1 ==============================
2 LLVM Language Reference Manual
3 ==============================
12 This document is a reference manual for the LLVM assembly language. LLVM
13 is a Static Single Assignment (SSA) based representation that provides
14 type safety, low-level operations, flexibility, and the capability of
15 representing 'all' high-level languages cleanly. It is the common code
16 representation used throughout all phases of the LLVM compilation
22 The LLVM code representation is designed to be used in three different
23 forms: as an in-memory compiler IR, as an on-disk bitcode representation
24 (suitable for fast loading by a Just-In-Time compiler), and as a human
25 readable assembly language representation. This allows LLVM to provide a
26 powerful intermediate representation for efficient compiler
27 transformations and analysis, while providing a natural means to debug
28 and visualize the transformations. The three different forms of LLVM are
29 all equivalent. This document describes the human readable
30 representation and notation.
32 The LLVM representation aims to be light-weight and low-level while
33 being expressive, typed, and extensible at the same time. It aims to be
34 a "universal IR" of sorts, by being at a low enough level that
35 high-level ideas may be cleanly mapped to it (similar to how
36 microprocessors are "universal IR's", allowing many source languages to
37 be mapped to them). By providing type information, LLVM can be used as
38 the target of optimizations: for example, through pointer analysis, it
39 can be proven that a C automatic variable is never accessed outside of
40 the current function, allowing it to be promoted to a simple SSA value
41 instead of a memory location.
48 It is important to note that this document describes 'well formed' LLVM
49 assembly language. There is a difference between what the parser accepts
50 and what is considered 'well formed'. For example, the following
51 instruction is syntactically okay, but not well formed:
57 because the definition of ``%x`` does not dominate all of its uses. The
58 LLVM infrastructure provides a verification pass that may be used to
59 verify that an LLVM module is well formed. This pass is automatically
60 run by the parser after parsing input assembly and by the optimizer
61 before it outputs bitcode. The violations pointed out by the verifier
62 pass indicate bugs in transformation passes or input to the parser.
69 LLVM identifiers come in two basic types: global and local. Global
70 identifiers (functions, global variables) begin with the ``'@'``
71 character. Local identifiers (register names, types) begin with the
72 ``'%'`` character. Additionally, there are three different formats for
73 identifiers, for different purposes:
75 #. Named values are represented as a string of characters with their
76 prefix. For example, ``%foo``, ``@DivisionByZero``,
77 ``%a.really.long.identifier``. The actual regular expression used is
78 '``[%@][-a-zA-Z$._][-a-zA-Z$._0-9]*``'. Identifiers that require other
79 characters in their names can be surrounded with quotes. Special
80 characters may be escaped using ``"\xx"`` where ``xx`` is the ASCII
81 code for the character in hexadecimal. In this way, any character can
82 be used in a name value, even quotes themselves. The ``"\01"`` prefix
83 can be used on global variables to suppress mangling.
84 #. Unnamed values are represented as an unsigned numeric value with
85 their prefix. For example, ``%12``, ``@2``, ``%44``.
86 #. Constants, which are described in the section Constants_ below.
88 LLVM requires that values start with a prefix for two reasons: Compilers
89 don't need to worry about name clashes with reserved words, and the set
90 of reserved words may be expanded in the future without penalty.
91 Additionally, unnamed identifiers allow a compiler to quickly come up
92 with a temporary variable without having to avoid symbol table
95 Reserved words in LLVM are very similar to reserved words in other
96 languages. There are keywords for different opcodes ('``add``',
97 '``bitcast``', '``ret``', etc...), for primitive type names ('``void``',
98 '``i32``', etc...), and others. These reserved words cannot conflict
99 with variable names, because none of them start with a prefix character
100 (``'%'`` or ``'@'``).
102 Here is an example of LLVM code to multiply the integer variable
109 %result = mul i32 %X, 8
111 After strength reduction:
115 %result = shl i32 %X, 3
121 %0 = add i32 %X, %X ; yields i32:%0
122 %1 = add i32 %0, %0 ; yields i32:%1
123 %result = add i32 %1, %1
125 This last way of multiplying ``%X`` by 8 illustrates several important
126 lexical features of LLVM:
128 #. Comments are delimited with a '``;``' and go until the end of line.
129 #. Unnamed temporaries are created when the result of a computation is
130 not assigned to a named value.
131 #. Unnamed temporaries are numbered sequentially (using a per-function
132 incrementing counter, starting with 0). Note that basic blocks and unnamed
133 function parameters are included in this numbering. For example, if the
134 entry basic block is not given a label name and all function parameters are
135 named, then it will get number 0.
137 It also shows a convention that we follow in this document. When
138 demonstrating instructions, we will follow an instruction with a comment
139 that defines the type and name of value produced.
147 LLVM programs are composed of ``Module``'s, each of which is a
148 translation unit of the input programs. Each module consists of
149 functions, global variables, and symbol table entries. Modules may be
150 combined together with the LLVM linker, which merges function (and
151 global variable) definitions, resolves forward declarations, and merges
152 symbol table entries. Here is an example of the "hello world" module:
156 ; Declare the string constant as a global constant.
157 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
159 ; External declaration of the puts function
160 declare i32 @puts(i8* nocapture) nounwind
162 ; Definition of main function
163 define i32 @main() { ; i32()*
164 ; Convert [13 x i8]* to i8 *...
165 %cast210 = getelementptr [13 x i8], [13 x i8]* @.str, i64 0, i64 0
167 ; Call puts function to write out the string to stdout.
168 call i32 @puts(i8* %cast210)
173 !0 = !{i32 42, null, !"string"}
176 This example is made up of a :ref:`global variable <globalvars>` named
177 "``.str``", an external declaration of the "``puts``" function, a
178 :ref:`function definition <functionstructure>` for "``main``" and
179 :ref:`named metadata <namedmetadatastructure>` "``foo``".
181 In general, a module is made up of a list of global values (where both
182 functions and global variables are global values). Global values are
183 represented by a pointer to a memory location (in this case, a pointer
184 to an array of char, and a pointer to a function), and have one of the
185 following :ref:`linkage types <linkage>`.
192 All Global Variables and Functions have one of the following types of
196 Global values with "``private``" linkage are only directly
197 accessible by objects in the current module. In particular, linking
198 code into a module with an private global value may cause the
199 private to be renamed as necessary to avoid collisions. Because the
200 symbol is private to the module, all references can be updated. This
201 doesn't show up in any symbol table in the object file.
