``' instruction uses three parameters: an integer
3580 comparison value '``value``', a default '``label``' destination, and an
3581 array of pairs of comparison value constants and '``label``'s. The table
3582 is not allowed to contain duplicate constant entries.
3587 The ``switch`` instruction specifies a table of values and destinations.
3588 When the '``switch``' instruction is executed, this table is searched
3589 for the given value. If the value is found, control flow is transferred
3590 to the corresponding destination; otherwise, control flow is transferred
3591 to the default destination.
3596 Depending on properties of the target machine and the particular
3597 ``switch`` instruction, this instruction may be code generated in
3598 different ways. For example, it could be generated as a series of
3599 chained conditional branches or with a lookup table.
3604 .. code-block:: llvm
3606 ; Emulate a conditional br instruction
3607 %Val = zext i1 %value to i32
3608 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
3610 ; Emulate an unconditional br instruction
3611 switch i32 0, label %dest [ ]
3613 ; Implement a jump table:
3614 switch i32 %val, label %otherwise [ i32 0, label %onzero
3616 i32 2, label %ontwo ]
3620 '``indirectbr``' Instruction
3621 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3628 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
3633 The '``indirectbr``' instruction implements an indirect branch to a
3634 label within the current function, whose address is specified by
3635 "``address``". Address must be derived from a
3636 :ref:`blockaddress <blockaddress>` constant.
3641 The '``address``' argument is the address of the label to jump to. The
3642 rest of the arguments indicate the full set of possible destinations
3643 that the address may point to. Blocks are allowed to occur multiple
3644 times in the destination list, though this isn't particularly useful.
3646 This destination list is required so that dataflow analysis has an
3647 accurate understanding of the CFG.
3652 Control transfers to the block specified in the address argument. All
3653 possible destination blocks must be listed in the label list, otherwise
3654 this instruction has undefined behavior. This implies that jumps to
3655 labels defined in other functions have undefined behavior as well.
3660 This is typically implemented with a jump through a register.
3665 .. code-block:: llvm
3667 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
3671 '``invoke``' Instruction
3672 ^^^^^^^^^^^^^^^^^^^^^^^^
3679 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
3680 to label <normal label> unwind label <exception label>
3685 The '``invoke``' instruction causes control to transfer to a specified
3686 function, with the possibility of control flow transfer to either the
3687 '``normal``' label or the '``exception``' label. If the callee function
3688 returns with the "``ret``" instruction, control flow will return to the
3689 "normal" label. If the callee (or any indirect callees) returns via the
3690 ":ref:`resume <i_resume>`" instruction or other exception handling
3691 mechanism, control is interrupted and continued at the dynamically
3692 nearest "exception" label.
3694 The '``exception``' label is a `landing
3695 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
3696 '``exception``' label is required to have the
3697 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
3698 information about the behavior of the program after unwinding happens,
3699 as its first non-PHI instruction. The restrictions on the
3700 "``landingpad``" instruction's tightly couples it to the "``invoke``"
3701 instruction, so that the important information contained within the
3702 "``landingpad``" instruction can't be lost through normal code motion.
3707 This instruction requires several arguments:
3709 #. The optional "cconv" marker indicates which :ref:`calling
3710 convention <callingconv>` the call should use. If none is
3711 specified, the call defaults to using C calling conventions.
3712 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
3713 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
3715 #. '``ptr to function ty``': shall be the signature of the pointer to
3716 function value being invoked. In most cases, this is a direct
3717 function invocation, but indirect ``invoke``'s are just as possible,
3718 branching off an arbitrary pointer to function value.
3719 #. '``function ptr val``': An LLVM value containing a pointer to a
3720 function to be invoked.
3721 #. '``function args``': argument list whose types match the function
3722 signature argument types and parameter attributes. All arguments must
3723 be of :ref:`first class <t_firstclass>` type. If the function signature
3724 indicates the function accepts a variable number of arguments, the
3725 extra arguments can be specified.
3726 #. '``normal label``': the label reached when the called function
3727 executes a '``ret``' instruction.
3728 #. '``exception label``': the label reached when a callee returns via
3729 the :ref:`resume <i_resume>` instruction or other exception handling
3731 #. The optional :ref:`function attributes <fnattrs>` list. Only
3732 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
3733 attributes are valid here.
3738 This instruction is designed to operate as a standard '``call``'
3739 instruction in most regards. The primary difference is that it
3740 establishes an association with a label, which is used by the runtime
3741 library to unwind the stack.
3743 This instruction is used in languages with destructors to ensure that
3744 proper cleanup is performed in the case of either a ``longjmp`` or a
3745 thrown exception. Additionally, this is important for implementation of
3746 '``catch``' clauses in high-level languages that support them.
3748 For the purposes of the SSA form, the definition of the value returned
3749 by the '``invoke``' instruction is deemed to occur on the edge from the
3750 current block to the "normal" label. If the callee unwinds then no
3751 return value is available.
3756 .. code-block:: llvm
3758 %retval = invoke i32 @Test(i32 15) to label %Continue
3759 unwind label %TestCleanup ; i32:retval set
3760 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
3761 unwind label %TestCleanup ; i32:retval set
3765 '``resume``' Instruction
3766 ^^^^^^^^^^^^^^^^^^^^^^^^
3773 resume <type> <value>
3778 The '``resume``' instruction is a terminator instruction that has no
3784 The '``resume``' instruction requires one argument, which must have the
3785 same type as the result of any '``landingpad``' instruction in the same
3791 The '``resume``' instruction resumes propagation of an existing
3792 (in-flight) exception whose unwinding was interrupted with a
3793 :ref:`landingpad <i_landingpad>` instruction.
3798 .. code-block:: llvm
3800 resume { i8*, i32 } %exn
3804 '``unreachable``' Instruction
3805 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3817 The '``unreachable``' instruction has no defined semantics. This
3818 instruction is used to inform the optimizer that a particular portion of
3819 the code is not reachable. This can be used to indicate that the code
3820 after a no-return function cannot be reached, and other facts.
3825 The '``unreachable``' instruction has no defined semantics.
3832 Binary operators are used to do most of the computation in a program.
3833 They require two operands of the same type, execute an operation on
3834 them, and produce a single value. The operands might represent multiple
3835 data, as is the case with the :ref:`vector <t_vector>` data type. The
3836 result value has the same type as its operands.
3838 There are several different binary operators:
3842 '``add``' Instruction
3843 ^^^^^^^^^^^^^^^^^^^^^
3850 <result> = add <ty> <op1>, <op2> ; yields ty:result
3851 <result> = add nuw <ty> <op1>, <op2> ; yields ty:result
3852 <result> = add nsw <ty> <op1>, <op2> ; yields ty:result
3853 <result> = add nuw nsw <ty> <op1>, <op2> ; yields ty:result
3858 The '``add``' instruction returns the sum of its two operands.
3863 The two arguments to the '``add``' instruction must be
3864 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3865 arguments must have identical types.
3870 The value produced is the integer sum of the two operands.
3872 If the sum has unsigned overflow, the result returned is the
3873 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3876 Because LLVM integers use a two's complement representation, this
3877 instruction is appropriate for both signed and unsigned integers.
3879 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3880 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3881 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
3882 unsigned and/or signed overflow, respectively, occurs.
3887 .. code-block:: llvm
3889 <result> = add i32 4, %var ; yields i32:result = 4 + %var
3893 '``fadd``' Instruction
3894 ^^^^^^^^^^^^^^^^^^^^^^
3901 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
3906 The '``fadd``' instruction returns the sum of its two operands.
3911 The two arguments to the '``fadd``' instruction must be :ref:`floating
3912 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3913 Both arguments must have identical types.
3918 The value produced is the floating point sum of the two operands. This
3919 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
3920 which are optimization hints to enable otherwise unsafe floating point
3926 .. code-block:: llvm
3928 <result> = fadd float 4.0, %var ; yields float:result = 4.0 + %var
3930 '``sub``' Instruction
3931 ^^^^^^^^^^^^^^^^^^^^^
3938 <result> = sub <ty> <op1>, <op2> ; yields ty:result
3939 <result> = sub nuw <ty> <op1>, <op2> ; yields ty:result
3940 <result> = sub nsw <ty> <op1>, <op2> ; yields ty:result
3941 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields ty:result
3946 The '``sub``' instruction returns the difference of its two operands.
3948 Note that the '``sub``' instruction is used to represent the '``neg``'
3949 instruction present in most other intermediate representations.
3954 The two arguments to the '``sub``' instruction must be
3955 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3956 arguments must have identical types.
3961 The value produced is the integer difference of the two operands.
3963 If the difference has unsigned overflow, the result returned is the
3964 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3967 Because LLVM integers use a two's complement representation, this
3968 instruction is appropriate for both signed and unsigned integers.
3970 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3971 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3972 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
3973 unsigned and/or signed overflow, respectively, occurs.
3978 .. code-block:: llvm
3980 <result> = sub i32 4, %var ; yields i32:result = 4 - %var
3981 <result> = sub i32 0, %val ; yields i32:result = -%var
3985 '``fsub``' Instruction
3986 ^^^^^^^^^^^^^^^^^^^^^^
3993 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
3998 The '``fsub``' instruction returns the difference of its two operands.
4000 Note that the '``fsub``' instruction is used to represent the '``fneg``'
4001 instruction present in most other intermediate representations.
4006 The two arguments to the '``fsub``' instruction must be :ref:`floating
4007 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4008 Both arguments must have identical types.
4013 The value produced is the floating point difference of the two operands.
4014 This instruction can also take any number of :ref:`fast-math
4015 flags <fastmath>`, which are optimization hints to enable otherwise
4016 unsafe floating point optimizations:
4021 .. code-block:: llvm
4023 <result> = fsub float 4.0, %var ; yields float:result = 4.0 - %var
4024 <result> = fsub float -0.0, %val ; yields float:result = -%var
4026 '``mul``' Instruction
4027 ^^^^^^^^^^^^^^^^^^^^^
4034 <result> = mul <ty> <op1>, <op2> ; yields ty:result
4035 <result> = mul nuw <ty> <op1>, <op2> ; yields ty:result
4036 <result> = mul nsw <ty> <op1>, <op2> ; yields ty:result
4037 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields ty:result
4042 The '``mul``' instruction returns the product of its two operands.
4047 The two arguments to the '``mul``' instruction must be
4048 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4049 arguments must have identical types.
4054 The value produced is the integer product of the two operands.
4056 If the result of the multiplication has unsigned overflow, the result
4057 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
4058 bit width of the result.
4060 Because LLVM integers use a two's complement representation, and the
4061 result is the same width as the operands, this instruction returns the
4062 correct result for both signed and unsigned integers. If a full product
4063 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
4064 sign-extended or zero-extended as appropriate to the width of the full
4067 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4068 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4069 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
4070 unsigned and/or signed overflow, respectively, occurs.
4075 .. code-block:: llvm
4077 <result> = mul i32 4, %var ; yields i32:result = 4 * %var
4081 '``fmul``' Instruction
4082 ^^^^^^^^^^^^^^^^^^^^^^
4089 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4094 The '``fmul``' instruction returns the product of its two operands.
4099 The two arguments to the '``fmul``' instruction must be :ref:`floating
4100 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4101 Both arguments must have identical types.
4106 The value produced is the floating point product of the two operands.
4107 This instruction can also take any number of :ref:`fast-math
4108 flags <fastmath>`, which are optimization hints to enable otherwise
4109 unsafe floating point optimizations:
4114 .. code-block:: llvm
4116 <result> = fmul float 4.0, %var ; yields float:result = 4.0 * %var
4118 '``udiv``' Instruction
4119 ^^^^^^^^^^^^^^^^^^^^^^
4126 <result> = udiv <ty> <op1>, <op2> ; yields ty:result
4127 <result> = udiv exact <ty> <op1>, <op2> ; yields ty:result
4132 The '``udiv``' instruction returns the quotient of its two operands.
4137 The two arguments to the '``udiv``' instruction must be
4138 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4139 arguments must have identical types.
4144 The value produced is the unsigned integer quotient of the two operands.
4146 Note that unsigned integer division and signed integer division are
4147 distinct operations; for signed integer division, use '``sdiv``'.
4149 Division by zero leads to undefined behavior.
4151 If the ``exact`` keyword is present, the result value of the ``udiv`` is
4152 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
4153 such, "((a udiv exact b) mul b) == a").
4158 .. code-block:: llvm
4160 <result> = udiv i32 4, %var ; yields i32:result = 4 / %var
4162 '``sdiv``' Instruction
4163 ^^^^^^^^^^^^^^^^^^^^^^
4170 <result> = sdiv <ty> <op1>, <op2> ; yields ty:result
4171 <result> = sdiv exact <ty> <op1>, <op2> ; yields ty:result
4176 The '``sdiv``' instruction returns the quotient of its two operands.
4181 The two arguments to the '``sdiv``' instruction must be
4182 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4183 arguments must have identical types.
4188 The value produced is the signed integer quotient of the two operands
4189 rounded towards zero.
4191 Note that signed integer division and unsigned integer division are
4192 distinct operations; for unsigned integer division, use '``udiv``'.
4194 Division by zero leads to undefined behavior. Overflow also leads to
4195 undefined behavior; this is a rare case, but can occur, for example, by
4196 doing a 32-bit division of -2147483648 by -1.
4198 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
4199 a :ref:`poison value <poisonvalues>` if the result would be rounded.
4204 .. code-block:: llvm
4206 <result> = sdiv i32 4, %var ; yields i32:result = 4 / %var
4210 '``fdiv``' Instruction
4211 ^^^^^^^^^^^^^^^^^^^^^^
4218 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4223 The '``fdiv``' instruction returns the quotient of its two operands.
4228 The two arguments to the '``fdiv``' instruction must be :ref:`floating
4229 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4230 Both arguments must have identical types.
4235 The value produced is the floating point quotient of the two operands.
4236 This instruction can also take any number of :ref:`fast-math
4237 flags <fastmath>`, which are optimization hints to enable otherwise
4238 unsafe floating point optimizations:
4243 .. code-block:: llvm
4245 <result> = fdiv float 4.0, %var ; yields float:result = 4.0 / %var
4247 '``urem``' Instruction
4248 ^^^^^^^^^^^^^^^^^^^^^^
4255 <result> = urem <ty> <op1>, <op2> ; yields ty:result
4260 The '``urem``' instruction returns the remainder from the unsigned
4261 division of its two arguments.
4266 The two arguments to the '``urem``' instruction must be
4267 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4268 arguments must have identical types.
4273 This instruction returns the unsigned integer *remainder* of a division.
4274 This instruction always performs an unsigned division to get the
4277 Note that unsigned integer remainder and signed integer remainder are
4278 distinct operations; for signed integer remainder, use '``srem``'.
4280 Taking the remainder of a division by zero leads to undefined behavior.
4285 .. code-block:: llvm
4287 <result> = urem i32 4, %var ; yields i32:result = 4 % %var
4289 '``srem``' Instruction
4290 ^^^^^^^^^^^^^^^^^^^^^^
4297 <result> = srem <ty> <op1>, <op2> ; yields ty:result
4302 The '``srem``' instruction returns the remainder from the signed
4303 division of its two operands. This instruction can also take
4304 :ref:`vector <t_vector>` versions of the values in which case the elements
4310 The two arguments to the '``srem``' instruction must be
4311 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4312 arguments must have identical types.