203 Similar to private, but the value shows as a local symbol
204 (``STB_LOCAL`` in the case of ELF) in the object file. This
205 corresponds to the notion of the '``static``' keyword in C.
206 ``available_externally``
207 Globals with "``available_externally``" linkage are never emitted
208 into the object file corresponding to the LLVM module. They exist to
209 allow inlining and other optimizations to take place given knowledge
210 of the definition of the global, which is known to be somewhere
211 outside the module. Globals with ``available_externally`` linkage
212 are allowed to be discarded at will, and are otherwise the same as
213 ``linkonce_odr``. This linkage type is only allowed on definitions,
216 Globals with "``linkonce``" linkage are merged with other globals of
217 the same name when linkage occurs. This can be used to implement
218 some forms of inline functions, templates, or other code which must
219 be generated in each translation unit that uses it, but where the
220 body may be overridden with a more definitive definition later.
221 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
222 that ``linkonce`` linkage does not actually allow the optimizer to
223 inline the body of this function into callers because it doesn't
224 know if this definition of the function is the definitive definition
225 within the program or whether it will be overridden by a stronger
226 definition. To enable inlining and other optimizations, use
227 "``linkonce_odr``" linkage.
229 "``weak``" linkage has the same merging semantics as ``linkonce``
230 linkage, except that unreferenced globals with ``weak`` linkage may
231 not be discarded. This is used for globals that are declared "weak"
234 "``common``" linkage is most similar to "``weak``" linkage, but they
235 are used for tentative definitions in C, such as "``int X;``" at
236 global scope. Symbols with "``common``" linkage are merged in the
237 same way as ``weak symbols``, and they may not be deleted if
238 unreferenced. ``common`` symbols may not have an explicit section,
239 must have a zero initializer, and may not be marked
240 ':ref:`constant <globalvars>`'. Functions and aliases may not have
243 .. _linkage_appending:
246 "``appending``" linkage may only be applied to global variables of
247 pointer to array type. When two global variables with appending
248 linkage are linked together, the two global arrays are appended
249 together. This is the LLVM, typesafe, equivalent of having the
250 system linker append together "sections" with identical names when
253 The semantics of this linkage follow the ELF object file model: the
254 symbol is weak until linked, if not linked, the symbol becomes null
255 instead of being an undefined reference.
256 ``linkonce_odr``, ``weak_odr``
257 Some languages allow differing globals to be merged, such as two
258 functions with different semantics. Other languages, such as
259 ``C++``, ensure that only equivalent globals are ever merged (the
260 "one definition rule" --- "ODR"). Such languages can use the
261 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
262 global will only be merged with equivalent globals. These linkage
263 types are otherwise the same as their non-``odr`` versions.
265 If none of the above identifiers are used, the global is externally
266 visible, meaning that it participates in linkage and can be used to
267 resolve external symbol references.
269 It is illegal for a function *declaration* to have any linkage type
270 other than ``external`` or ``extern_weak``.
277 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
278 :ref:`invokes <i_invoke>` can all have an optional calling convention
279 specified for the call. The calling convention of any pair of dynamic
280 caller/callee must match, or the behavior of the program is undefined.
281 The following calling conventions are supported by LLVM, and more may be
284 "``ccc``" - The C calling convention
285 This calling convention (the default if no other calling convention
286 is specified) matches the target C calling conventions. This calling
287 convention supports varargs function calls and tolerates some
288 mismatch in the declared prototype and implemented declaration of
289 the function (as does normal C).
290 "``fastcc``" - The fast calling convention
291 This calling convention attempts to make calls as fast as possible
292 (e.g. by passing things in registers). This calling convention
293 allows the target to use whatever tricks it wants to produce fast
294 code for the target, without having to conform to an externally
295 specified ABI (Application Binary Interface). `Tail calls can only
296 be optimized when this, the GHC or the HiPE convention is
297 used. <CodeGenerator.html#id80>`_ This calling convention does not
298 support varargs and requires the prototype of all callees to exactly
299 match the prototype of the function definition.
300 "``coldcc``" - The cold calling convention
301 This calling convention attempts to make code in the caller as
302 efficient as possible under the assumption that the call is not
303 commonly executed. As such, these calls often preserve all registers
304 so that the call does not break any live ranges in the caller side.
305 This calling convention does not support varargs and requires the
306 prototype of all callees to exactly match the prototype of the
307 function definition. Furthermore the inliner doesn't consider such function
309 "``cc 10``" - GHC convention
310 This calling convention has been implemented specifically for use by
311 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
312 It passes everything in registers, going to extremes to achieve this
313 by disabling callee save registers. This calling convention should
314 not be used lightly but only for specific situations such as an
315 alternative to the *register pinning* performance technique often
316 used when implementing functional programming languages. At the
317 moment only X86 supports this convention and it has the following
320 - On *X86-32* only supports up to 4 bit type parameters. No
321 floating point types are supported.
322 - On *X86-64* only supports up to 10 bit type parameters and 6
323 floating point parameters.
325 This calling convention supports `tail call
326 optimization <CodeGenerator.html#id80>`_ but requires both the
327 caller and callee are using it.
328 "``cc 11``" - The HiPE calling convention
329 This calling convention has been implemented specifically for use by
330 the `High-Performance Erlang
331 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
332 native code compiler of the `Ericsson's Open Source Erlang/OTP
333 system <http://www.erlang.org/download.shtml>`_. It uses more
334 registers for argument passing than the ordinary C calling
335 convention and defines no callee-saved registers. The calling
336 convention properly supports `tail call
337 optimization <CodeGenerator.html#id80>`_ but requires that both the
338 caller and the callee use it. It uses a *register pinning*
339 mechanism, similar to GHC's convention, for keeping frequently
340 accessed runtime components pinned to specific hardware registers.
341 At the moment only X86 supports this convention (both 32 and 64
343 "``webkit_jscc``" - WebKit's JavaScript calling convention
344 This calling convention has been implemented for `WebKit FTL JIT
345 <https://trac.webkit.org/wiki/FTLJIT>`_. It passes arguments on the
346 stack right to left (as cdecl does), and returns a value in the
347 platform's customary return register.
348 "``anyregcc``" - Dynamic calling convention for code patching
349 This is a special convention that supports patching an arbitrary code
350 sequence in place of a call site. This convention forces the call
351 arguments into registers but allows them to be dynamcially
352 allocated. This can currently only be used with calls to
353 llvm.experimental.patchpoint because only this intrinsic records
354 the location of its arguments in a side table. See :doc:`StackMaps`.
355 "``preserve_mostcc``" - The `PreserveMost` calling convention
356 This calling convention attempts to make the code in the caller as little
357 intrusive as possible. This calling convention behaves identical to the `C`
358 calling convention on how arguments and return values are passed, but it
359 uses a different set of caller/callee-saved registers. This alleviates the
360 burden of saving and recovering a large register set before and after the
361 call in the caller. If the arguments are passed in callee-saved registers,
362 then they will be preserved by the callee across the call. This doesn't
363 apply for values returned in callee-saved registers.
365 - On X86-64 the callee preserves all general purpose registers, except for
366 R11. R11 can be used as a scratch register. Floating-point registers
367 (XMMs/YMMs) are not preserved and need to be saved by the caller.
369 The idea behind this convention is to support calls to runtime functions
370 that have a hot path and a cold path. The hot path is usually a small piece
371 of code that doesn't use many registers. The cold path might need to call out to
372 another function and therefore only needs to preserve the caller-saved
373 registers, which haven't already been saved by the caller. The
374 `PreserveMost` calling convention is very similar to the `cold` calling
375 convention in terms of caller/callee-saved registers, but they are used for
376 different types of function calls. `coldcc` is for function calls that are
377 rarely executed, whereas `preserve_mostcc` function calls are intended to be
378 on the hot path and definitely executed a lot. Furthermore `preserve_mostcc`
379 doesn't prevent the inliner from inlining the function call.
381 This calling convention will be used by a future version of the ObjectiveC
382 runtime and should therefore still be considered experimental at this time.
383 Although this convention was created to optimize certain runtime calls to
384 the ObjectiveC runtime, it is not limited to this runtime and might be used
385 by other runtimes in the future too. The current implementation only
386 supports X86-64, but the intention is to support more architectures in the
388 "``preserve_allcc``" - The `PreserveAll` calling convention
389 This calling convention attempts to make the code in the caller even less
390 intrusive than the `PreserveMost` calling convention. This calling
391 convention also behaves identical to the `C` calling convention on how
392 arguments and return values are passed, but it uses a different set of
393 caller/callee-saved registers. This removes the burden of saving and
394 recovering a large register set before and after the call in the caller. If
395 the arguments are passed in callee-saved registers, then they will be
396 preserved by the callee across the call. This doesn't apply for values
397 returned in callee-saved registers.
399 - On X86-64 the callee preserves all general purpose registers, except for
400 R11. R11 can be used as a scratch register. Furthermore it also preserves
401 all floating-point registers (XMMs/YMMs).
403 The idea behind this convention is to support calls to runtime functions
404 that don't need to call out to any other functions.
406 This calling convention, like the `PreserveMost` calling convention, will be
407 used by a future version of the ObjectiveC runtime and should be considered
408 experimental at this time.
409 "``cc <n>``" - Numbered convention
410 Any calling convention may be specified by number, allowing
411 target-specific calling conventions to be used. Target specific
412 calling conventions start at 64.
414 More calling conventions can be added/defined on an as-needed basis, to
415 support Pascal conventions or any other well-known target-independent
418 .. _visibilitystyles:
423 All Global Variables and Functions have one of the following visibility
426 "``default``" - Default style
427 On targets that use the ELF object file format, default visibility
428 means that the declaration is visible to other modules and, in
429 shared libraries, means that the declared entity may be overridden.
430 On Darwin, default visibility means that the declaration is visible
431 to other modules. Default visibility corresponds to "external
432 linkage" in the language.
433 "``hidden``" - Hidden style
434 Two declarations of an object with hidden visibility refer to the
435 same object if they are in the same shared object. Usually, hidden
436 visibility indicates that the symbol will not be placed into the
437 dynamic symbol table, so no other module (executable or shared
438 library) can reference it directly.
439 "``protected``" - Protected style
440 On ELF, protected visibility indicates that the symbol will be
441 placed in the dynamic symbol table, but that references within the
442 defining module will bind to the local symbol. That is, the symbol
443 cannot be overridden by another module.
445 A symbol with ``internal`` or ``private`` linkage must have ``default``
453 All Global Variables, Functions and Aliases can have one of the following
457 "``dllimport``" causes the compiler to reference a function or variable via
458 a global pointer to a pointer that is set up by the DLL exporting the
459 symbol. On Microsoft Windows targets, the pointer name is formed by
460 combining ``__imp_`` and the function or variable name.
462 "``dllexport``" causes the compiler to provide a global pointer to a pointer
463 in a DLL, so that it can be referenced with the ``dllimport`` attribute. On
464 Microsoft Windows targets, the pointer name is formed by combining
465 ``__imp_`` and the function or variable name. Since this storage class
466 exists for defining a dll interface, the compiler, assembler and linker know
467 it is externally referenced and must refrain from deleting the symbol.
471 Thread Local Storage Models
472 ---------------------------
474 A variable may be defined as ``thread_local``, which means that it will
475 not be shared by threads (each thread will have a separated copy of the
476 variable). Not all targets support thread-local variables. Optionally, a
477 TLS model may be specified:
480 For variables that are only used within the current shared library.
482 For variables in modules that will not be loaded dynamically.
484 For variables defined in the executable and only used within it.
486 If no explicit model is given, the "general dynamic" model is used.
488 The models correspond to the ELF TLS models; see `ELF Handling For
489 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
490 more information on under which circumstances the different models may
491 be used. The target may choose a different TLS model if the specified
492 model is not supported, or if a better choice of model can be made.
494 A model can also be specified in a alias, but then it only governs how
495 the alias is accessed. It will not have any effect in the aliasee.
502 LLVM IR allows you to specify both "identified" and "literal" :ref:`structure
503 types <t_struct>`. Literal types are uniqued structurally, but identified types
504 are never uniqued. An :ref:`opaque structural type <t_opaque>` can also be used
505 to forward declare a type that is not yet available.
507 An example of a identified structure specification is:
511 %mytype = type { %mytype*, i32 }
513 Prior to the LLVM 3.0 release, identified types were structurally uniqued. Only
514 literal types are uniqued in recent versions of LLVM.
521 Global variables define regions of memory allocated at compilation time
524 Global variable definitions must be initialized.
526 Global variables in other translation units can also be declared, in which
527 case they don't have an initializer.
529 Either global variable definitions or declarations may have an explicit section
530 to be placed in and may have an optional explicit alignment specified.