4317 This instruction returns the *remainder* of a division (where the result
4318 is either zero or has the same sign as the dividend, ``op1``), not the
4319 *modulo* operator (where the result is either zero or has the same sign
4320 as the divisor, ``op2``) of a value. For more information about the
4321 difference, see `The Math
4322 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
4323 table of how this is implemented in various languages, please see
4325 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
4327 Note that signed integer remainder and unsigned integer remainder are
4328 distinct operations; for unsigned integer remainder, use '``urem``'.
4330 Taking the remainder of a division by zero leads to undefined behavior.
4331 Overflow also leads to undefined behavior; this is a rare case, but can
4332 occur, for example, by taking the remainder of a 32-bit division of
4333 -2147483648 by -1. (The remainder doesn't actually overflow, but this
4334 rule lets srem be implemented using instructions that return both the
4335 result of the division and the remainder.)
4340 .. code-block:: llvm
4342 <result> = srem i32 4, %var ; yields i32:result = 4 % %var
4346 '``frem``' Instruction
4347 ^^^^^^^^^^^^^^^^^^^^^^
4354 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4359 The '``frem``' instruction returns the remainder from the division of
4365 The two arguments to the '``frem``' instruction must be :ref:`floating
4366 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4367 Both arguments must have identical types.
4372 This instruction returns the *remainder* of a division. The remainder
4373 has the same sign as the dividend. This instruction can also take any
4374 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
4375 to enable otherwise unsafe floating point optimizations:
4380 .. code-block:: llvm
4382 <result> = frem float 4.0, %var ; yields float:result = 4.0 % %var
4386 Bitwise Binary Operations
4387 -------------------------
4389 Bitwise binary operators are used to do various forms of bit-twiddling
4390 in a program. They are generally very efficient instructions and can
4391 commonly be strength reduced from other instructions. They require two
4392 operands of the same type, execute an operation on them, and produce a
4393 single value. The resulting value is the same type as its operands.
4395 '``shl``' Instruction
4396 ^^^^^^^^^^^^^^^^^^^^^
4403 <result> = shl <ty> <op1>, <op2> ; yields ty:result
4404 <result> = shl nuw <ty> <op1>, <op2> ; yields ty:result
4405 <result> = shl nsw <ty> <op1>, <op2> ; yields ty:result
4406 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields ty:result
4411 The '``shl``' instruction returns the first operand shifted to the left
4412 a specified number of bits.
4417 Both arguments to the '``shl``' instruction must be the same
4418 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4419 '``op2``' is treated as an unsigned value.
4424 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
4425 where ``n`` is the width of the result. If ``op2`` is (statically or
4426 dynamically) negative or equal to or larger than the number of bits in
4427 ``op1``, the result is undefined. If the arguments are vectors, each
4428 vector element of ``op1`` is shifted by the corresponding shift amount
4431 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
4432 value <poisonvalues>` if it shifts out any non-zero bits. If the
4433 ``nsw`` keyword is present, then the shift produces a :ref:`poison
4434 value <poisonvalues>` if it shifts out any bits that disagree with the
4435 resultant sign bit. As such, NUW/NSW have the same semantics as they
4436 would if the shift were expressed as a mul instruction with the same
4437 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
4442 .. code-block:: llvm
4444 <result> = shl i32 4, %var ; yields i32: 4 << %var
4445 <result> = shl i32 4, 2 ; yields i32: 16
4446 <result> = shl i32 1, 10 ; yields i32: 1024
4447 <result> = shl i32 1, 32 ; undefined
4448 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
4450 '``lshr``' Instruction
4451 ^^^^^^^^^^^^^^^^^^^^^^
4458 <result> = lshr <ty> <op1>, <op2> ; yields ty:result
4459 <result> = lshr exact <ty> <op1>, <op2> ; yields ty:result
4464 The '``lshr``' instruction (logical shift right) returns the first
4465 operand shifted to the right a specified number of bits with zero fill.
4470 Both arguments to the '``lshr``' instruction must be the same
4471 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4472 '``op2``' is treated as an unsigned value.
4477 This instruction always performs a logical shift right operation. The
4478 most significant bits of the result will be filled with zero bits after
4479 the shift. If ``op2`` is (statically or dynamically) equal to or larger
4480 than the number of bits in ``op1``, the result is undefined. If the
4481 arguments are vectors, each vector element of ``op1`` is shifted by the
4482 corresponding shift amount in ``op2``.
4484 If the ``exact`` keyword is present, the result value of the ``lshr`` is
4485 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4491 .. code-block:: llvm
4493 <result> = lshr i32 4, 1 ; yields i32:result = 2
4494 <result> = lshr i32 4, 2 ; yields i32:result = 1
4495 <result> = lshr i8 4, 3 ; yields i8:result = 0
4496 <result> = lshr i8 -2, 1 ; yields i8:result = 0x7F
4497 <result> = lshr i32 1, 32 ; undefined
4498 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
4500 '``ashr``' Instruction
4501 ^^^^^^^^^^^^^^^^^^^^^^
4508 <result> = ashr <ty> <op1>, <op2> ; yields ty:result
4509 <result> = ashr exact <ty> <op1>, <op2> ; yields ty:result
4514 The '``ashr``' instruction (arithmetic shift right) returns the first
4515 operand shifted to the right a specified number of bits with sign
4521 Both arguments to the '``ashr``' instruction must be the same
4522 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4523 '``op2``' is treated as an unsigned value.
4528 This instruction always performs an arithmetic shift right operation,
4529 The most significant bits of the result will be filled with the sign bit
4530 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
4531 than the number of bits in ``op1``, the result is undefined. If the
4532 arguments are vectors, each vector element of ``op1`` is shifted by the
4533 corresponding shift amount in ``op2``.
4535 If the ``exact`` keyword is present, the result value of the ``ashr`` is
4536 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4542 .. code-block:: llvm
4544 <result> = ashr i32 4, 1 ; yields i32:result = 2
4545 <result> = ashr i32 4, 2 ; yields i32:result = 1
4546 <result> = ashr i8 4, 3 ; yields i8:result = 0
4547 <result> = ashr i8 -2, 1 ; yields i8:result = -1
4548 <result> = ashr i32 1, 32 ; undefined
4549 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
4551 '``and``' Instruction
4552 ^^^^^^^^^^^^^^^^^^^^^
4559 <result> = and <ty> <op1>, <op2> ; yields ty:result
4564 The '``and``' instruction returns the bitwise logical and of its two
4570 The two arguments to the '``and``' instruction must be
4571 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4572 arguments must have identical types.
4577 The truth table used for the '``and``' instruction is:
4594 .. code-block:: llvm
4596 <result> = and i32 4, %var ; yields i32:result = 4 & %var
4597 <result> = and i32 15, 40 ; yields i32:result = 8
4598 <result> = and i32 4, 8 ; yields i32:result = 0
4600 '``or``' Instruction
4601 ^^^^^^^^^^^^^^^^^^^^
4608 <result> = or <ty> <op1>, <op2> ; yields ty:result
4613 The '``or``' instruction returns the bitwise logical inclusive or of its
4619 The two arguments to the '``or``' instruction must be
4620 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4621 arguments must have identical types.
4626 The truth table used for the '``or``' instruction is:
4645 <result> = or i32 4, %var ; yields i32:result = 4 | %var
4646 <result> = or i32 15, 40 ; yields i32:result = 47
4647 <result> = or i32 4, 8 ; yields i32:result = 12
4649 '``xor``' Instruction
4650 ^^^^^^^^^^^^^^^^^^^^^
4657 <result> = xor <ty> <op1>, <op2> ; yields ty:result
4662 The '``xor``' instruction returns the bitwise logical exclusive or of
4663 its two operands. The ``xor`` is used to implement the "one's
4664 complement" operation, which is the "~" operator in C.
4669 The two arguments to the '``xor``' instruction must be
4670 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4671 arguments must have identical types.
4676 The truth table used for the '``xor``' instruction is:
4693 .. code-block:: llvm
4695 <result> = xor i32 4, %var ; yields i32:result = 4 ^ %var
4696 <result> = xor i32 15, 40 ; yields i32:result = 39
4697 <result> = xor i32 4, 8 ; yields i32:result = 12
4698 <result> = xor i32 %V, -1 ; yields i32:result = ~%V
4703 LLVM supports several instructions to represent vector operations in a
4704 target-independent manner. These instructions cover the element-access
4705 and vector-specific operations needed to process vectors effectively.
4706 While LLVM does directly support these vector operations, many
4707 sophisticated algorithms will want to use target-specific intrinsics to
4708 take full advantage of a specific target.
4710 .. _i_extractelement:
4712 '``extractelement``' Instruction
4713 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4720 <result> = extractelement <n x <ty>> <val>, <ty2> <idx> ; yields <ty>
4725 The '``extractelement``' instruction extracts a single scalar element
4726 from a vector at a specified index.
4731 The first operand of an '``extractelement``' instruction is a value of
4732 :ref:`vector <t_vector>` type. The second operand is an index indicating
4733 the position from which to extract the element. The index may be a
4734 variable of any integer type.
4739 The result is a scalar of the same type as the element type of ``val``.
4740 Its value is the value at position ``idx`` of ``val``. If ``idx``
4741 exceeds the length of ``val``, the results are undefined.
4746 .. code-block:: llvm
4748 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
4750 .. _i_insertelement:
4752 '``insertelement``' Instruction
4753 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4760 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, <ty2> <idx> ; yields <n x <ty>>
4765 The '``insertelement``' instruction inserts a scalar element into a
4766 vector at a specified index.
4771 The first operand of an '``insertelement``' instruction is a value of
4772 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
4773 type must equal the element type of the first operand. The third operand
4774 is an index indicating the position at which to insert the value. The
4775 index may be a variable of any integer type.
4780 The result is a vector of the same type as ``val``. Its element values
4781 are those of ``val`` except at position ``idx``, where it gets the value
4782 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
4788 .. code-block:: llvm
4790 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
4792 .. _i_shufflevector:
4794 '``shufflevector``' Instruction
4795 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4802 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
4807 The '``shufflevector``' instruction constructs a permutation of elements
4808 from two input vectors, returning a vector with the same element type as
4809 the input and length that is the same as the shuffle mask.
4814 The first two operands of a '``shufflevector``' instruction are vectors
4815 with the same type. The third argument is a shuffle mask whose element
4816 type is always 'i32'. The result of the instruction is a vector whose
4817 length is the same as the shuffle mask and whose element type is the
4818 same as the element type of the first two operands.
4820 The shuffle mask operand is required to be a constant vector with either
4821 constant integer or undef values.
4826 The elements of the two input vectors are numbered from left to right
4827 across both of the vectors. The shuffle mask operand specifies, for each
4828 element of the result vector, which element of the two input vectors the
4829 result element gets. The element selector may be undef (meaning "don't
4830 care") and the second operand may be undef if performing a shuffle from
4836 .. code-block:: llvm
4838 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4839 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
4840 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
4841 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
4842 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
4843 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
4844 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4845 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
4847 Aggregate Operations
4848 --------------------
4850 LLVM supports several instructions for working with
4851 :ref:`aggregate <t_aggregate>` values.
4855 '``extractvalue``' Instruction
4856 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4863 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
4868 The '``extractvalue``' instruction extracts the value of a member field
4869 from an :ref:`aggregate <t_aggregate>` value.
4874 The first operand of an '``extractvalue``' instruction is a value of
4875 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
4876 constant indices to specify which value to extract in a similar manner
4877 as indices in a '``getelementptr``' instruction.
4879 The major differences to ``getelementptr`` indexing are:
4881 - Since the value being indexed is not a pointer, the first index is
4882 omitted and assumed to be zero.
4883 - At least one index must be specified.
4884 - Not only struct indices but also array indices must be in bounds.
4889 The result is the value at the position in the aggregate specified by
4895 .. code-block:: llvm
4897 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
4901 '``insertvalue``' Instruction
4902 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4909 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
4914 The '``insertvalue``' instruction inserts a value into a member field in
4915 an :ref:`aggregate <t_aggregate>` value.
4920 The first operand of an '``insertvalue``' instruction is a value of
4921 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
4922 a first-class value to insert. The following operands are constant
4923 indices indicating the position at which to insert the value in a
4924 similar manner as indices in a '``extractvalue``' instruction. The value
4925 to insert must have the same type as the value identified by the
4931 The result is an aggregate of the same type as ``val``. Its value is
4932 that of ``val`` except that the value at the position specified by the
4933 indices is that of ``elt``.
4938 .. code-block:: llvm
4940 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
4941 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
4942 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 ; yields {i32 1, float %val}
4946 Memory Access and Addressing Operations
4947 ---------------------------------------
4949 A key design point of an SSA-based representation is how it represents
4950 memory. In LLVM, no memory locations are in SSA form, which makes things
4951 very simple. This section describes how to read, write, and allocate
4956 '``alloca``' Instruction
4957 ^^^^^^^^^^^^^^^^^^^^^^^^
4964 <result> = alloca [inalloca] <type> [, <ty> <NumElements>] [, align <alignment>] ; yields type*:result
4969 The '``alloca``' instruction allocates memory on the stack frame of the
4970 currently executing function, to be automatically released when this
4971 function returns to its caller. The object is always allocated in the
4972 generic address space (address space zero).
4977 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
4978 bytes of memory on the runtime stack, returning a pointer of the
4979 appropriate type to the program. If "NumElements" is specified, it is
4980 the number of elements allocated, otherwise "NumElements" is defaulted
4981 to be one. If a constant alignment is specified, the value result of the
4982 allocation is guaranteed to be aligned to at least that boundary. The
4983 alignment may not be greater than ``1 << 29``. If not specified, or if
4984 zero, the target can choose to align the allocation on any convenient
4985 boundary compatible with the type.
4987 '``type``' may be any sized type.
4992 Memory is allocated; a pointer is returned. The operation is undefined
4993 if there is insufficient stack space for the allocation. '``alloca``'d
4994 memory is automatically released when the function returns. The
4995 '``alloca``' instruction is commonly used to represent automatic
4996 variables that must have an address available. When the function returns
4997 (either with the ``ret`` or ``resume`` instructions), the memory is
4998 reclaimed. Allocating zero bytes is legal, but the result is undefined.
4999 The order in which memory is allocated (ie., which way the stack grows)
5005 .. code-block:: llvm
5007 %ptr = alloca i32 ; yields i32*:ptr
5008 %ptr = alloca i32, i32 4 ; yields i32*:ptr
5009 %ptr = alloca i32, i32 4, align 1024 ; yields i32*:ptr
5010 %ptr = alloca i32, align 1024 ; yields i32*:ptr
5014 '``load``' Instruction
5015 ^^^^^^^^^^^^^^^^^^^^^^
5022 <result> = load [volatile] <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>]
5023 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
5024 !<index> = !{ i32 1 }
5029 The '``load``' instruction is used to read from memory.
5034 The argument to the ``load`` instruction specifies the memory address
5035 from which to load. The pointer must point to a :ref:`first
5036 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
5037 then the optimizer is not allowed to modify the number or order of
5038 execution of this ``load`` with other :ref:`volatile
5039 operations <volatile>`.