532 A variable may be defined as a global ``constant``, which indicates that
533 the contents of the variable will **never** be modified (enabling better
534 optimization, allowing the global data to be placed in the read-only
535 section of an executable, etc). Note that variables that need runtime
536 initialization cannot be marked ``constant`` as there is a store to the
539 LLVM explicitly allows *declarations* of global variables to be marked
540 constant, even if the final definition of the global is not. This
541 capability can be used to enable slightly better optimization of the
542 program, but requires the language definition to guarantee that
543 optimizations based on the 'constantness' are valid for the translation
544 units that do not include the definition.
546 As SSA values, global variables define pointer values that are in scope
547 (i.e. they dominate) all basic blocks in the program. Global variables
548 always define a pointer to their "content" type because they describe a
549 region of memory, and all memory objects in LLVM are accessed through
552 Global variables can be marked with ``unnamed_addr`` which indicates
553 that the address is not significant, only the content. Constants marked
554 like this can be merged with other constants if they have the same
555 initializer. Note that a constant with significant address *can* be
556 merged with a ``unnamed_addr`` constant, the result being a constant
557 whose address is significant.
559 A global variable may be declared to reside in a target-specific
560 numbered address space. For targets that support them, address spaces
561 may affect how optimizations are performed and/or what target
562 instructions are used to access the variable. The default address space
563 is zero. The address space qualifier must precede any other attributes.
565 LLVM allows an explicit section to be specified for globals. If the
566 target supports it, it will emit globals to the section specified.
567 Additionally, the global can placed in a comdat if the target has the necessary
570 By default, global initializers are optimized by assuming that global
571 variables defined within the module are not modified from their
572 initial values before the start of the global initializer. This is
573 true even for variables potentially accessible from outside the
574 module, including those with external linkage or appearing in
575 ``@llvm.used`` or dllexported variables. This assumption may be suppressed
576 by marking the variable with ``externally_initialized``.
578 An explicit alignment may be specified for a global, which must be a
579 power of 2. If not present, or if the alignment is set to zero, the
580 alignment of the global is set by the target to whatever it feels
581 convenient. If an explicit alignment is specified, the global is forced
582 to have exactly that alignment. Targets and optimizers are not allowed
583 to over-align the global if the global has an assigned section. In this
584 case, the extra alignment could be observable: for example, code could
585 assume that the globals are densely packed in their section and try to
586 iterate over them as an array, alignment padding would break this
587 iteration. The maximum alignment is ``1 << 29``.
589 Globals can also have a :ref:`DLL storage class <dllstorageclass>`.
591 Variables and aliases can have a
592 :ref:`Thread Local Storage Model <tls_model>`.
596 [@<GlobalVarName> =] [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal]
597 [unnamed_addr] [AddrSpace] [ExternallyInitialized]
598 <global | constant> <Type> [<InitializerConstant>]
599 [, section "name"] [, comdat [($name)]]
600 [, align <Alignment>]
602 For example, the following defines a global in a numbered address space
603 with an initializer, section, and alignment:
607 @G = addrspace(5) constant float 1.0, section "foo", align 4
609 The following example just declares a global variable
613 @G = external global i32
615 The following example defines a thread-local global with the
616 ``initialexec`` TLS model:
620 @G = thread_local(initialexec) global i32 0, align 4
622 .. _functionstructure:
627 LLVM function definitions consist of the "``define``" keyword, an
628 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
629 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
630 an optional :ref:`calling convention <callingconv>`,
631 an optional ``unnamed_addr`` attribute, a return type, an optional
632 :ref:`parameter attribute <paramattrs>` for the return type, a function
633 name, a (possibly empty) argument list (each with optional :ref:`parameter
634 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
635 an optional section, an optional alignment,
636 an optional :ref:`comdat <langref_comdats>`,
637 an optional :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
638 an optional :ref:`prologue <prologuedata>`, an opening
639 curly brace, a list of basic blocks, and a closing curly brace.
641 LLVM function declarations consist of the "``declare``" keyword, an
642 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
643 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
644 an optional :ref:`calling convention <callingconv>`,
645 an optional ``unnamed_addr`` attribute, a return type, an optional
646 :ref:`parameter attribute <paramattrs>` for the return type, a function
647 name, a possibly empty list of arguments, an optional alignment, an optional
648 :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
649 and an optional :ref:`prologue <prologuedata>`.
651 A function definition contains a list of basic blocks, forming the CFG (Control
652 Flow Graph) for the function. Each basic block may optionally start with a label
653 (giving the basic block a symbol table entry), contains a list of instructions,
654 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
655 function return). If an explicit label is not provided, a block is assigned an
656 implicit numbered label, using the next value from the same counter as used for
657 unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
658 entry block does not have an explicit label, it will be assigned label "%0",
659 then the first unnamed temporary in that block will be "%1", etc.
661 The first basic block in a function is special in two ways: it is
662 immediately executed on entrance to the function, and it is not allowed
663 to have predecessor basic blocks (i.e. there can not be any branches to
664 the entry block of a function). Because the block can have no
665 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
667 LLVM allows an explicit section to be specified for functions. If the
668 target supports it, it will emit functions to the section specified.
669 Additionally, the function can be placed in a COMDAT.
671 An explicit alignment may be specified for a function. If not present,
672 or if the alignment is set to zero, the alignment of the function is set
673 by the target to whatever it feels convenient. If an explicit alignment
674 is specified, the function is forced to have at least that much
675 alignment. All alignments must be a power of 2.
677 If the ``unnamed_addr`` attribute is given, the address is known to not
678 be significant and two identical functions can be merged.
682 define [linkage] [visibility] [DLLStorageClass]
684 <ResultType> @<FunctionName> ([argument list])
685 [unnamed_addr] [fn Attrs] [section "name"] [comdat [($name)]]
686 [align N] [gc] [prefix Constant] [prologue Constant] { ... }
688 The argument list is a comma seperated sequence of arguments where each
689 argument is of the following form
693 <type> [parameter Attrs] [name]
701 Aliases, unlike function or variables, don't create any new data. They
702 are just a new symbol and metadata for an existing position.
704 Aliases have a name and an aliasee that is either a global value or a
707 Aliases may have an optional :ref:`linkage type <linkage>`, an optional
708 :ref:`visibility style <visibility>`, an optional :ref:`DLL storage class
709 <dllstorageclass>` and an optional :ref:`tls model <tls_model>`.
713 @<Name> = [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal] [unnamed_addr] alias <AliaseeTy> @<Aliasee>
715 The linkage must be one of ``private``, ``internal``, ``linkonce``, ``weak``,
716 ``linkonce_odr``, ``weak_odr``, ``external``. Note that some system linkers
717 might not correctly handle dropping a weak symbol that is aliased.
719 Aliases that are not ``unnamed_addr`` are guaranteed to have the same address as
720 the aliasee expression. ``unnamed_addr`` ones are only guaranteed to point
723 Since aliases are only a second name, some restrictions apply, of which
724 some can only be checked when producing an object file:
726 * The expression defining the aliasee must be computable at assembly
727 time. Since it is just a name, no relocations can be used.
729 * No alias in the expression can be weak as the possibility of the
730 intermediate alias being overridden cannot be represented in an
733 * No global value in the expression can be a declaration, since that
734 would require a relocation, which is not possible.
741 Comdat IR provides access to COFF and ELF object file COMDAT functionality.
743 Comdats have a name which represents the COMDAT key. All global objects that
744 specify this key will only end up in the final object file if the linker chooses
745 that key over some other key. Aliases are placed in the same COMDAT that their
746 aliasee computes to, if any.
748 Comdats have a selection kind to provide input on how the linker should
749 choose between keys in two different object files.
753 $<Name> = comdat SelectionKind
755 The selection kind must be one of the following:
758 The linker may choose any COMDAT key, the choice is arbitrary.
760 The linker may choose any COMDAT key but the sections must contain the
763 The linker will choose the section containing the largest COMDAT key.
765 The linker requires that only section with this COMDAT key exist.
767 The linker may choose any COMDAT key but the sections must contain the
770 Note that the Mach-O platform doesn't support COMDATs and ELF only supports
771 ``any`` as a selection kind.
773 Here is an example of a COMDAT group where a function will only be selected if
774 the COMDAT key's section is the largest:
778 $foo = comdat largest
779 @foo = global i32 2, comdat($foo)
781 define void @bar() comdat($foo) {
785 As a syntactic sugar the ``$name`` can be omitted if the name is the same as
791 @foo = global i32 2, comdat
794 In a COFF object file, this will create a COMDAT section with selection kind
795 ``IMAGE_COMDAT_SELECT_LARGEST`` containing the contents of the ``@foo`` symbol
796 and another COMDAT section with selection kind
797 ``IMAGE_COMDAT_SELECT_ASSOCIATIVE`` which is associated with the first COMDAT
798 section and contains the contents of the ``@bar`` symbol.
800 There are some restrictions on the properties of the global object.
801 It, or an alias to it, must have the same name as the COMDAT group when
803 The contents and size of this object may be used during link-time to determine
804 which COMDAT groups get selected depending on the selection kind.
805 Because the name of the object must match the name of the COMDAT group, the
806 linkage of the global object must not be local; local symbols can get renamed
807 if a collision occurs in the symbol table.
809 The combined use of COMDATS and section attributes may yield surprising results.
816 @g1 = global i32 42, section "sec", comdat($foo)
817 @g2 = global i32 42, section "sec", comdat($bar)
819 From the object file perspective, this requires the creation of two sections
820 with the same name. This is necessary because both globals belong to different
821 COMDAT groups and COMDATs, at the object file level, are represented by
824 Note that certain IR constructs like global variables and functions may create
825 COMDATs in the object file in addition to any which are specified using COMDAT
826 IR. This arises, for example, when a global variable has linkonce_odr linkage.
828 .. _namedmetadatastructure:
833 Named metadata is a collection of metadata. :ref:`Metadata
834 nodes <metadata>` (but not metadata strings) are the only valid
835 operands for a named metadata.
839 ; Some unnamed metadata nodes, which are referenced by the named metadata.
844 !name = !{!0, !1, !2}
851 The return type and each parameter of a function type may have a set of
852 *parameter attributes* associated with them. Parameter attributes are
853 used to communicate additional information about the result or
854 parameters of a function. Parameter attributes are considered to be part
855 of the function, not of the function type, so functions with different
856 parameter attributes can have the same function type.
858 Parameter attributes are simple keywords that follow the type specified.
859 If multiple parameter attributes are needed, they are space separated.
864 declare i32 @printf(i8* noalias nocapture, ...)
865 declare i32 @atoi(i8 zeroext)
866 declare signext i8 @returns_signed_char()
868 Note that any attributes for the function result (``nounwind``,
869 ``readonly``) come immediately after the argument list.
871 Currently, only the following parameter attributes are defined:
874 This indicates to the code generator that the parameter or return
875 value should be zero-extended to the extent required by the target's
876 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
877 the caller (for a parameter) or the callee (for a return value).
879 This indicates to the code generator that the parameter or return
880 value should be sign-extended to the extent required by the target's
881 ABI (which is usually 32-bits) by the caller (for a parameter) or
882 the callee (for a return value).
884 This indicates that this parameter or return value should be treated
885 in a special target-dependent fashion during while emitting code for
886 a function call or return (usually, by putting it in a register as
887 opposed to memory, though some targets use it to distinguish between
888 two different kinds of registers). Use of this attribute is
891 This indicates that the pointer parameter should really be passed by
892 value to the function. The attribute implies that a hidden copy of
893 the pointee is made between the caller and the callee, so the callee
894 is unable to modify the value in the caller. This attribute is only
895 valid on LLVM pointer arguments. It is generally used to pass
896 structs and arrays by value, but is also valid on pointers to
897 scalars. The copy is considered to belong to the caller not the
898 callee (for example, ``readonly`` functions should not write to
899 ``byval`` parameters). This is not a valid attribute for return
902 The byval attribute also supports specifying an alignment with the
903 align attribute. It indicates the alignment of the stack slot to
904 form and the known alignment of the pointer specified to the call
905 site. If the alignment is not specified, then the code generator
906 makes a target-specific assumption.
912 The ``inalloca`` argument attribute allows the caller to take the
913 address of outgoing stack arguments. An ``inalloca`` argument must
914 be a pointer to stack memory produced by an ``alloca`` instruction.
915 The alloca, or argument allocation, must also be tagged with the
916 inalloca keyword. Only the last argument may have the ``inalloca``
917 attribute, and that argument is guaranteed to be passed in memory.