5041 If the ``load`` is marked as ``atomic``, it takes an extra
5042 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5043 ``release`` and ``acq_rel`` orderings are not valid on ``load``
5044 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5045 when they may see multiple atomic stores. The type of the pointee must
5046 be an integer type whose bit width is a power of two greater than or
5047 equal to eight and less than or equal to a target-specific size limit.
5048 ``align`` must be explicitly specified on atomic loads, and the load has
5049 undefined behavior if the alignment is not set to a value which is at
5050 least the size in bytes of the pointee. ``!nontemporal`` does not have
5051 any defined semantics for atomic loads.
5053 The optional constant ``align`` argument specifies the alignment of the
5054 operation (that is, the alignment of the memory address). A value of 0
5055 or an omitted ``align`` argument means that the operation has the ABI
5056 alignment for the target. It is the responsibility of the code emitter
5057 to ensure that the alignment information is correct. Overestimating the
5058 alignment results in undefined behavior. Underestimating the alignment
5059 may produce less efficient code. An alignment of 1 is always safe. The
5060 maximum possible alignment is ``1 << 29``.
5062 The optional ``!nontemporal`` metadata must reference a single
5063 metadata name ``<index>`` corresponding to a metadata node with one
5064 ``i32`` entry of value 1. The existence of the ``!nontemporal``
5065 metadata on the instruction tells the optimizer and code generator
5066 that this load is not expected to be reused in the cache. The code
5067 generator may select special instructions to save cache bandwidth, such
5068 as the ``MOVNT`` instruction on x86.
5070 The optional ``!invariant.load`` metadata must reference a single
5071 metadata name ``<index>`` corresponding to a metadata node with no
5072 entries. The existence of the ``!invariant.load`` metadata on the
5073 instruction tells the optimizer and code generator that this load
5074 address points to memory which does not change value during program
5075 execution. The optimizer may then move this load around, for example, by
5076 hoisting it out of loops using loop invariant code motion.
5081 The location of memory pointed to is loaded. If the value being loaded
5082 is of scalar type then the number of bytes read does not exceed the
5083 minimum number of bytes needed to hold all bits of the type. For
5084 example, loading an ``i24`` reads at most three bytes. When loading a
5085 value of a type like ``i20`` with a size that is not an integral number
5086 of bytes, the result is undefined if the value was not originally
5087 written using a store of the same type.
5092 .. code-block:: llvm
5094 %ptr = alloca i32 ; yields i32*:ptr
5095 store i32 3, i32* %ptr ; yields void
5096 %val = load i32* %ptr ; yields i32:val = i32 3
5100 '``store``' Instruction
5101 ^^^^^^^^^^^^^^^^^^^^^^^
5108 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields void
5109 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields void
5114 The '``store``' instruction is used to write to memory.
5119 There are two arguments to the ``store`` instruction: a value to store
5120 and an address at which to store it. The type of the ``<pointer>``
5121 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
5122 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
5123 then the optimizer is not allowed to modify the number or order of
5124 execution of this ``store`` with other :ref:`volatile
5125 operations <volatile>`.
5127 If the ``store`` is marked as ``atomic``, it takes an extra
5128 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5129 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
5130 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5131 when they may see multiple atomic stores. The type of the pointee must
5132 be an integer type whose bit width is a power of two greater than or
5133 equal to eight and less than or equal to a target-specific size limit.
5134 ``align`` must be explicitly specified on atomic stores, and the store
5135 has undefined behavior if the alignment is not set to a value which is
5136 at least the size in bytes of the pointee. ``!nontemporal`` does not
5137 have any defined semantics for atomic stores.
5139 The optional constant ``align`` argument specifies the alignment of the
5140 operation (that is, the alignment of the memory address). A value of 0
5141 or an omitted ``align`` argument means that the operation has the ABI
5142 alignment for the target. It is the responsibility of the code emitter
5143 to ensure that the alignment information is correct. Overestimating the
5144 alignment results in undefined behavior. Underestimating the
5145 alignment may produce less efficient code. An alignment of 1 is always
5146 safe. The maximum possible alignment is ``1 << 29``.
5148 The optional ``!nontemporal`` metadata must reference a single metadata
5149 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
5150 value 1. The existence of the ``!nontemporal`` metadata on the instruction
5151 tells the optimizer and code generator that this load is not expected to
5152 be reused in the cache. The code generator may select special
5153 instructions to save cache bandwidth, such as the MOVNT instruction on
5159 The contents of memory are updated to contain ``<value>`` at the
5160 location specified by the ``<pointer>`` operand. If ``<value>`` is
5161 of scalar type then the number of bytes written does not exceed the
5162 minimum number of bytes needed to hold all bits of the type. For
5163 example, storing an ``i24`` writes at most three bytes. When writing a
5164 value of a type like ``i20`` with a size that is not an integral number
5165 of bytes, it is unspecified what happens to the extra bits that do not
5166 belong to the type, but they will typically be overwritten.
5171 .. code-block:: llvm
5173 %ptr = alloca i32 ; yields i32*:ptr
5174 store i32 3, i32* %ptr ; yields void
5175 %val = load i32* %ptr ; yields i32:val = i32 3
5179 '``fence``' Instruction
5180 ^^^^^^^^^^^^^^^^^^^^^^^
5187 fence [singlethread] <ordering> ; yields void
5192 The '``fence``' instruction is used to introduce happens-before edges
5198 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
5199 defines what *synchronizes-with* edges they add. They can only be given
5200 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
5205 A fence A which has (at least) ``release`` ordering semantics
5206 *synchronizes with* a fence B with (at least) ``acquire`` ordering
5207 semantics if and only if there exist atomic operations X and Y, both
5208 operating on some atomic object M, such that A is sequenced before X, X
5209 modifies M (either directly or through some side effect of a sequence
5210 headed by X), Y is sequenced before B, and Y observes M. This provides a
5211 *happens-before* dependency between A and B. Rather than an explicit
5212 ``fence``, one (but not both) of the atomic operations X or Y might
5213 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
5214 still *synchronize-with* the explicit ``fence`` and establish the
5215 *happens-before* edge.
5217 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
5218 ``acquire`` and ``release`` semantics specified above, participates in
5219 the global program order of other ``seq_cst`` operations and/or fences.
5221 The optional ":ref:`singlethread <singlethread>`" argument specifies
5222 that the fence only synchronizes with other fences in the same thread.
5223 (This is useful for interacting with signal handlers.)
5228 .. code-block:: llvm
5230 fence acquire ; yields void
5231 fence singlethread seq_cst ; yields void
5235 '``cmpxchg``' Instruction
5236 ^^^^^^^^^^^^^^^^^^^^^^^^^
5243 cmpxchg [weak] [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <success ordering> <failure ordering> ; yields { ty, i1 }
5248 The '``cmpxchg``' instruction is used to atomically modify memory. It
5249 loads a value in memory and compares it to a given value. If they are
5250 equal, it tries to store a new value into the memory.
5255 There are three arguments to the '``cmpxchg``' instruction: an address
5256 to operate on, a value to compare to the value currently be at that
5257 address, and a new value to place at that address if the compared values
5258 are equal. The type of '<cmp>' must be an integer type whose bit width
5259 is a power of two greater than or equal to eight and less than or equal
5260 to a target-specific size limit. '<cmp>' and '<new>' must have the same
5261 type, and the type of '<pointer>' must be a pointer to that type. If the
5262 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
5263 to modify the number or order of execution of this ``cmpxchg`` with
5264 other :ref:`volatile operations <volatile>`.
5266 The success and failure :ref:`ordering <ordering>` arguments specify how this
5267 ``cmpxchg`` synchronizes with other atomic operations. Both ordering parameters
5268 must be at least ``monotonic``, the ordering constraint on failure must be no
5269 stronger than that on success, and the failure ordering cannot be either
5270 ``release`` or ``acq_rel``.
5272 The optional "``singlethread``" argument declares that the ``cmpxchg``
5273 is only atomic with respect to code (usually signal handlers) running in
5274 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
5275 respect to all other code in the system.
5277 The pointer passed into cmpxchg must have alignment greater than or
5278 equal to the size in memory of the operand.
5283 The contents of memory at the location specified by the '``<pointer>``' operand
5284 is read and compared to '``<cmp>``'; if the read value is the equal, the
5285 '``<new>``' is written. The original value at the location is returned, together
5286 with a flag indicating success (true) or failure (false).
5288 If the cmpxchg operation is marked as ``weak`` then a spurious failure is
5289 permitted: the operation may not write ``<new>`` even if the comparison
5292 If the cmpxchg operation is strong (the default), the i1 value is 1 if and only
5293 if the value loaded equals ``cmp``.
5295 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose of
5296 identifying release sequences. A failed ``cmpxchg`` is equivalent to an atomic
5297 load with an ordering parameter determined the second ordering parameter.
5302 .. code-block:: llvm
5305 %orig = atomic load i32* %ptr unordered ; yields i32
5309 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
5310 %squared = mul i32 %cmp, %cmp
5311 %val_success = cmpxchg i32* %ptr, i32 %cmp, i32 %squared acq_rel monotonic ; yields { i32, i1 }
5312 %value_loaded = extractvalue { i32, i1 } %val_success, 0
5313 %success = extractvalue { i32, i1 } %val_success, 1
5314 br i1 %success, label %done, label %loop
5321 '``atomicrmw``' Instruction
5322 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
5329 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields ty
5334 The '``atomicrmw``' instruction is used to atomically modify memory.
5339 There are three arguments to the '``atomicrmw``' instruction: an
5340 operation to apply, an address whose value to modify, an argument to the
5341 operation. The operation must be one of the following keywords:
5355 The type of '<value>' must be an integer type whose bit width is a power
5356 of two greater than or equal to eight and less than or equal to a
5357 target-specific size limit. The type of the '``<pointer>``' operand must
5358 be a pointer to that type. If the ``atomicrmw`` is marked as
5359 ``volatile``, then the optimizer is not allowed to modify the number or
5360 order of execution of this ``atomicrmw`` with other :ref:`volatile
5361 operations <volatile>`.
5366 The contents of memory at the location specified by the '``<pointer>``'
5367 operand are atomically read, modified, and written back. The original
5368 value at the location is returned. The modification is specified by the
5371 - xchg: ``*ptr = val``
5372 - add: ``*ptr = *ptr + val``
5373 - sub: ``*ptr = *ptr - val``
5374 - and: ``*ptr = *ptr & val``
5375 - nand: ``*ptr = ~(*ptr & val)``
5376 - or: ``*ptr = *ptr | val``
5377 - xor: ``*ptr = *ptr ^ val``
5378 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
5379 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
5380 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
5382 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
5388 .. code-block:: llvm
5390 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields i32
5392 .. _i_getelementptr:
5394 '``getelementptr``' Instruction
5395 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5402 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
5403 <result> = getelementptr inbounds <pty>* <ptrval>{, <ty> <idx>}*
5404 <result> = getelementptr <ptr vector> ptrval, <vector index type> idx
5409 The '``getelementptr``' instruction is used to get the address of a
5410 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
5411 address calculation only and does not access memory.
5416 The first argument is always a pointer or a vector of pointers, and
5417 forms the basis of the calculation. The remaining arguments are indices
5418 that indicate which of the elements of the aggregate object are indexed.
5419 The interpretation of each index is dependent on the type being indexed
5420 into. The first index always indexes the pointer value given as the
5421 first argument, the second index indexes a value of the type pointed to
5422 (not necessarily the value directly pointed to, since the first index
5423 can be non-zero), etc. The first type indexed into must be a pointer
5424 value, subsequent types can be arrays, vectors, and structs. Note that
5425 subsequent types being indexed into can never be pointers, since that
5426 would require loading the pointer before continuing calculation.
5428 The type of each index argument depends on the type it is indexing into.
5429 When indexing into a (optionally packed) structure, only ``i32`` integer
5430 **constants** are allowed (when using a vector of indices they must all
5431 be the **same** ``i32`` integer constant). When indexing into an array,
5432 pointer or vector, integers of any width are allowed, and they are not
5433 required to be constant. These integers are treated as signed values
5436 For example, let's consider a C code fragment and how it gets compiled
5452 int *foo(struct ST *s) {
5453 return &s[1].Z.B[5][13];
5456 The LLVM code generated by Clang is:
5458 .. code-block:: llvm
5460 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
5461 %struct.ST = type { i32, double, %struct.RT }
5463 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
5465 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5472 In the example above, the first index is indexing into the
5473 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
5474 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
5475 indexes into the third element of the structure, yielding a
5476 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
5477 structure. The third index indexes into the second element of the
5478 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
5479 dimensions of the array are subscripted into, yielding an '``i32``'
5480 type. The '``getelementptr``' instruction returns a pointer to this
5481 element, thus computing a value of '``i32*``' type.
5483 Note that it is perfectly legal to index partially through a structure,
5484 returning a pointer to an inner element. Because of this, the LLVM code
5485 for the given testcase is equivalent to:
5487 .. code-block:: llvm
5489 define i32* @foo(%struct.ST* %s) {
5490 %t1 = getelementptr %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
5491 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
5492 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
5493 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
5494 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
5498 If the ``inbounds`` keyword is present, the result value of the
5499 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
5500 pointer is not an *in bounds* address of an allocated object, or if any
5501 of the addresses that would be formed by successive addition of the
5502 offsets implied by the indices to the base address with infinitely
5503 precise signed arithmetic are not an *in bounds* address of that
5504 allocated object. The *in bounds* addresses for an allocated object are
5505 all the addresses that point into the object, plus the address one byte
5506 past the end. In cases where the base is a vector of pointers the
5507 ``inbounds`` keyword applies to each of the computations element-wise.
5509 If the ``inbounds`` keyword is not present, the offsets are added to the
5510 base address with silently-wrapping two's complement arithmetic. If the
5511 offsets have a different width from the pointer, they are sign-extended
5512 or truncated to the width of the pointer. The result value of the
5513 ``getelementptr`` may be outside the object pointed to by the base
5514 pointer. The result value may not necessarily be used to access memory
5515 though, even if it happens to point into allocated storage. See the
5516 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
5519 The getelementptr instruction is often confusing. For some more insight
5520 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
5525 .. code-block:: llvm
5527 ; yields [12 x i8]*:aptr
5528 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5530 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
5532 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
5534 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
5536 In cases where the pointer argument is a vector of pointers, each index
5537 must be a vector with the same number of elements. For example:
5539 .. code-block:: llvm
5541 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5543 Conversion Operations
5544 ---------------------
5546 The instructions in this category are the conversion instructions
5547 (casting) which all take a single operand and a type. They perform
5548 various bit conversions on the operand.
5550 '``trunc .. to``' Instruction
5551 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5558 <result> = trunc <ty> <value> to <ty2> ; yields ty2
5563 The '``trunc``' instruction truncates its operand to the type ``ty2``.
5568 The '``trunc``' instruction takes a value to trunc, and a type to trunc
5569 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
5570 of the same number of integers. The bit size of the ``value`` must be
5571 larger than the bit size of the destination type, ``ty2``. Equal sized
5572 types are not allowed.
5577 The '``trunc``' instruction truncates the high order bits in ``value``
5578 and converts the remaining bits to ``ty2``. Since the source size must
5579 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
5580 It will always truncate bits.
5585 .. code-block:: llvm
5587 %X = trunc i32 257 to i8 ; yields i8:1
5588 %Y = trunc i32 123 to i1 ; yields i1:true
5589 %Z = trunc i32 122 to i1 ; yields i1:false
5590 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
5592 '``zext .. to``' Instruction
5593 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5600 <result> = zext <ty> <value> to <ty2> ; yields ty2
5605 The '``zext``' instruction zero extends its operand to type ``ty2``.
5610 The '``zext``' instruction takes a value to cast, and a type to cast it
5611 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5612 the same number of integers. The bit size of the ``value`` must be
5613 smaller than the bit size of the destination type, ``ty2``.
5618 The ``zext`` fills the high order bits of the ``value`` with zero bits
5619 until it reaches the size of the destination type, ``ty2``.
5621 When zero extending from i1, the result will always be either 0 or 1.
5626 .. code-block:: llvm
5628 %X = zext i32 257 to i64 ; yields i64:257
5629 %Y = zext i1 true to i32 ; yields i32:1
5630 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5632 '``sext .. to``' Instruction
5633 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5640 <result> = sext <ty> <value> to <ty2> ; yields ty2
5645 The '``sext``' sign extends ``value`` to the type ``ty2``.
5650 The '``sext``' instruction takes a value to cast, and a type to cast it
5651 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5652 the same number of integers. The bit size of the ``value`` must be
5653 smaller than the bit size of the destination type, ``ty2``.
5658 The '``sext``' instruction performs a sign extension by copying the sign
5659 bit (highest order bit) of the ``value`` until it reaches the bit size
5660 of the type ``ty2``.
5662 When sign extending from i1, the extension always results in -1 or 0.
5667 .. code-block:: llvm
5669 %X = sext i8 -1 to i16 ; yields i16 :65535
5670 %Y = sext i1 true to i32 ; yields i32:-1
5671 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5673 '``fptrunc .. to``' Instruction
5674 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5681 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
5686 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
5691 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
5692 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
5693 The size of ``value`` must be larger than the size of ``ty2``. This
5694 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
5699 The '``fptrunc``' instruction truncates a ``value`` from a larger
5700 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
5701 point <t_floating>` type. If the value cannot fit within the
5702 destination type, ``ty2``, then the results are undefined.
5707 .. code-block:: llvm
5709 %X = fptrunc double 123.0 to float ; yields float:123.0
5710 %Y = fptrunc double 1.0E+300 to float ; yields undefined
5712 '``fpext .. to``' Instruction
5713 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5720 <result> = fpext <ty> <value> to <ty2> ; yields ty2
5725 The '``fpext``' extends a floating point ``value`` to a larger floating
5731 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
5732 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
5733 to. The source type must be smaller than the destination type.
5738 The '``fpext``' instruction extends the ``value`` from a smaller
5739 :ref:`floating point <t_floating>` type to a larger :ref:`floating
5740 point <t_floating>` type. The ``fpext`` cannot be used to make a
5741 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
5742 *no-op cast* for a floating point cast.
5747 .. code-block:: llvm
5749 %X = fpext float 3.125 to double ; yields double:3.125000e+00
5750 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
5752 '``fptoui .. to``' Instruction
5753 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5760 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
5765 The '``fptoui``' converts a floating point ``value`` to its unsigned
5766 integer equivalent of type ``ty2``.
5771 The '``fptoui``' instruction takes a value to cast, which must be a
5772 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5773 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5774 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5775 type with the same number of elements as ``ty``
5780 The '``fptoui``' instruction converts its :ref:`floating
5781 point <t_floating>` operand into the nearest (rounding towards zero)
5782 unsigned integer value. If the value cannot fit in ``ty2``, the results
5788 .. code-block:: llvm
5790 %X = fptoui double 123.0 to i32 ; yields i32:123
5791 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
5792 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
5794 '``fptosi .. to``' Instruction
5795 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5802 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
5807 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
5808 ``value`` to type ``ty2``.
5813 The '``fptosi``' instruction takes a value to cast, which must be a
5814 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5815 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5816 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5817 type with the same number of elements as ``ty``
5822 The '``fptosi``' instruction converts its :ref:`floating
5823 point <t_floating>` operand into the nearest (rounding towards zero)
5824 signed integer value. If the value cannot fit in ``ty2``, the results
5830 .. code-block:: llvm
5832 %X = fptosi double -123.0 to i32 ; yields i32:-123
5833 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
5834 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
5836 '``uitofp .. to``' Instruction
5837 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5844 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
5849 The '``uitofp``' instruction regards ``value`` as an unsigned integer
5850 and converts that value to the ``ty2`` type.
5855 The '``uitofp``' instruction takes a value to cast, which must be a
5856 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5857 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5858 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5859 type with the same number of elements as ``ty``
5864 The '``uitofp``' instruction interprets its operand as an unsigned
5865 integer quantity and converts it to the corresponding floating point
5866 value. If the value cannot fit in the floating point value, the results
5872 .. code-block:: llvm
5874 %X = uitofp i32 257 to float ; yields float:257.0
5875 %Y = uitofp i8 -1 to double ; yields double:255.0
5877 '``sitofp .. to``' Instruction
5878 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5885 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
5890 The '``sitofp``' instruction regards ``value`` as a signed integer and
5891 converts that value to the ``ty2`` type.
5896 The '``sitofp``' instruction takes a value to cast, which must be a
5897 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5898 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5899 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5900 type with the same number of elements as ``ty``
5905 The '``sitofp``' instruction interprets its operand as a signed integer
5906 quantity and converts it to the corresponding floating point value. If
5907 the value cannot fit in the floating point value, the results are
5913 .. code-block:: llvm
5915 %X = sitofp i32 257 to float ; yields float:257.0
5916 %Y = sitofp i8 -1 to double ; yields double:-1.0
5920 '``ptrtoint .. to``' Instruction
5921 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5928 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
5933 The '``ptrtoint``' instruction converts the pointer or a vector of
5934 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
5939 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
5940 a a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
5941 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
5942 a vector of integers type.
5947 The '``ptrtoint``' instruction converts ``value`` to integer type
5948 ``ty2`` by interpreting the pointer value as an integer and either
5949 truncating or zero extending that value to the size of the integer type.
5950 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
5951 ``value`` is larger than ``ty2`` then a truncation is done. If they are
5952 the same size, then nothing is done (*no-op cast*) other than a type
5958 .. code-block:: llvm
5960 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
5961 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
5962 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
5966 '``inttoptr .. to``' Instruction
5967 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5974 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
5979 The '``inttoptr``' instruction converts an integer ``value`` to a
5980 pointer type, ``ty2``.
5985 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
5986 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
5992 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
5993 applying either a zero extension or a truncation depending on the size
5994 of the integer ``value``. If ``value`` is larger than the size of a
5995 pointer then a truncation is done. If ``value`` is smaller than the size
5996 of a pointer then a zero extension is done. If they are the same size,
5997 nothing is done (*no-op cast*).
6002 .. code-block:: llvm
6004 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
6005 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
6006 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
6007 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
6011 '``bitcast .. to``' Instruction
6012 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6019 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
6024 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
6030 The '``bitcast``' instruction takes a value to cast, which must be a
6031 non-aggregate first class value, and a type to cast it to, which must
6032 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
6033 bit sizes of ``value`` and the destination type, ``ty2``, must be
6034 identical. If the source type is a pointer, the destination type must
6035 also be a pointer of the same size. This instruction supports bitwise
6036 conversion of vectors to integers and to vectors of other types (as
6037 long as they have the same size).
6042 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
6043 is always a *no-op cast* because no bits change with this
6044 conversion. The conversion is done as if the ``value`` had been stored
6045 to memory and read back as type ``ty2``. Pointer (or vector of
6046 pointers) types may only be converted to other pointer (or vector of
6047 pointers) types with the same address space through this instruction.
6048 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
6049 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
6054 .. code-block:: llvm
6056 %X = bitcast i8 255 to i8 ; yields i8 :-1
6057 %Y = bitcast i32* %x to sint* ; yields sint*:%x
6058 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
6059 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
6061 .. _i_addrspacecast:
6063 '``addrspacecast .. to``' Instruction
6064 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6071 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
6076 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
6077 address space ``n`` to type ``pty2`` in address space ``m``.
6082 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
6083 to cast and a pointer type to cast it to, which must have a different
6089 The '``addrspacecast``' instruction converts the pointer value
6090 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
6091 value modification, depending on the target and the address space
6092 pair. Pointer conversions within the same address space must be
6093 performed with the ``bitcast`` instruction. Note that if the address space
6094 conversion is legal then both result and operand refer to the same memory
6100 .. code-block:: llvm
6102 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
6103 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
6104 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
6111 The instructions in this category are the "miscellaneous" instructions,
6112 which defy better classification.
6116 '``icmp``' Instruction
6117 ^^^^^^^^^^^^^^^^^^^^^^
6124 <result> = icmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6129 The '``icmp``' instruction returns a boolean value or a vector of
6130 boolean values based on comparison of its two integer, integer vector,
6131 pointer, or pointer vector operands.
6136 The '``icmp``' instruction takes three operands. The first operand is
6137 the condition code indicating the kind of comparison to perform. It is
6138 not a value, just a keyword. The possible condition code are:
6141 #. ``ne``: not equal
6142 #. ``ugt``: unsigned greater than
6143 #. ``uge``: unsigned greater or equal
6144 #. ``ult``: unsigned less than
6145 #. ``ule``: unsigned less or equal
6146 #. ``sgt``: signed greater than
6147 #. ``sge``: signed greater or equal
6148 #. ``slt``: signed less than
6149 #. ``sle``: signed less or equal
6151 The remaining two arguments must be :ref:`integer <t_integer>` or
6152 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
6153 must also be identical types.
6158 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
6159 code given as ``cond``. The comparison performed always yields either an
6160 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
6162 #. ``eq``: yields ``true`` if the operands are equal, ``false``
6163 otherwise. No sign interpretation is necessary or performed.
6164 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
6165 otherwise. No sign interpretation is necessary or performed.
6166 #. ``ugt``: interprets the operands as unsigned values and yields
6167 ``true`` if ``op1`` is greater than ``op2``.
6168 #. ``uge``: interprets the operands as unsigned values and yields
6169 ``true`` if ``op1`` is greater than or equal to ``op2``.
6170 #. ``ult``: interprets the operands as unsigned values and yields
6171 ``true`` if ``op1`` is less than ``op2``.
6172 #. ``ule``: interprets the operands as unsigned values and yields
6173 ``true`` if ``op1`` is less than or equal to ``op2``.
6174 #. ``sgt``: interprets the operands as signed values and yields ``true``
6175 if ``op1`` is greater than ``op2``.
6176 #. ``sge``: interprets the operands as signed values and yields ``true``
6177 if ``op1`` is greater than or equal to ``op2``.
6178 #. ``slt``: interprets the operands as signed values and yields ``true``
6179 if ``op1`` is less than ``op2``.
6180 #. ``sle``: interprets the operands as signed values and yields ``true``
6181 if ``op1`` is less than or equal to ``op2``.
6183 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
6184 are compared as if they were integers.
6186 If the operands are integer vectors, then they are compared element by
6187 element. The result is an ``i1`` vector with the same number of elements
6188 as the values being compared. Otherwise, the result is an ``i1``.
6193 .. code-block:: llvm
6195 <result> = icmp eq i32 4, 5 ; yields: result=false
6196 <result> = icmp ne float* %X, %X ; yields: result=false
6197 <result> = icmp ult i16 4, 5 ; yields: result=true
6198 <result> = icmp sgt i16 4, 5 ; yields: result=false
6199 <result> = icmp ule i16 -4, 5 ; yields: result=false
6200 <result> = icmp sge i16 4, 5 ; yields: result=false
6202 Note that the code generator does not yet support vector types with the
6203 ``icmp`` instruction.
6207 '``fcmp``' Instruction
6208 ^^^^^^^^^^^^^^^^^^^^^^
6215 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6220 The '``fcmp``' instruction returns a boolean value or vector of boolean
6221 values based on comparison of its operands.
6223 If the operands are floating point scalars, then the result type is a
6224 boolean (:ref:`i1 <t_integer>`).
6226 If the operands are floating point vectors, then the result type is a
6227 vector of boolean with the same number of elements as the operands being
6233 The '``fcmp``' instruction takes three operands. The first operand is
6234 the condition code indicating the kind of comparison to perform. It is
6235 not a value, just a keyword. The possible condition code are:
6237 #. ``false``: no comparison, always returns false
6238 #. ``oeq``: ordered and equal
6239 #. ``ogt``: ordered and greater than
6240 #. ``oge``: ordered and greater than or equal
6241 #. ``olt``: ordered and less than
6242 #. ``ole``: ordered and less than or equal
6243 #. ``one``: ordered and not equal
6244 #. ``ord``: ordered (no nans)
6245 #. ``ueq``: unordered or equal
6246 #. ``ugt``: unordered or greater than
6247 #. ``uge``: unordered or greater than or equal
6248 #. ``ult``: unordered or less than
6249 #. ``ule``: unordered or less than or equal
6250 #. ``une``: unordered or not equal
6251 #. ``uno``: unordered (either nans)
6252 #. ``true``: no comparison, always returns true
6254 *Ordered* means that neither operand is a QNAN while *unordered* means
6255 that either operand may be a QNAN.
6257 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
6258 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
6259 type. They must have identical types.
6264 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
6265 condition code given as ``cond``. If the operands are vectors, then the
6266 vectors are compared element by element. Each comparison performed
6267 always yields an :ref:`i1 <t_integer>` result, as follows:
6269 #. ``false``: always yields ``false``, regardless of operands.
6270 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
6271 is equal to ``op2``.
6272 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
6273 is greater than ``op2``.
6274 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
6275 is greater than or equal to ``op2``.
6276 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
6277 is less than ``op2``.
6278 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
6279 is less than or equal to ``op2``.
6280 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
6281 is not equal to ``op2``.
6282 #. ``ord``: yields ``true`` if both operands are not a QNAN.
6283 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
6285 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
6286 greater than ``op2``.
6287 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
6288 greater than or equal to ``op2``.
6289 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
6291 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
6292 less than or equal to ``op2``.
6293 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
6294 not equal to ``op2``.
6295 #. ``uno``: yields ``true`` if either operand is a QNAN.
6296 #. ``true``: always yields ``true``, regardless of operands.
6301 .. code-block:: llvm
6303 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
6304 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
6305 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
6306 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
6308 Note that the code generator does not yet support vector types with the
6309 ``fcmp`` instruction.
6313 '``phi``' Instruction
6314 ^^^^^^^^^^^^^^^^^^^^^
6321 <result> = phi <ty> [ <val0>, <label0>], ...
6326 The '``phi``' instruction is used to implement the φ node in the SSA
6327 graph representing the function.
6332 The type of the incoming values is specified with the first type field.
6333 After this, the '``phi``' instruction takes a list of pairs as
6334 arguments, with one pair for each predecessor basic block of the current
6335 block. Only values of :ref:`first class <t_firstclass>` type may be used as
6336 the value arguments to the PHI node. Only labels may be used as the
6339 There must be no non-phi instructions between the start of a basic block
6340 and the PHI instructions: i.e. PHI instructions must be first in a basic
6343 For the purposes of the SSA form, the use of each incoming value is
6344 deemed to occur on the edge from the corresponding predecessor block to
6345 the current block (but after any definition of an '``invoke``'
6346 instruction's return value on the same edge).