919 An argument allocation may be used by a call at most once because
920 the call may deallocate it. The ``inalloca`` attribute cannot be
921 used in conjunction with other attributes that affect argument
922 storage, like ``inreg``, ``nest``, ``sret``, or ``byval``. The
923 ``inalloca`` attribute also disables LLVM's implicit lowering of
924 large aggregate return values, which means that frontend authors
925 must lower them with ``sret`` pointers.
927 When the call site is reached, the argument allocation must have
928 been the most recent stack allocation that is still live, or the
929 results are undefined. It is possible to allocate additional stack
930 space after an argument allocation and before its call site, but it
931 must be cleared off with :ref:`llvm.stackrestore
934 See :doc:`InAlloca` for more information on how to use this
938 This indicates that the pointer parameter specifies the address of a
939 structure that is the return value of the function in the source
940 program. This pointer must be guaranteed by the caller to be valid:
941 loads and stores to the structure may be assumed by the callee
942 not to trap and to be properly aligned. This may only be applied to
943 the first parameter. This is not a valid attribute for return
947 This indicates that the pointer value may be assumed by the optimizer to
948 have the specified alignment.
950 Note that this attribute has additional semantics when combined with the
956 This indicates that objects accessed via pointer values
957 :ref:`based <pointeraliasing>` on the argument or return value are not also
958 accessed, during the execution of the function, via pointer values not
959 *based* on the argument or return value. The attribute on a return value
960 also has additional semantics described below. The caller shares the
961 responsibility with the callee for ensuring that these requirements are met.
962 For further details, please see the discussion of the NoAlias response in
963 :ref:`alias analysis <Must, May, or No>`.
965 Note that this definition of ``noalias`` is intentionally similar
966 to the definition of ``restrict`` in C99 for function arguments.
968 For function return values, C99's ``restrict`` is not meaningful,
969 while LLVM's ``noalias`` is. Furthermore, the semantics of the ``noalias``
970 attribute on return values are stronger than the semantics of the attribute
971 when used on function arguments. On function return values, the ``noalias``
972 attribute indicates that the function acts like a system memory allocation
973 function, returning a pointer to allocated storage disjoint from the
974 storage for any other object accessible to the caller.
977 This indicates that the callee does not make any copies of the
978 pointer that outlive the callee itself. This is not a valid
979 attribute for return values.
984 This indicates that the pointer parameter can be excised using the
985 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
986 attribute for return values and can only be applied to one parameter.
989 This indicates that the function always returns the argument as its return
990 value. This is an optimization hint to the code generator when generating
991 the caller, allowing tail call optimization and omission of register saves
992 and restores in some cases; it is not checked or enforced when generating
993 the callee. The parameter and the function return type must be valid
994 operands for the :ref:`bitcast instruction <i_bitcast>`. This is not a
995 valid attribute for return values and can only be applied to one parameter.
998 This indicates that the parameter or return pointer is not null. This
999 attribute may only be applied to pointer typed parameters. This is not
1000 checked or enforced by LLVM, the caller must ensure that the pointer
1001 passed in is non-null, or the callee must ensure that the returned pointer
1004 ``dereferenceable(<n>)``
1005 This indicates that the parameter or return pointer is dereferenceable. This
1006 attribute may only be applied to pointer typed parameters. A pointer that
1007 is dereferenceable can be loaded from speculatively without a risk of
1008 trapping. The number of bytes known to be dereferenceable must be provided
1009 in parentheses. It is legal for the number of bytes to be less than the
1010 size of the pointee type. The ``nonnull`` attribute does not imply
1011 dereferenceability (consider a pointer to one element past the end of an
1012 array), however ``dereferenceable(<n>)`` does imply ``nonnull`` in
1013 ``addrspace(0)`` (which is the default address space).
1017 Garbage Collector Strategy Names
1018 --------------------------------
1020 Each function may specify a garbage collector strategy name, which is simply a
1023 .. code-block:: llvm
1025 define void @f() gc "name" { ... }
1027 The supported values of *name* includes those :ref:`built in to LLVM
1028 <builtin-gc-strategies>` and any provided by loaded plugins. Specifying a GC
1029 strategy will cause the compiler to alter its output in order to support the
1030 named garbage collection algorithm. Note that LLVM itself does not contain a
1031 garbage collector, this functionality is restricted to generating machine code
1032 which can interoperate with a collector provided externally.
1039 Prefix data is data associated with a function which the code
1040 generator will emit immediately before the function's entrypoint.
1041 The purpose of this feature is to allow frontends to associate
1042 language-specific runtime metadata with specific functions and make it
1043 available through the function pointer while still allowing the
1044 function pointer to be called.
1046 To access the data for a given function, a program may bitcast the
1047 function pointer to a pointer to the constant's type and dereference
1048 index -1. This implies that the IR symbol points just past the end of
1049 the prefix data. For instance, take the example of a function annotated
1050 with a single ``i32``,
1052 .. code-block:: llvm
1054 define void @f() prefix i32 123 { ... }
1056 The prefix data can be referenced as,
1058 .. code-block:: llvm
1060 %0 = bitcast void* () @f to i32*
1061 %a = getelementptr inbounds i32, i32* %0, i32 -1
1062 %b = load i32, i32* %a
1064 Prefix data is laid out as if it were an initializer for a global variable
1065 of the prefix data's type. The function will be placed such that the
1066 beginning of the prefix data is aligned. This means that if the size
1067 of the prefix data is not a multiple of the alignment size, the
1068 function's entrypoint will not be aligned. If alignment of the
1069 function's entrypoint is desired, padding must be added to the prefix
1072 A function may have prefix data but no body. This has similar semantics
1073 to the ``available_externally`` linkage in that the data may be used by the
1074 optimizers but will not be emitted in the object file.
1081 The ``prologue`` attribute allows arbitrary code (encoded as bytes) to
1082 be inserted prior to the function body. This can be used for enabling
1083 function hot-patching and instrumentation.
1085 To maintain the semantics of ordinary function calls, the prologue data must
1086 have a particular format. Specifically, it must begin with a sequence of
1087 bytes which decode to a sequence of machine instructions, valid for the
1088 module's target, which transfer control to the point immediately succeeding
1089 the prologue data, without performing any other visible action. This allows
1090 the inliner and other passes to reason about the semantics of the function
1091 definition without needing to reason about the prologue data. Obviously this
1092 makes the format of the prologue data highly target dependent.