6351 At runtime, the '``phi``' instruction logically takes on the value
6352 specified by the pair corresponding to the predecessor basic block that
6353 executed just prior to the current block.
6358 .. code-block:: llvm
6360 Loop: ; Infinite loop that counts from 0 on up...
6361 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
6362 %nextindvar = add i32 %indvar, 1
6367 '``select``' Instruction
6368 ^^^^^^^^^^^^^^^^^^^^^^^^
6375 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
6377 selty is either i1 or {<N x i1>}
6382 The '``select``' instruction is used to choose one value based on a
6383 condition, without IR-level branching.
6388 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
6389 values indicating the condition, and two values of the same :ref:`first
6390 class <t_firstclass>` type. If the val1/val2 are vectors and the
6391 condition is a scalar, then entire vectors are selected, not individual
6397 If the condition is an i1 and it evaluates to 1, the instruction returns
6398 the first value argument; otherwise, it returns the second value
6401 If the condition is a vector of i1, then the value arguments must be
6402 vectors of the same size, and the selection is done element by element.
6407 .. code-block:: llvm
6409 %X = select i1 true, i8 17, i8 42 ; yields i8:17
6413 '``call``' Instruction
6414 ^^^^^^^^^^^^^^^^^^^^^^
6421 <result> = [tail | musttail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
6426 The '``call``' instruction represents a simple function call.
6431 This instruction requires several arguments:
6433 #. The optional ``tail`` and ``musttail`` markers indicate that the optimizers
6434 should perform tail call optimization. The ``tail`` marker is a hint that
6435 `can be ignored <CodeGenerator.html#sibcallopt>`_. The ``musttail`` marker
6436 means that the call must be tail call optimized in order for the program to
6437 be correct. The ``musttail`` marker provides these guarantees:
6439 #. The call will not cause unbounded stack growth if it is part of a
6440 recursive cycle in the call graph.
6441 #. Arguments with the :ref:`inalloca <attr_inalloca>` attribute are
6444 Both markers imply that the callee does not access allocas or varargs from
6445 the caller. Calls marked ``musttail`` must obey the following additional
6448 - The call must immediately precede a :ref:`ret <i_ret>` instruction,
6449 or a pointer bitcast followed by a ret instruction.
6450 - The ret instruction must return the (possibly bitcasted) value
6451 produced by the call or void.
6452 - The caller and callee prototypes must match. Pointer types of
6453 parameters or return types may differ in pointee type, but not
6455 - The calling conventions of the caller and callee must match.
6456 - All ABI-impacting function attributes, such as sret, byval, inreg,
6457 returned, and inalloca, must match.
6459 Tail call optimization for calls marked ``tail`` is guaranteed to occur if
6460 the following conditions are met:
6462 - Caller and callee both have the calling convention ``fastcc``.
6463 - The call is in tail position (ret immediately follows call and ret
6464 uses value of call or is void).
6465 - Option ``-tailcallopt`` is enabled, or
6466 ``llvm::GuaranteedTailCallOpt`` is ``true``.
6467 - `Platform-specific constraints are
6468 met. <CodeGenerator.html#tailcallopt>`_
6470 #. The optional "cconv" marker indicates which :ref:`calling
6471 convention <callingconv>` the call should use. If none is
6472 specified, the call defaults to using C calling conventions. The
6473 calling convention of the call must match the calling convention of
6474 the target function, or else the behavior is undefined.
6475 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
6476 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
6478 #. '``ty``': the type of the call instruction itself which is also the
6479 type of the return value. Functions that return no value are marked
6481 #. '``fnty``': shall be the signature of the pointer to function value
6482 being invoked. The argument types must match the types implied by
6483 this signature. This type can be omitted if the function is not
6484 varargs and if the function type does not return a pointer to a
6486 #. '``fnptrval``': An LLVM value containing a pointer to a function to
6487 be invoked. In most cases, this is a direct function invocation, but
6488 indirect ``call``'s are just as possible, calling an arbitrary pointer
6490 #. '``function args``': argument list whose types match the function
6491 signature argument types and parameter attributes. All arguments must
6492 be of :ref:`first class <t_firstclass>` type. If the function signature
6493 indicates the function accepts a variable number of arguments, the
6494 extra arguments can be specified.
6495 #. The optional :ref:`function attributes <fnattrs>` list. Only
6496 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
6497 attributes are valid here.
6502 The '``call``' instruction is used to cause control flow to transfer to
6503 a specified function, with its incoming arguments bound to the specified
6504 values. Upon a '``ret``' instruction in the called function, control
6505 flow continues with the instruction after the function call, and the
6506 return value of the function is bound to the result argument.
6511 .. code-block:: llvm
6513 %retval = call i32 @test(i32 %argc)
6514 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
6515 %X = tail call i32 @foo() ; yields i32
6516 %Y = tail call fastcc i32 @foo() ; yields i32
6517 call void %foo(i8 97 signext)
6519 %struct.A = type { i32, i8 }
6520 %r = call %struct.A @foo() ; yields { i32, i8 }
6521 %gr = extractvalue %struct.A %r, 0 ; yields i32
6522 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
6523 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
6524 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
6526 llvm treats calls to some functions with names and arguments that match
6527 the standard C99 library as being the C99 library functions, and may
6528 perform optimizations or generate code for them under that assumption.
6529 This is something we'd like to change in the future to provide better
6530 support for freestanding environments and non-C-based languages.
6534 '``va_arg``' Instruction
6535 ^^^^^^^^^^^^^^^^^^^^^^^^
6542 <resultval> = va_arg <va_list*> <arglist>, <argty>
6547 The '``va_arg``' instruction is used to access arguments passed through
6548 the "variable argument" area of a function call. It is used to implement
6549 the ``va_arg`` macro in C.
6554 This instruction takes a ``va_list*`` value and the type of the
6555 argument. It returns a value of the specified argument type and
6556 increments the ``va_list`` to point to the next argument. The actual
6557 type of ``va_list`` is target specific.
6562 The '``va_arg``' instruction loads an argument of the specified type
6563 from the specified ``va_list`` and causes the ``va_list`` to point to
6564 the next argument. For more information, see the variable argument
6565 handling :ref:`Intrinsic Functions <int_varargs>`.
6567 It is legal for this instruction to be called in a function which does
6568 not take a variable number of arguments, for example, the ``vfprintf``
6571 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
6572 function <intrinsics>` because it takes a type as an argument.
6577 See the :ref:`variable argument processing <int_varargs>` section.
6579 Note that the code generator does not yet fully support va\_arg on many
6580 targets. Also, it does not currently support va\_arg with aggregate
6581 types on any target.
6585 '``landingpad``' Instruction
6586 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6593 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
6594 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
6596 <clause> := catch <type> <value>
6597 <clause> := filter <array constant type> <array constant>
6602 The '``landingpad``' instruction is used by `LLVM's exception handling
6603 system <ExceptionHandling.html#overview>`_ to specify that a basic block
6604 is a landing pad --- one where the exception lands, and corresponds to the
6605 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
6606 defines values supplied by the personality function (``pers_fn``) upon
6607 re-entry to the function. The ``resultval`` has the type ``resultty``.
6612 This instruction takes a ``pers_fn`` value. This is the personality
6613 function associated with the unwinding mechanism. The optional
6614 ``cleanup`` flag indicates that the landing pad block is a cleanup.
6616 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
6617 contains the global variable representing the "type" that may be caught
6618 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
6619 clause takes an array constant as its argument. Use
6620 "``[0 x i8**] undef``" for a filter which cannot throw. The
6621 '``landingpad``' instruction must contain *at least* one ``clause`` or
6622 the ``cleanup`` flag.
6627 The '``landingpad``' instruction defines the values which are set by the
6628 personality function (``pers_fn``) upon re-entry to the function, and
6629 therefore the "result type" of the ``landingpad`` instruction. As with
6630 calling conventions, how the personality function results are
6631 represented in LLVM IR is target specific.
6633 The clauses are applied in order from top to bottom. If two
6634 ``landingpad`` instructions are merged together through inlining, the
6635 clauses from the calling function are appended to the list of clauses.
6636 When the call stack is being unwound due to an exception being thrown,
6637 the exception is compared against each ``clause`` in turn. If it doesn't
6638 match any of the clauses, and the ``cleanup`` flag is not set, then
6639 unwinding continues further up the call stack.
6641 The ``landingpad`` instruction has several restrictions:
6643 - A landing pad block is a basic block which is the unwind destination
6644 of an '``invoke``' instruction.
6645 - A landing pad block must have a '``landingpad``' instruction as its
6646 first non-PHI instruction.
6647 - There can be only one '``landingpad``' instruction within the landing
6649 - A basic block that is not a landing pad block may not include a
6650 '``landingpad``' instruction.
6651 - All '``landingpad``' instructions in a function must have the same
6652 personality function.
6657 .. code-block:: llvm
6659 ;; A landing pad which can catch an integer.
6660 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6662 ;; A landing pad that is a cleanup.
6663 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6665 ;; A landing pad which can catch an integer and can only throw a double.
6666 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6668 filter [1 x i8**] [@_ZTId]
6675 LLVM supports the notion of an "intrinsic function". These functions
6676 have well known names and semantics and are required to follow certain
6677 restrictions. Overall, these intrinsics represent an extension mechanism
6678 for the LLVM language that does not require changing all of the
6679 transformations in LLVM when adding to the language (or the bitcode
6680 reader/writer, the parser, etc...).
6682 Intrinsic function names must all start with an "``llvm.``" prefix. This
6683 prefix is reserved in LLVM for intrinsic names; thus, function names may
6684 not begin with this prefix. Intrinsic functions must always be external
6685 functions: you cannot define the body of intrinsic functions. Intrinsic
6686 functions may only be used in call or invoke instructions: it is illegal
6687 to take the address of an intrinsic function. Additionally, because
6688 intrinsic functions are part of the LLVM language, it is required if any
6689 are added that they be documented here.
6691 Some intrinsic functions can be overloaded, i.e., the intrinsic
6692 represents a family of functions that perform the same operation but on
6693 different data types. Because LLVM can represent over 8 million
6694 different integer types, overloading is used commonly to allow an
6695 intrinsic function to operate on any integer type. One or more of the
6696 argument types or the result type can be overloaded to accept any
6697 integer type. Argument types may also be defined as exactly matching a
6698 previous argument's type or the result type. This allows an intrinsic
6699 function which accepts multiple arguments, but needs all of them to be
6700 of the same type, to only be overloaded with respect to a single
6701 argument or the result.
6703 Overloaded intrinsics will have the names of its overloaded argument
6704 types encoded into its function name, each preceded by a period. Only
6705 those types which are overloaded result in a name suffix. Arguments
6706 whose type is matched against another type do not. For example, the
6707 ``llvm.ctpop`` function can take an integer of any width and returns an
6708 integer of exactly the same integer width. This leads to a family of
6709 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
6710 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
6711 overloaded, and only one type suffix is required. Because the argument's
6712 type is matched against the return type, it does not require its own
6715 To learn how to add an intrinsic function, please see the `Extending
6716 LLVM Guide <ExtendingLLVM.html>`_.
6720 Variable Argument Handling Intrinsics
6721 -------------------------------------
6723 Variable argument support is defined in LLVM with the
6724 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
6725 functions. These functions are related to the similarly named macros
6726 defined in the ``<stdarg.h>`` header file.
6728 All of these functions operate on arguments that use a target-specific
6729 value type "``va_list``". The LLVM assembly language reference manual
6730 does not define what this type is, so all transformations should be
6731 prepared to handle these functions regardless of the type used.
6733 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
6734 variable argument handling intrinsic functions are used.
6736 .. code-block:: llvm
6738 define i32 @test(i32 %X, ...) {
6739 ; Initialize variable argument processing
6741 %ap2 = bitcast i8** %ap to i8*
6742 call void @llvm.va_start(i8* %ap2)
6744 ; Read a single integer argument
6745 %tmp = va_arg i8** %ap, i32
6747 ; Demonstrate usage of llvm.va_copy and llvm.va_end
6749 %aq2 = bitcast i8** %aq to i8*
6750 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
6751 call void @llvm.va_end(i8* %aq2)
6753 ; Stop processing of arguments.
6754 call void @llvm.va_end(i8* %ap2)
6758 declare void @llvm.va_start(i8*)
6759 declare void @llvm.va_copy(i8*, i8*)
6760 declare void @llvm.va_end(i8*)
6764 '``llvm.va_start``' Intrinsic
6765 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6772 declare void @llvm.va_start(i8* <arglist>)
6777 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
6778 subsequent use by ``va_arg``.
6783 The argument is a pointer to a ``va_list`` element to initialize.
6788 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
6789 available in C. In a target-dependent way, it initializes the
6790 ``va_list`` element to which the argument points, so that the next call
6791 to ``va_arg`` will produce the first variable argument passed to the
6792 function. Unlike the C ``va_start`` macro, this intrinsic does not need
6793 to know the last argument of the function as the compiler can figure
6796 '``llvm.va_end``' Intrinsic
6797 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6804 declare void @llvm.va_end(i8* <arglist>)
6809 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
6810 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
6815 The argument is a pointer to a ``va_list`` to destroy.
6820 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
6821 available in C. In a target-dependent way, it destroys the ``va_list``
6822 element to which the argument points. Calls to
6823 :ref:`llvm.va_start <int_va_start>` and
6824 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
6829 '``llvm.va_copy``' Intrinsic
6830 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6837 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
6842 The '``llvm.va_copy``' intrinsic copies the current argument position
6843 from the source argument list to the destination argument list.
6848 The first argument is a pointer to a ``va_list`` element to initialize.
6849 The second argument is a pointer to a ``va_list`` element to copy from.
6854 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
6855 available in C. In a target-dependent way, it copies the source
6856 ``va_list`` element into the destination ``va_list`` element. This
6857 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
6858 arbitrarily complex and require, for example, memory allocation.
6860 Accurate Garbage Collection Intrinsics
6861 --------------------------------------
6863 LLVM support for `Accurate Garbage Collection <GarbageCollection.html>`_
6864 (GC) requires the implementation and generation of these intrinsics.
6865 These intrinsics allow identification of :ref:`GC roots on the
6866 stack <int_gcroot>`, as well as garbage collector implementations that
6867 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
6868 Front-ends for type-safe garbage collected languages should generate
6869 these intrinsics to make use of the LLVM garbage collectors. For more
6870 details, see `Accurate Garbage Collection with
6871 LLVM <GarbageCollection.html>`_.
6873 The garbage collection intrinsics only operate on objects in the generic
6874 address space (address space zero).
6878 '``llvm.gcroot``' Intrinsic
6879 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6886 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
6891 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
6892 the code generator, and allows some metadata to be associated with it.
6897 The first argument specifies the address of a stack object that contains
6898 the root pointer. The second pointer (which must be either a constant or
6899 a global value address) contains the meta-data to be associated with the
6905 At runtime, a call to this intrinsic stores a null pointer into the
6906 "ptrloc" location. At compile-time, the code generator generates
6907 information to allow the runtime to find the pointer at GC safe points.
6908 The '``llvm.gcroot``' intrinsic may only be used in a function which
6909 :ref:`specifies a GC algorithm <gc>`.
6913 '``llvm.gcread``' Intrinsic
6914 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6921 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
6926 The '``llvm.gcread``' intrinsic identifies reads of references from heap
6927 locations, allowing garbage collector implementations that require read
6933 The second argument is the address to read from, which should be an
6934 address allocated from the garbage collector. The first object is a
6935 pointer to the start of the referenced object, if needed by the language
6936 runtime (otherwise null).
6941 The '``llvm.gcread``' intrinsic has the same semantics as a load
6942 instruction, but may be replaced with substantially more complex code by
6943 the garbage collector runtime, as needed. The '``llvm.gcread``'
6944 intrinsic may only be used in a function which :ref:`specifies a GC
6949 '``llvm.gcwrite``' Intrinsic
6950 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6957 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
6962 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
6963 locations, allowing garbage collector implementations that require write
6964 barriers (such as generational or reference counting collectors).
6969 The first argument is the reference to store, the second is the start of
6970 the object to store it to, and the third is the address of the field of
6971 Obj to store to. If the runtime does not require a pointer to the
6972 object, Obj may be null.
6977 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
6978 instruction, but may be replaced with substantially more complex code by
6979 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
6980 intrinsic may only be used in a function which :ref:`specifies a GC
6983 Code Generator Intrinsics
6984 -------------------------
6986 These intrinsics are provided by LLVM to expose special features that
6987 may only be implemented with code generator support.
6989 '``llvm.returnaddress``' Intrinsic
6990 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6997 declare i8 *@llvm.returnaddress(i32 <level>)
7002 The '``llvm.returnaddress``' intrinsic attempts to compute a
7003 target-specific value indicating the return address of the current
7004 function or one of its callers.
7009 The argument to this intrinsic indicates which function to return the
7010 address for. Zero indicates the calling function, one indicates its
7011 caller, etc. The argument is **required** to be a constant integer
7017 The '``llvm.returnaddress``' intrinsic either returns a pointer
7018 indicating the return address of the specified call frame, or zero if it
7019 cannot be identified. The value returned by this intrinsic is likely to
7020 be incorrect or 0 for arguments other than zero, so it should only be
7021 used for debugging purposes.
7023 Note that calling this intrinsic does not prevent function inlining or
7024 other aggressive transformations, so the value returned may not be that
7025 of the obvious source-language caller.
7027 '``llvm.frameaddress``' Intrinsic
7028 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7035 declare i8* @llvm.frameaddress(i32 <level>)
7040 The '``llvm.frameaddress``' intrinsic attempts to return the
7041 target-specific frame pointer value for the specified stack frame.
7046 The argument to this intrinsic indicates which function to return the
7047 frame pointer for. Zero indicates the calling function, one indicates
7048 its caller, etc. The argument is **required** to be a constant integer
7054 The '``llvm.frameaddress``' intrinsic either returns a pointer
7055 indicating the frame address of the specified call frame, or zero if it
7056 cannot be identified. The value returned by this intrinsic is likely to
7057 be incorrect or 0 for arguments other than zero, so it should only be
7058 used for debugging purposes.
7060 Note that calling this intrinsic does not prevent function inlining or
7061 other aggressive transformations, so the value returned may not be that
7062 of the obvious source-language caller.
7064 .. _int_read_register:
7065 .. _int_write_register:
7067 '``llvm.read_register``' and '``llvm.write_register``' Intrinsics
7068 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7075 declare i32 @llvm.read_register.i32(metadata)
7076 declare i64 @llvm.read_register.i64(metadata)
7077 declare void @llvm.write_register.i32(metadata, i32 @value)
7078 declare void @llvm.write_register.i64(metadata, i64 @value)
7079 !0 = metadata !{metadata !"sp\00"}
7084 The '``llvm.read_register``' and '``llvm.write_register``' intrinsics
7085 provides access to the named register. The register must be valid on
7086 the architecture being compiled to. The type needs to be compatible
7087 with the register being read.
7092 The '``llvm.read_register``' intrinsic returns the current value of the
7093 register, where possible. The '``llvm.write_register``' intrinsic sets
7094 the current value of the register, where possible.
7096 This is useful to implement named register global variables that need
7097 to always be mapped to a specific register, as is common practice on
7098 bare-metal programs including OS kernels.
7100 The compiler doesn't check for register availability or use of the used
7101 register in surrounding code, including inline assembly. Because of that,
7102 allocatable registers are not supported.
7104 Warning: So far it only works with the stack pointer on selected
7105 architectures (ARM, AArch64, PowerPC and x86_64). Significant amount of
7106 work is needed to support other registers and even more so, allocatable
7111 '``llvm.stacksave``' Intrinsic
7112 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7119 declare i8* @llvm.stacksave()
7124 The '``llvm.stacksave``' intrinsic is used to remember the current state
7125 of the function stack, for use with
7126 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
7127 implementing language features like scoped automatic variable sized
7133 This intrinsic returns a opaque pointer value that can be passed to
7134 :ref:`llvm.stackrestore <int_stackrestore>`. When an
7135 ``llvm.stackrestore`` intrinsic is executed with a value saved from
7136 ``llvm.stacksave``, it effectively restores the state of the stack to
7137 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
7138 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
7139 were allocated after the ``llvm.stacksave`` was executed.
7141 .. _int_stackrestore:
7143 '``llvm.stackrestore``' Intrinsic
7144 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7151 declare void @llvm.stackrestore(i8* %ptr)
7156 The '``llvm.stackrestore``' intrinsic is used to restore the state of
7157 the function stack to the state it was in when the corresponding
7158 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
7159 useful for implementing language features like scoped automatic variable
7160 sized arrays in C99.
7165 See the description for :ref:`llvm.stacksave <int_stacksave>`.
7167 '``llvm.prefetch``' Intrinsic
7168 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7175 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
7180 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
7181 insert a prefetch instruction if supported; otherwise, it is a noop.
7182 Prefetches have no effect on the behavior of the program but can change
7183 its performance characteristics.
7188 ``address`` is the address to be prefetched, ``rw`` is the specifier
7189 determining if the fetch should be for a read (0) or write (1), and
7190 ``locality`` is a temporal locality specifier ranging from (0) - no
7191 locality, to (3) - extremely local keep in cache. The ``cache type``
7192 specifies whether the prefetch is performed on the data (1) or
7193 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
7194 arguments must be constant integers.
7199 This intrinsic does not modify the behavior of the program. In
7200 particular, prefetches cannot trap and do not produce a value. On
7201 targets that support this intrinsic, the prefetch can provide hints to
7202 the processor cache for better performance.
7204 '``llvm.pcmarker``' Intrinsic
7205 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7212 declare void @llvm.pcmarker(i32 <id>)
7217 The '``llvm.pcmarker``' intrinsic is a method to export a Program
7218 Counter (PC) in a region of code to simulators and other tools. The
7219 method is target specific, but it is expected that the marker will use
7220 exported symbols to transmit the PC of the marker. The marker makes no
7221 guarantees that it will remain with any specific instruction after
7222 optimizations. It is possible that the presence of a marker will inhibit
7223 optimizations. The intended use is to be inserted after optimizations to
7224 allow correlations of simulation runs.
7229 ``id`` is a numerical id identifying the marker.
7234 This intrinsic does not modify the behavior of the program. Backends
7235 that do not support this intrinsic may ignore it.
7237 '``llvm.readcyclecounter``' Intrinsic
7238 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7245 declare i64 @llvm.readcyclecounter()
7250 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
7251 counter register (or similar low latency, high accuracy clocks) on those
7252 targets that support it. On X86, it should map to RDTSC. On Alpha, it
7253 should map to RPCC. As the backing counters overflow quickly (on the
7254 order of 9 seconds on alpha), this should only be used for small
7260 When directly supported, reading the cycle counter should not modify any
7261 memory. Implementations are allowed to either return a application
7262 specific value or a system wide value. On backends without support, this
7263 is lowered to a constant 0.
7265 Note that runtime support may be conditional on the privilege-level code is
7266 running at and the host platform.
7268 '``llvm.clear_cache``' Intrinsic
7269 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7276 declare void @llvm.clear_cache(i8*, i8*)
7281 The '``llvm.clear_cache``' intrinsic ensures visibility of modifications
7282 in the specified range to the execution unit of the processor. On
7283 targets with non-unified instruction and data cache, the implementation
7284 flushes the instruction cache.
7289 On platforms with coherent instruction and data caches (e.g. x86), this
7290 intrinsic is a nop. On platforms with non-coherent instruction and data
7291 cache (e.g. ARM, MIPS), the intrinsic is lowered either to appropriate
7292 instructions or a system call, if cache flushing requires special
7295 The default behavior is to emit a call to ``__clear_cache`` from the run
7298 This instrinsic does *not* empty the instruction pipeline. Modifications
7299 of the current function are outside the scope of the intrinsic.
7301 Standard C Library Intrinsics
7302 -----------------------------
7304 LLVM provides intrinsics for a few important standard C library
7305 functions. These intrinsics allow source-language front-ends to pass
7306 information about the alignment of the pointer arguments to the code
7307 generator, providing opportunity for more efficient code generation.
7311 '``llvm.memcpy``' Intrinsic
7312 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7317 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
7318 integer bit width and for different address spaces. Not all targets
7319 support all bit widths however.
7323 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7324 i32 <len>, i32 <align>, i1 <isvolatile>)
7325 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7326 i64 <len>, i32 <align>, i1 <isvolatile>)
7331 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7332 source location to the destination location.
7334 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
7335 intrinsics do not return a value, takes extra alignment/isvolatile
7336 arguments and the pointers can be in specified address spaces.
7341 The first argument is a pointer to the destination, the second is a
7342 pointer to the source. The third argument is an integer argument
7343 specifying the number of bytes to copy, the fourth argument is the
7344 alignment of the source and destination locations, and the fifth is a
7345 boolean indicating a volatile access.
7347 If the call to this intrinsic has an alignment value that is not 0 or 1,
7348 then the caller guarantees that both the source and destination pointers
7349 are aligned to that boundary.
7351 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
7352 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7353 very cleanly specified and it is unwise to depend on it.
7358 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7359 source location to the destination location, which are not allowed to
7360 overlap. It copies "len" bytes of memory over. If the argument is known
7361 to be aligned to some boundary, this can be specified as the fourth
7362 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
7364 '``llvm.memmove``' Intrinsic
7365 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7370 This is an overloaded intrinsic. You can use llvm.memmove on any integer
7371 bit width and for different address space. Not all targets support all
7376 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7377 i32 <len>, i32 <align>, i1 <isvolatile>)
7378 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7379 i64 <len>, i32 <align>, i1 <isvolatile>)
7384 The '``llvm.memmove.*``' intrinsics move a block of memory from the
7385 source location to the destination location. It is similar to the
7386 '``llvm.memcpy``' intrinsic but allows the two memory locations to
7389 Note that, unlike the standard libc function, the ``llvm.memmove.*``
7390 intrinsics do not return a value, takes extra alignment/isvolatile
7391 arguments and the pointers can be in specified address spaces.
7396 The first argument is a pointer to the destination, the second is a
7397 pointer to the source. The third argument is an integer argument
7398 specifying the number of bytes to copy, the fourth argument is the
7399 alignment of the source and destination locations, and the fifth is a
7400 boolean indicating a volatile access.
7402 If the call to this intrinsic has an alignment value that is not 0 or 1,
7403 then the caller guarantees that the source and destination pointers are
7404 aligned to that boundary.
7406 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
7407 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
7408 not very cleanly specified and it is unwise to depend on it.
7413 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
7414 source location to the destination location, which may overlap. It
7415 copies "len" bytes of memory over. If the argument is known to be
7416 aligned to some boundary, this can be specified as the fourth argument,
7417 otherwise it should be set to 0 or 1 (both meaning no alignment).
7419 '``llvm.memset.*``' Intrinsics
7420 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7425 This is an overloaded intrinsic. You can use llvm.memset on any integer
7426 bit width and for different address spaces. However, not all targets
7427 support all bit widths.
7431 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
7432 i32 <len>, i32 <align>, i1 <isvolatile>)
7433 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
7434 i64 <len>, i32 <align>, i1 <isvolatile>)
7439 The '``llvm.memset.*``' intrinsics fill a block of memory with a
7440 particular byte value.
7442 Note that, unlike the standard libc function, the ``llvm.memset``
7443 intrinsic does not return a value and takes extra alignment/volatile
7444 arguments. Also, the destination can be in an arbitrary address space.
7449 The first argument is a pointer to the destination to fill, the second
7450 is the byte value with which to fill it, the third argument is an
7451 integer argument specifying the number of bytes to fill, and the fourth
7452 argument is the known alignment of the destination location.
7454 If the call to this intrinsic has an alignment value that is not 0 or 1,
7455 then the caller guarantees that the destination pointer is aligned to
7458 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
7459 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7460 very cleanly specified and it is unwise to depend on it.
7465 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
7466 at the destination location. If the argument is known to be aligned to
7467 some boundary, this can be specified as the fourth argument, otherwise
7468 it should be set to 0 or 1 (both meaning no alignment).
7470 '``llvm.sqrt.*``' Intrinsic
7471 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7476 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
7477 floating point or vector of floating point type. Not all targets support
7482 declare float @llvm.sqrt.f32(float %Val)
7483 declare double @llvm.sqrt.f64(double %Val)
7484 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
7485 declare fp128 @llvm.sqrt.f128(fp128 %Val)
7486 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
7491 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
7492 returning the same value as the libm '``sqrt``' functions would. Unlike
7493 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
7494 negative numbers other than -0.0 (which allows for better optimization,
7495 because there is no need to worry about errno being set).
7496 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
7501 The argument and return value are floating point numbers of the same
7507 This function returns the sqrt of the specified operand if it is a
7508 nonnegative floating point number.
7510 '``llvm.powi.*``' Intrinsic
7511 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7516 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
7517 floating point or vector of floating point type. Not all targets support
7522 declare float @llvm.powi.f32(float %Val, i32 %power)
7523 declare double @llvm.powi.f64(double %Val, i32 %power)
7524 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7525 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7526 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
7531 The '``llvm.powi.*``' intrinsics return the first operand raised to the
7532 specified (positive or negative) power. The order of evaluation of
7533 multiplications is not defined. When a vector of floating point type is
7534 used, the second argument remains a scalar integer value.
7539 The second argument is an integer power, and the first is a value to
7540 raise to that power.
7545 This function returns the first value raised to the second power with an
7546 unspecified sequence of rounding operations.
7548 '``llvm.sin.*``' Intrinsic
7549 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7554 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
7555 floating point or vector of floating point type. Not all targets support
7560 declare float @llvm.sin.f32(float %Val)
7561 declare double @llvm.sin.f64(double %Val)
7562 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7563 declare fp128 @llvm.sin.f128(fp128 %Val)
7564 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7569 The '``llvm.sin.*``' intrinsics return the sine of the operand.
7574 The argument and return value are floating point numbers of the same
7580 This function returns the sine of the specified operand, returning the
7581 same values as the libm ``sin`` functions would, and handles error
7582 conditions in the same way.
7584 '``llvm.cos.*``' Intrinsic
7585 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7590 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
7591 floating point or vector of floating point type. Not all targets support
7596 declare float @llvm.cos.f32(float %Val)
7597 declare double @llvm.cos.f64(double %Val)
7598 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7599 declare fp128 @llvm.cos.f128(fp128 %Val)
7600 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7605 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
7610 The argument and return value are floating point numbers of the same
7616 This function returns the cosine of the specified operand, returning the
7617 same values as the libm ``cos`` functions would, and handles error
7618 conditions in the same way.