1094 A trivial example of valid prologue data for the x86 architecture is ``i8 144``,
1095 which encodes the ``nop`` instruction:
1097 .. code-block:: llvm
1099 define void @f() prologue i8 144 { ... }
1101 Generally prologue data can be formed by encoding a relative branch instruction
1102 which skips the metadata, as in this example of valid prologue data for the
1103 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
1105 .. code-block:: llvm
1107 %0 = type <{ i8, i8, i8* }>
1109 define void @f() prologue %0 <{ i8 235, i8 8, i8* @md}> { ... }
1111 A function may have prologue data but no body. This has similar semantics
1112 to the ``available_externally`` linkage in that the data may be used by the
1113 optimizers but will not be emitted in the object file.
1120 Attribute groups are groups of attributes that are referenced by objects within
1121 the IR. They are important for keeping ``.ll`` files readable, because a lot of
1122 functions will use the same set of attributes. In the degenerative case of a
1123 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
1124 group will capture the important command line flags used to build that file.
1126 An attribute group is a module-level object. To use an attribute group, an
1127 object references the attribute group's ID (e.g. ``#37``). An object may refer
1128 to more than one attribute group. In that situation, the attributes from the
1129 different groups are merged.
1131 Here is an example of attribute groups for a function that should always be
1132 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
1134 .. code-block:: llvm
1136 ; Target-independent attributes:
1137 attributes #0 = { alwaysinline alignstack=4 }
1139 ; Target-dependent attributes:
1140 attributes #1 = { "no-sse" }
1142 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
1143 define void @f() #0 #1 { ... }
1150 Function attributes are set to communicate additional information about
1151 a function. Function attributes are considered to be part of the
1152 function, not of the function type, so functions with different function
1153 attributes can have the same function type.
1155 Function attributes are simple keywords that follow the type specified.
1156 If multiple attributes are needed, they are space separated. For
1159 .. code-block:: llvm
1161 define void @f() noinline { ... }
1162 define void @f() alwaysinline { ... }
1163 define void @f() alwaysinline optsize { ... }
1164 define void @f() optsize { ... }
1167 This attribute indicates that, when emitting the prologue and
1168 epilogue, the backend should forcibly align the stack pointer.
1169 Specify the desired alignment, which must be a power of two, in
1172 This attribute indicates that the inliner should attempt to inline
1173 this function into callers whenever possible, ignoring any active
1174 inlining size threshold for this caller.
1176 This indicates that the callee function at a call site should be
1177 recognized as a built-in function, even though the function's declaration
1178 uses the ``nobuiltin`` attribute. This is only valid at call sites for
1179 direct calls to functions that are declared with the ``nobuiltin``
1182 This attribute indicates that this function is rarely called. When
1183 computing edge weights, basic blocks post-dominated by a cold
1184 function call are also considered to be cold; and, thus, given low
1187 This attribute indicates that the source code contained a hint that
1188 inlining this function is desirable (such as the "inline" keyword in
1189 C/C++). It is just a hint; it imposes no requirements on the
1192 This attribute indicates that the function should be added to a
1193 jump-instruction table at code-generation time, and that all address-taken
1194 references to this function should be replaced with a reference to the
1195 appropriate jump-instruction-table function pointer. Note that this creates
1196 a new pointer for the original function, which means that code that depends
1197 on function-pointer identity can break. So, any function annotated with
1198 ``jumptable`` must also be ``unnamed_addr``.
1200 This attribute suggests that optimization passes and code generator
1201 passes make choices that keep the code size of this function as small
1202 as possible and perform optimizations that may sacrifice runtime
1203 performance in order to minimize the size of the generated code.
1205 This attribute disables prologue / epilogue emission for the
1206 function. This can have very system-specific consequences.
1208 This indicates that the callee function at a call site is not recognized as
1209 a built-in function. LLVM will retain the original call and not replace it
1210 with equivalent code based on the semantics of the built-in function, unless
1211 the call site uses the ``builtin`` attribute. This is valid at call sites
1212 and on function declarations and definitions.
1214 This attribute indicates that calls to the function cannot be
1215 duplicated. A call to a ``noduplicate`` function may be moved
1216 within its parent function, but may not be duplicated within
1217 its parent function.
1219 A function containing a ``noduplicate`` call may still
1220 be an inlining candidate, provided that the call is not
1221 duplicated by inlining. That implies that the function has
1222 internal linkage and only has one call site, so the original
1223 call is dead after inlining.
1225 This attributes disables implicit floating point instructions.
1227 This attribute indicates that the inliner should never inline this
1228 function in any situation. This attribute may not be used together
1229 with the ``alwaysinline`` attribute.
1231 This attribute suppresses lazy symbol binding for the function. This
1232 may make calls to the function faster, at the cost of extra program
1233 startup time if the function is not called during program startup.
1235 This attribute indicates that the code generator should not use a
1236 red zone, even if the target-specific ABI normally permits it.
1238 This function attribute indicates that the function never returns
1239 normally. This produces undefined behavior at runtime if the
1240 function ever does dynamically return.
1242 This function attribute indicates that the function never raises an
1243 exception. If the function does raise an exception, its runtime
1244 behavior is undefined. However, functions marked nounwind may still
1245 trap or generate asynchronous exceptions. Exception handling schemes
1246 that are recognized by LLVM to handle asynchronous exceptions, such
1247 as SEH, will still provide their implementation defined semantics.
1249 This function attribute indicates that the function is not optimized
1250 by any optimization or code generator passes with the
1251 exception of interprocedural optimization passes.
1252 This attribute cannot be used together with the ``alwaysinline``
1253 attribute; this attribute is also incompatible
1254 with the ``minsize`` attribute and the ``optsize`` attribute.
1256 This attribute requires the ``noinline`` attribute to be specified on
1257 the function as well, so the function is never inlined into any caller.
1258 Only functions with the ``alwaysinline`` attribute are valid
1259 candidates for inlining into the body of this function.
1261 This attribute suggests that optimization passes and code generator
1262 passes make choices that keep the code size of this function low,
1263 and otherwise do optimizations specifically to reduce code size as
1264 long as they do not significantly impact runtime performance.
1266 On a function, this attribute indicates that the function computes its
1267 result (or decides to unwind an exception) based strictly on its arguments,
1268 without dereferencing any pointer arguments or otherwise accessing
1269 any mutable state (e.g. memory, control registers, etc) visible to
1270 caller functions. It does not write through any pointer arguments
1271 (including ``byval`` arguments) and never changes any state visible
1272 to callers. This means that it cannot unwind exceptions by calling
1273 the ``C++`` exception throwing methods.
1275 On an argument, this attribute indicates that the function does not
1276 dereference that pointer argument, even though it may read or write the
1277 memory that the pointer points to if accessed through other pointers.