7620 '``llvm.pow.*``' Intrinsic
7621 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7626 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
7627 floating point or vector of floating point type. Not all targets support
7632 declare float @llvm.pow.f32(float %Val, float %Power)
7633 declare double @llvm.pow.f64(double %Val, double %Power)
7634 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7635 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7636 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7641 The '``llvm.pow.*``' intrinsics return the first operand raised to the
7642 specified (positive or negative) power.
7647 The second argument is a floating point power, and the first is a value
7648 to raise to that power.
7653 This function returns the first value raised to the second power,
7654 returning the same values as the libm ``pow`` functions would, and
7655 handles error conditions in the same way.
7657 '``llvm.exp.*``' Intrinsic
7658 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7663 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
7664 floating point or vector of floating point type. Not all targets support
7669 declare float @llvm.exp.f32(float %Val)
7670 declare double @llvm.exp.f64(double %Val)
7671 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7672 declare fp128 @llvm.exp.f128(fp128 %Val)
7673 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7678 The '``llvm.exp.*``' intrinsics perform the exp function.
7683 The argument and return value are floating point numbers of the same
7689 This function returns the same values as the libm ``exp`` functions
7690 would, and handles error conditions in the same way.
7692 '``llvm.exp2.*``' Intrinsic
7693 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7698 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
7699 floating point or vector of floating point type. Not all targets support
7704 declare float @llvm.exp2.f32(float %Val)
7705 declare double @llvm.exp2.f64(double %Val)
7706 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
7707 declare fp128 @llvm.exp2.f128(fp128 %Val)
7708 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
7713 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
7718 The argument and return value are floating point numbers of the same
7724 This function returns the same values as the libm ``exp2`` functions
7725 would, and handles error conditions in the same way.
7727 '``llvm.log.*``' Intrinsic
7728 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7733 This is an overloaded intrinsic. You can use ``llvm.log`` on any
7734 floating point or vector of floating point type. Not all targets support
7739 declare float @llvm.log.f32(float %Val)
7740 declare double @llvm.log.f64(double %Val)
7741 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7742 declare fp128 @llvm.log.f128(fp128 %Val)
7743 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7748 The '``llvm.log.*``' intrinsics perform the log function.
7753 The argument and return value are floating point numbers of the same
7759 This function returns the same values as the libm ``log`` functions
7760 would, and handles error conditions in the same way.
7762 '``llvm.log10.*``' Intrinsic
7763 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7768 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
7769 floating point or vector of floating point type. Not all targets support
7774 declare float @llvm.log10.f32(float %Val)
7775 declare double @llvm.log10.f64(double %Val)
7776 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
7777 declare fp128 @llvm.log10.f128(fp128 %Val)
7778 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
7783 The '``llvm.log10.*``' intrinsics perform the log10 function.
7788 The argument and return value are floating point numbers of the same
7794 This function returns the same values as the libm ``log10`` functions
7795 would, and handles error conditions in the same way.
7797 '``llvm.log2.*``' Intrinsic
7798 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7803 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
7804 floating point or vector of floating point type. Not all targets support
7809 declare float @llvm.log2.f32(float %Val)
7810 declare double @llvm.log2.f64(double %Val)
7811 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
7812 declare fp128 @llvm.log2.f128(fp128 %Val)
7813 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
7818 The '``llvm.log2.*``' intrinsics perform the log2 function.
7823 The argument and return value are floating point numbers of the same
7829 This function returns the same values as the libm ``log2`` functions
7830 would, and handles error conditions in the same way.
7832 '``llvm.fma.*``' Intrinsic
7833 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7838 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
7839 floating point or vector of floating point type. Not all targets support
7844 declare float @llvm.fma.f32(float %a, float %b, float %c)
7845 declare double @llvm.fma.f64(double %a, double %b, double %c)
7846 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7847 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7848 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7853 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
7859 The argument and return value are floating point numbers of the same
7865 This function returns the same values as the libm ``fma`` functions
7866 would, and does not set errno.
7868 '``llvm.fabs.*``' Intrinsic
7869 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7874 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
7875 floating point or vector of floating point type. Not all targets support
7880 declare float @llvm.fabs.f32(float %Val)
7881 declare double @llvm.fabs.f64(double %Val)
7882 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
7883 declare fp128 @llvm.fabs.f128(fp128 %Val)
7884 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
7889 The '``llvm.fabs.*``' intrinsics return the absolute value of the
7895 The argument and return value are floating point numbers of the same
7901 This function returns the same values as the libm ``fabs`` functions
7902 would, and handles error conditions in the same way.
7904 '``llvm.copysign.*``' Intrinsic
7905 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7910 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
7911 floating point or vector of floating point type. Not all targets support
7916 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
7917 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
7918 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
7919 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
7920 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
7925 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
7926 first operand and the sign of the second operand.
7931 The arguments and return value are floating point numbers of the same
7937 This function returns the same values as the libm ``copysign``
7938 functions would, and handles error conditions in the same way.
7940 '``llvm.floor.*``' Intrinsic
7941 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7946 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
7947 floating point or vector of floating point type. Not all targets support
7952 declare float @llvm.floor.f32(float %Val)
7953 declare double @llvm.floor.f64(double %Val)
7954 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
7955 declare fp128 @llvm.floor.f128(fp128 %Val)
7956 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
7961 The '``llvm.floor.*``' intrinsics return the floor of the operand.
7966 The argument and return value are floating point numbers of the same
7972 This function returns the same values as the libm ``floor`` functions
7973 would, and handles error conditions in the same way.
7975 '``llvm.ceil.*``' Intrinsic
7976 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7981 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
7982 floating point or vector of floating point type. Not all targets support
7987 declare float @llvm.ceil.f32(float %Val)
7988 declare double @llvm.ceil.f64(double %Val)
7989 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
7990 declare fp128 @llvm.ceil.f128(fp128 %Val)
7991 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
7996 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
8001 The argument and return value are floating point numbers of the same
8007 This function returns the same values as the libm ``ceil`` functions
8008 would, and handles error conditions in the same way.
8010 '``llvm.trunc.*``' Intrinsic
8011 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8016 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
8017 floating point or vector of floating point type. Not all targets support
8022 declare float @llvm.trunc.f32(float %Val)
8023 declare double @llvm.trunc.f64(double %Val)
8024 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
8025 declare fp128 @llvm.trunc.f128(fp128 %Val)
8026 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
8031 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
8032 nearest integer not larger in magnitude than the operand.
8037 The argument and return value are floating point numbers of the same
8043 This function returns the same values as the libm ``trunc`` functions
8044 would, and handles error conditions in the same way.
8046 '``llvm.rint.*``' Intrinsic
8047 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8052 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
8053 floating point or vector of floating point type. Not all targets support
8058 declare float @llvm.rint.f32(float %Val)
8059 declare double @llvm.rint.f64(double %Val)
8060 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
8061 declare fp128 @llvm.rint.f128(fp128 %Val)
8062 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
8067 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
8068 nearest integer. It may raise an inexact floating-point exception if the
8069 operand isn't an integer.
8074 The argument and return value are floating point numbers of the same
8080 This function returns the same values as the libm ``rint`` functions
8081 would, and handles error conditions in the same way.
8083 '``llvm.nearbyint.*``' Intrinsic
8084 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8089 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
8090 floating point or vector of floating point type. Not all targets support
8095 declare float @llvm.nearbyint.f32(float %Val)
8096 declare double @llvm.nearbyint.f64(double %Val)
8097 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
8098 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
8099 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
8104 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
8110 The argument and return value are floating point numbers of the same
8116 This function returns the same values as the libm ``nearbyint``
8117 functions would, and handles error conditions in the same way.
8119 '``llvm.round.*``' Intrinsic
8120 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8125 This is an overloaded intrinsic. You can use ``llvm.round`` on any
8126 floating point or vector of floating point type. Not all targets support
8131 declare float @llvm.round.f32(float %Val)
8132 declare double @llvm.round.f64(double %Val)
8133 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
8134 declare fp128 @llvm.round.f128(fp128 %Val)
8135 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
8140 The '``llvm.round.*``' intrinsics returns the operand rounded to the
8146 The argument and return value are floating point numbers of the same
8152 This function returns the same values as the libm ``round``
8153 functions would, and handles error conditions in the same way.
8155 Bit Manipulation Intrinsics
8156 ---------------------------
8158 LLVM provides intrinsics for a few important bit manipulation
8159 operations. These allow efficient code generation for some algorithms.
8161 '``llvm.bswap.*``' Intrinsics
8162 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8167 This is an overloaded intrinsic function. You can use bswap on any
8168 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
8172 declare i16 @llvm.bswap.i16(i16 <id>)
8173 declare i32 @llvm.bswap.i32(i32 <id>)
8174 declare i64 @llvm.bswap.i64(i64 <id>)
8179 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
8180 values with an even number of bytes (positive multiple of 16 bits).
8181 These are useful for performing operations on data that is not in the
8182 target's native byte order.
8187 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
8188 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
8189 intrinsic returns an i32 value that has the four bytes of the input i32
8190 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
8191 returned i32 will have its bytes in 3, 2, 1, 0 order. The
8192 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
8193 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
8196 '``llvm.ctpop.*``' Intrinsic
8197 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8202 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
8203 bit width, or on any vector with integer elements. Not all targets
8204 support all bit widths or vector types, however.
8208 declare i8 @llvm.ctpop.i8(i8 <src>)
8209 declare i16 @llvm.ctpop.i16(i16 <src>)
8210 declare i32 @llvm.ctpop.i32(i32 <src>)
8211 declare i64 @llvm.ctpop.i64(i64 <src>)
8212 declare i256 @llvm.ctpop.i256(i256 <src>)
8213 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
8218 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
8224 The only argument is the value to be counted. The argument may be of any
8225 integer type, or a vector with integer elements. The return type must
8226 match the argument type.
8231 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
8232 each element of a vector.
8234 '``llvm.ctlz.*``' Intrinsic
8235 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8240 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
8241 integer bit width, or any vector whose elements are integers. Not all
8242 targets support all bit widths or vector types, however.
8246 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
8247 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
8248 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
8249 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
8250 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
8251 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
8256 The '``llvm.ctlz``' family of intrinsic functions counts the number of
8257 leading zeros in a variable.
8262 The first argument is the value to be counted. This argument may be of
8263 any integer type, or a vectory with integer element type. The return
8264 type must match the first argument type.
8266 The second argument must be a constant and is a flag to indicate whether
8267 the intrinsic should ensure that a zero as the first argument produces a
8268 defined result. Historically some architectures did not provide a
8269 defined result for zero values as efficiently, and many algorithms are
8270 now predicated on avoiding zero-value inputs.
8275 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
8276 zeros in a variable, or within each element of the vector. If
8277 ``src == 0`` then the result is the size in bits of the type of ``src``
8278 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
8279 ``llvm.ctlz(i32 2) = 30``.
8281 '``llvm.cttz.*``' Intrinsic
8282 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8287 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
8288 integer bit width, or any vector of integer elements. Not all targets
8289 support all bit widths or vector types, however.
8293 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
8294 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
8295 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
8296 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
8297 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
8298 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
8303 The '``llvm.cttz``' family of intrinsic functions counts the number of
8309 The first argument is the value to be counted. This argument may be of
8310 any integer type, or a vectory with integer element type. The return
8311 type must match the first argument type.
8313 The second argument must be a constant and is a flag to indicate whether
8314 the intrinsic should ensure that a zero as the first argument produces a
8315 defined result. Historically some architectures did not provide a
8316 defined result for zero values as efficiently, and many algorithms are
8317 now predicated on avoiding zero-value inputs.
8322 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
8323 zeros in a variable, or within each element of a vector. If ``src == 0``
8324 then the result is the size in bits of the type of ``src`` if
8325 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
8326 ``llvm.cttz(2) = 1``.
8328 Arithmetic with Overflow Intrinsics
8329 -----------------------------------
8331 LLVM provides intrinsics for some arithmetic with overflow operations.
8333 '``llvm.sadd.with.overflow.*``' Intrinsics
8334 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8339 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
8340 on any integer bit width.
8344 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
8345 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8346 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
8351 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8352 a signed addition of the two arguments, and indicate whether an overflow
8353 occurred during the signed summation.
8358 The arguments (%a and %b) and the first element of the result structure
8359 may be of integer types of any bit width, but they must have the same
8360 bit width. The second element of the result structure must be of type
8361 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8367 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8368 a signed addition of the two variables. They return a structure --- the
8369 first element of which is the signed summation, and the second element
8370 of which is a bit specifying if the signed summation resulted in an
8376 .. code-block:: llvm
8378 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8379 %sum = extractvalue {i32, i1} %res, 0
8380 %obit = extractvalue {i32, i1} %res, 1
8381 br i1 %obit, label %overflow, label %normal
8383 '``llvm.uadd.with.overflow.*``' Intrinsics
8384 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8389 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
8390 on any integer bit width.
8394 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
8395 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8396 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
8401 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8402 an unsigned addition of the two arguments, and indicate whether a carry
8403 occurred during the unsigned summation.
8408 The arguments (%a and %b) and the first element of the result structure
8409 may be of integer types of any bit width, but they must have the same
8410 bit width. The second element of the result structure must be of type
8411 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8417 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8418 an unsigned addition of the two arguments. They return a structure --- the
8419 first element of which is the sum, and the second element of which is a
8420 bit specifying if the unsigned summation resulted in a carry.
8425 .. code-block:: llvm
8427 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8428 %sum = extractvalue {i32, i1} %res, 0
8429 %obit = extractvalue {i32, i1} %res, 1
8430 br i1 %obit, label %carry, label %normal
8432 '``llvm.ssub.with.overflow.*``' Intrinsics
8433 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8438 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
8439 on any integer bit width.
8443 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
8444 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8445 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
8450 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8451 a signed subtraction of the two arguments, and indicate whether an
8452 overflow occurred during the signed subtraction.
8457 The arguments (%a and %b) and the first element of the result structure
8458 may be of integer types of any bit width, but they must have the same
8459 bit width. The second element of the result structure must be of type
8460 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8466 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8467 a signed subtraction of the two arguments. They return a structure --- the
8468 first element of which is the subtraction, and the second element of
8469 which is a bit specifying if the signed subtraction resulted in an
8475 .. code-block:: llvm
8477 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8478 %sum = extractvalue {i32, i1} %res, 0
8479 %obit = extractvalue {i32, i1} %res, 1
8480 br i1 %obit, label %overflow, label %normal
8482 '``llvm.usub.with.overflow.*``' Intrinsics
8483 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8488 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
8489 on any integer bit width.
8493 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
8494 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8495 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
8500 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8501 an unsigned subtraction of the two arguments, and indicate whether an
8502 overflow occurred during the unsigned subtraction.
8507 The arguments (%a and %b) and the first element of the result structure
8508 may be of integer types of any bit width, but they must have the same
8509 bit width. The second element of the result structure must be of type
8510 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8516 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8517 an unsigned subtraction of the two arguments. They return a structure ---
8518 the first element of which is the subtraction, and the second element of
8519 which is a bit specifying if the unsigned subtraction resulted in an
8525 .. code-block:: llvm
8527 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8528 %sum = extractvalue {i32, i1} %res, 0
8529 %obit = extractvalue {i32, i1} %res, 1
8530 br i1 %obit, label %overflow, label %normal
8532 '``llvm.smul.with.overflow.*``' Intrinsics
8533 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8538 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
8539 on any integer bit width.
8543 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
8544 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8545 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
8550 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8551 a signed multiplication of the two arguments, and indicate whether an
8552 overflow occurred during the signed multiplication.
8557 The arguments (%a and %b) and the first element of the result structure
8558 may be of integer types of any bit width, but they must have the same
8559 bit width. The second element of the result structure must be of type
8560 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8566 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8567 a signed multiplication of the two arguments. They return a structure ---
8568 the first element of which is the multiplication, and the second element
8569 of which is a bit specifying if the signed multiplication resulted in an
8575 .. code-block:: llvm
8577 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8578 %sum = extractvalue {i32, i1} %res, 0
8579 %obit = extractvalue {i32, i1} %res, 1
8580 br i1 %obit, label %overflow, label %normal
8582 '``llvm.umul.with.overflow.*``' Intrinsics
8583 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8588 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
8589 on any integer bit width.
8593 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
8594 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8595 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
8600 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
8601 a unsigned multiplication of the two arguments, and indicate whether an
8602 overflow occurred during the unsigned multiplication.
8607 The arguments (%a and %b) and the first element of the result structure
8608 may be of integer types of any bit width, but they must have the same
8609 bit width. The second element of the result structure must be of type
8610 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8616 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
8617 an unsigned multiplication of the two arguments. They return a structure ---
8618 the first element of which is the multiplication, and the second
8619 element of which is a bit specifying if the unsigned multiplication
8620 resulted in an overflow.
8625 .. code-block:: llvm
8627 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8628 %sum = extractvalue {i32, i1} %res, 0
8629 %obit = extractvalue {i32, i1} %res, 1
8630 br i1 %obit, label %overflow, label %normal
8632 Specialised Arithmetic Intrinsics
8633 ---------------------------------
8635 '``llvm.fmuladd.*``' Intrinsic
8636 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8643 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
8644 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
8649 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
8650 expressions that can be fused if the code generator determines that (a) the
8651 target instruction set has support for a fused operation, and (b) that the
8652 fused operation is more efficient than the equivalent, separate pair of mul
8653 and add instructions.
8658 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
8659 multiplicands, a and b, and an addend c.
8668 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
8670 is equivalent to the expression a \* b + c, except that rounding will
8671 not be performed between the multiplication and addition steps if the
8672 code generator fuses the operations. Fusion is not guaranteed, even if
8673 the target platform supports it. If a fused multiply-add is required the
8674 corresponding llvm.fma.\* intrinsic function should be used
8675 instead. This never sets errno, just as '``llvm.fma.*``'.
8680 .. code-block:: llvm
8682 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields float:r2 = (a * b) + c
8684 Half Precision Floating Point Intrinsics
8685 ----------------------------------------
8687 For most target platforms, half precision floating point is a
8688 storage-only format. This means that it is a dense encoding (in memory)
8689 but does not support computation in the format.
8691 This means that code must first load the half-precision floating point
8692 value as an i16, then convert it to float with
8693 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
8694 then be performed on the float value (including extending to double
8695 etc). To store the value back to memory, it is first converted to float
8696 if needed, then converted to i16 with
8697 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
8700 .. _int_convert_to_fp16:
8702 '``llvm.convert.to.fp16``' Intrinsic
8703 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8710 declare i16 @llvm.convert.to.fp16.f32(float %a)
8711 declare i16 @llvm.convert.to.fp16.f64(double %a)
8716 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
8717 conventional floating point type to half precision floating point format.
8722 The intrinsic function contains single argument - the value to be
8728 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
8729 conventional floating point format to half precision floating point format. The
8730 return value is an ``i16`` which contains the converted number.
8735 .. code-block:: llvm
8737 %res = call i16 @llvm.convert.to.fp16.f32(float %a)
8738 store i16 %res, i16* @x, align 2
8740 .. _int_convert_from_fp16:
8742 '``llvm.convert.from.fp16``' Intrinsic
8743 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8750 declare float @llvm.convert.from.fp16.f32(i16 %a)
8751 declare double @llvm.convert.from.fp16.f64(i16 %a)
8756 The '``llvm.convert.from.fp16``' intrinsic function performs a
8757 conversion from half precision floating point format to single precision
8758 floating point format.
8763 The intrinsic function contains single argument - the value to be
8769 The '``llvm.convert.from.fp16``' intrinsic function performs a
8770 conversion from half single precision floating point format to single
8771 precision floating point format. The input half-float value is
8772 represented by an ``i16`` value.
8777 .. code-block:: llvm
8779 %a = load i16* @x, align 2
8780 %res = call float @llvm.convert.from.fp16(i16 %a)
8785 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
8786 prefix), are described in the `LLVM Source Level
8787 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
8790 Exception Handling Intrinsics
8791 -----------------------------
8793 The LLVM exception handling intrinsics (which all start with
8794 ``llvm.eh.`` prefix), are described in the `LLVM Exception
8795 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
8799 Trampoline Intrinsics
8800 ---------------------
8802 These intrinsics make it possible to excise one parameter, marked with
8803 the :ref:`nest <nest>` attribute, from a function. The result is a
8804 callable function pointer lacking the nest parameter - the caller does
8805 not need to provide a value for it. Instead, the value to use is stored
8806 in advance in a "trampoline", a block of memory usually allocated on the
8807 stack, which also contains code to splice the nest value into the
8808 argument list. This is used to implement the GCC nested function address
8811 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
8812 then the resulting function pointer has signature ``i32 (i32, i32)*``.
8813 It can be created as follows:
8815 .. code-block:: llvm
8817 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
8818 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
8819 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
8820 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
8821 %fp = bitcast i8* %p to i32 (i32, i32)*
8823 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
8824 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
8828 '``llvm.init.trampoline``' Intrinsic
8829 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8836 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
8841 This fills the memory pointed to by ``tramp`` with executable code,
8842 turning it into a trampoline.
8847 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
8848 pointers. The ``tramp`` argument must point to a sufficiently large and
8849 sufficiently aligned block of memory; this memory is written to by the
8850 intrinsic. Note that the size and the alignment are target-specific -
8851 LLVM currently provides no portable way of determining them, so a
8852 front-end that generates this intrinsic needs to have some
8853 target-specific knowledge. The ``func`` argument must hold a function
8854 bitcast to an ``i8*``.
8859 The block of memory pointed to by ``tramp`` is filled with target
8860 dependent code, turning it into a function. Then ``tramp`` needs to be
8861 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
8862 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
8863 function's signature is the same as that of ``func`` with any arguments
8864 marked with the ``nest`` attribute removed. At most one such ``nest``
8865 argument is allowed, and it must be of pointer type. Calling the new
8866 function is equivalent to calling ``func`` with the same argument list,
8867 but with ``nval`` used for the missing ``nest`` argument. If, after
8868 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
8869 modified, then the effect of any later call to the returned function
8870 pointer is undefined.
8874 '``llvm.adjust.trampoline``' Intrinsic
8875 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8882 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
8887 This performs any required machine-specific adjustment to the address of
8888 a trampoline (passed as ``tramp``).
8893 ``tramp`` must point to a block of memory which already has trampoline
8894 code filled in by a previous call to
8895 :ref:`llvm.init.trampoline <int_it>`.
8900 On some architectures the address of the code to be executed needs to be
8901 different than the address where the trampoline is actually stored. This
8902 intrinsic returns the executable address corresponding to ``tramp``
8903 after performing the required machine specific adjustments. The pointer
8904 returned can then be :ref:`bitcast and executed <int_trampoline>`.
8909 This class of intrinsics provides information about the lifetime of
8910 memory objects and ranges where variables are immutable.
8914 '``llvm.lifetime.start``' Intrinsic
8915 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8922 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
8927 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
8933 The first argument is a constant integer representing the size of the
8934 object, or -1 if it is variable sized. The second argument is a pointer
8940 This intrinsic indicates that before this point in the code, the value
8941 of the memory pointed to by ``ptr`` is dead. This means that it is known
8942 to never be used and has an undefined value. A load from the pointer
8943 that precedes this intrinsic can be replaced with ``'undef'``.
8947 '``llvm.lifetime.end``' Intrinsic
8948 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8955 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
8960 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
8966 The first argument is a constant integer representing the size of the
8967 object, or -1 if it is variable sized. The second argument is a pointer
8973 This intrinsic indicates that after this point in the code, the value of
8974 the memory pointed to by ``ptr`` is dead. This means that it is known to
8975 never be used and has an undefined value. Any stores into the memory
8976 object following this intrinsic may be removed as dead.
8978 '``llvm.invariant.start``' Intrinsic
8979 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8986 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
8991 The '``llvm.invariant.start``' intrinsic specifies that the contents of
8992 a memory object will not change.
8997 The first argument is a constant integer representing the size of the
8998 object, or -1 if it is variable sized. The second argument is a pointer
9004 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
9005 the return value, the referenced memory location is constant and
9008 '``llvm.invariant.end``' Intrinsic
9009 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9016 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
9021 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
9022 memory object are mutable.
9027 The first argument is the matching ``llvm.invariant.start`` intrinsic.
9028 The second argument is a constant integer representing the size of the
9029 object, or -1 if it is variable sized and the third argument is a
9030 pointer to the object.
9035 This intrinsic indicates that the memory is mutable again.
9040 This class of intrinsics is designed to be generic and has no specific
9043 '``llvm.var.annotation``' Intrinsic
9044 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9051 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9056 The '``llvm.var.annotation``' intrinsic.
9061 The first argument is a pointer to a value, the second is a pointer to a
9062 global string, the third is a pointer to a global string which is the
9063 source file name, and the last argument is the line number.
9068 This intrinsic allows annotation of local variables with arbitrary
9069 strings. This can be useful for special purpose optimizations that want
9070 to look for these annotations. These have no other defined use; they are
9071 ignored by code generation and optimization.
9073 '``llvm.ptr.annotation.*``' Intrinsic
9074 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9079 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
9080 pointer to an integer of any width. *NOTE* you must specify an address space for
9081 the pointer. The identifier for the default address space is the integer
9086 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9087 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
9088 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
9089 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
9090 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
9095 The '``llvm.ptr.annotation``' intrinsic.
9100 The first argument is a pointer to an integer value of arbitrary bitwidth
9101 (result of some expression), the second is a pointer to a global string, the
9102 third is a pointer to a global string which is the source file name, and the
9103 last argument is the line number. It returns the value of the first argument.
9108 This intrinsic allows annotation of a pointer to an integer with arbitrary
9109 strings. This can be useful for special purpose optimizations that want to look
9110 for these annotations. These have no other defined use; they are ignored by code
9111 generation and optimization.
9113 '``llvm.annotation.*``' Intrinsic
9114 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9119 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
9120 any integer bit width.
9124 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
9125 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
9126 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
9127 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
9128 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
9133 The '``llvm.annotation``' intrinsic.
9138 The first argument is an integer value (result of some expression), the
9139 second is a pointer to a global string, the third is a pointer to a
9140 global string which is the source file name, and the last argument is
9141 the line number. It returns the value of the first argument.
9146 This intrinsic allows annotations to be put on arbitrary expressions
9147 with arbitrary strings. This can be useful for special purpose
9148 optimizations that want to look for these annotations. These have no
9149 other defined use; they are ignored by code generation and optimization.
9151 '``llvm.trap``' Intrinsic
9152 ^^^^^^^^^^^^^^^^^^^^^^^^^
9159 declare void @llvm.trap() noreturn nounwind
9164 The '``llvm.trap``' intrinsic.
9174 This intrinsic is lowered to the target dependent trap instruction. If
9175 the target does not have a trap instruction, this intrinsic will be
9176 lowered to a call of the ``abort()`` function.
9178 '``llvm.debugtrap``' Intrinsic
9179 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9186 declare void @llvm.debugtrap() nounwind
9191 The '``llvm.debugtrap``' intrinsic.
9201 This intrinsic is lowered to code which is intended to cause an
9202 execution trap with the intention of requesting the attention of a
9205 '``llvm.stackprotector``' Intrinsic
9206 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9213 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
9218 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
9219 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
9220 is placed on the stack before local variables.
9225 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
9226 The first argument is the value loaded from the stack guard
9227 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
9228 enough space to hold the value of the guard.
9233 This intrinsic causes the prologue/epilogue inserter to force the position of
9234 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
9235 to ensure that if a local variable on the stack is overwritten, it will destroy
9236 the value of the guard. When the function exits, the guard on the stack is
9237 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
9238 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
9239 calling the ``__stack_chk_fail()`` function.
9241 '``llvm.stackprotectorcheck``' Intrinsic
9242 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9249 declare void @llvm.stackprotectorcheck(i8** <guard>)
9254 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
9255 created stack protector and if they are not equal calls the
9256 ``__stack_chk_fail()`` function.
9261 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
9262 the variable ``@__stack_chk_guard``.
9267 This intrinsic is provided to perform the stack protector check by comparing
9268 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
9269 values do not match call the ``__stack_chk_fail()`` function.
9271 The reason to provide this as an IR level intrinsic instead of implementing it
9272 via other IR operations is that in order to perform this operation at the IR
9273 level without an intrinsic, one would need to create additional basic blocks to
9274 handle the success/failure cases. This makes it difficult to stop the stack
9275 protector check from disrupting sibling tail calls in Codegen. With this
9276 intrinsic, we are able to generate the stack protector basic blocks late in
9277 codegen after the tail call decision has occurred.
9279 '``llvm.objectsize``' Intrinsic
9280 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9287 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
9288 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
9293 The ``llvm.objectsize`` intrinsic is designed to provide information to
9294 the optimizers to determine at compile time whether a) an operation
9295 (like memcpy) will overflow a buffer that corresponds to an object, or
9296 b) that a runtime check for overflow isn't necessary. An object in this
9297 context means an allocation of a specific class, structure, array, or
9303 The ``llvm.objectsize`` intrinsic takes two arguments. The first
9304 argument is a pointer to or into the ``object``. The second argument is
9305 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
9306 or -1 (if false) when the object size is unknown. The second argument
9307 only accepts constants.
9312 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
9313 the size of the object concerned. If the size cannot be determined at
9314 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
9315 on the ``min`` argument).
9317 '``llvm.expect``' Intrinsic
9318 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9323 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
9328 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
9329 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
9330 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
9335 The ``llvm.expect`` intrinsic provides information about expected (the
9336 most probable) value of ``val``, which can be used by optimizers.
9341 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
9342 a value. The second argument is an expected value, this needs to be a
9343 constant value, variables are not allowed.
9348 This intrinsic is lowered to the ``val``.
9350 '``llvm.donothing``' Intrinsic
9351 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9358 declare void @llvm.donothing() nounwind readnone
9363 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's the
9364 only intrinsic that can be called with an invoke instruction.
9374 This intrinsic does nothing, and it's removed by optimizers and ignored
9377 Stack Map Intrinsics
9378 --------------------
9380 LLVM provides experimental intrinsics to support runtime patching
9381 mechanisms commonly desired in dynamic language JITs. These intrinsics
9382 are described in :doc:`StackMaps`.