1279 On a function, this attribute indicates that the function does not write
1280 through any pointer arguments (including ``byval`` arguments) or otherwise
1281 modify any state (e.g. memory, control registers, etc) visible to
1282 caller functions. It may dereference pointer arguments and read
1283 state that may be set in the caller. A readonly function always
1284 returns the same value (or unwinds an exception identically) when
1285 called with the same set of arguments and global state. It cannot
1286 unwind an exception by calling the ``C++`` exception throwing
1289 On an argument, this attribute indicates that the function does not write
1290 through this pointer argument, even though it may write to the memory that
1291 the pointer points to.
1293 This attribute indicates that this function can return twice. The C
1294 ``setjmp`` is an example of such a function. The compiler disables
1295 some optimizations (like tail calls) in the caller of these
1297 ``sanitize_address``
1298 This attribute indicates that AddressSanitizer checks
1299 (dynamic address safety analysis) are enabled for this function.
1301 This attribute indicates that MemorySanitizer checks (dynamic detection
1302 of accesses to uninitialized memory) are enabled for this function.
1304 This attribute indicates that ThreadSanitizer checks
1305 (dynamic thread safety analysis) are enabled for this function.
1307 This attribute indicates that the function should emit a stack
1308 smashing protector. It is in the form of a "canary" --- a random value
1309 placed on the stack before the local variables that's checked upon
1310 return from the function to see if it has been overwritten. A
1311 heuristic is used to determine if a function needs stack protectors
1312 or not. The heuristic used will enable protectors for functions with:
1314 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1315 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1316 - Calls to alloca() with variable sizes or constant sizes greater than
1317 ``ssp-buffer-size``.
1319 Variables that are identified as requiring a protector will be arranged
1320 on the stack such that they are adjacent to the stack protector guard.
1322 If a function that has an ``ssp`` attribute is inlined into a
1323 function that doesn't have an ``ssp`` attribute, then the resulting
1324 function will have an ``ssp`` attribute.
1326 This attribute indicates that the function should *always* emit a
1327 stack smashing protector. This overrides the ``ssp`` function
1330 Variables that are identified as requiring a protector will be arranged
1331 on the stack such that they are adjacent to the stack protector guard.
1332 The specific layout rules are:
1334 #. Large arrays and structures containing large arrays
1335 (``>= ssp-buffer-size``) are closest to the stack protector.
1336 #. Small arrays and structures containing small arrays
1337 (``< ssp-buffer-size``) are 2nd closest to the protector.
1338 #. Variables that have had their address taken are 3rd closest to the
1341 If a function that has an ``sspreq`` attribute is inlined into a
1342 function that doesn't have an ``sspreq`` attribute or which has an
1343 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1344 an ``sspreq`` attribute.
1346 This attribute indicates that the function should emit a stack smashing
1347 protector. This attribute causes a strong heuristic to be used when
1348 determining if a function needs stack protectors. The strong heuristic
1349 will enable protectors for functions with:
1351 - Arrays of any size and type
1352 - Aggregates containing an array of any size and type.
1353 - Calls to alloca().
1354 - Local variables that have had their address taken.
1356 Variables that are identified as requiring a protector will be arranged
1357 on the stack such that they are adjacent to the stack protector guard.
1358 The specific layout rules are:
1360 #. Large arrays and structures containing large arrays
1361 (``>= ssp-buffer-size``) are closest to the stack protector.
1362 #. Small arrays and structures containing small arrays
1363 (``< ssp-buffer-size``) are 2nd closest to the protector.
1364 #. Variables that have had their address taken are 3rd closest to the
1367 This overrides the ``ssp`` function attribute.
1369 If a function that has an ``sspstrong`` attribute is inlined into a
1370 function that doesn't have an ``sspstrong`` attribute, then the
1371 resulting function will have an ``sspstrong`` attribute.
1373 This attribute indicates that the function will delegate to some other
1374 function with a tail call. The prototype of a thunk should not be used for
1375 optimization purposes. The caller is expected to cast the thunk prototype to
1376 match the thunk target prototype.
1378 This attribute indicates that the ABI being targeted requires that
1379 an unwind table entry be produce for this function even if we can
1380 show that no exceptions passes by it. This is normally the case for
1381 the ELF x86-64 abi, but it can be disabled for some compilation
1386 Module-Level Inline Assembly
1387 ----------------------------
1389 Modules may contain "module-level inline asm" blocks, which corresponds
1390 to the GCC "file scope inline asm" blocks. These blocks are internally
1391 concatenated by LLVM and treated as a single unit, but may be separated
1392 in the ``.ll`` file if desired. The syntax is very simple:
1394 .. code-block:: llvm
1396 module asm "inline asm code goes here"
1397 module asm "more can go here"
1399 The strings can contain any character by escaping non-printable
1400 characters. The escape sequence used is simply "\\xx" where "xx" is the
1401 two digit hex code for the number.
1403 The inline asm code is simply printed to the machine code .s file when
1404 assembly code is generated.
1406 .. _langref_datalayout:
1411 A module may specify a target specific data layout string that specifies
1412 how data is to be laid out in memory. The syntax for the data layout is
1415 .. code-block:: llvm
1417 target datalayout = "layout specification"
1419 The *layout specification* consists of a list of specifications
1420 separated by the minus sign character ('-'). Each specification starts
1421 with a letter and may include other information after the letter to
1422 define some aspect of the data layout. The specifications accepted are
1426 Specifies that the target lays out data in big-endian form. That is,
1427 the bits with the most significance have the lowest address
1430 Specifies that the target lays out data in little-endian form. That
1431 is, the bits with the least significance have the lowest address
1434 Specifies the natural alignment of the stack in bits. Alignment
1435 promotion of stack variables is limited to the natural stack
1436 alignment to avoid dynamic stack realignment. The stack alignment
1437 must be a multiple of 8-bits. If omitted, the natural stack
1438 alignment defaults to "unspecified", which does not prevent any
1439 alignment promotions.
1440 ``p[n]:<size>:<abi>:<pref>``
1441 This specifies the *size* of a pointer and its ``<abi>`` and
1442 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1443 bits. The address space, ``n`` is optional, and if not specified,
1444 denotes the default address space 0. The value of ``n`` must be
1445 in the range [1,2^23).
1446 ``i<size>:<abi>:<pref>``
1447 This specifies the alignment for an integer type of a given bit
1448 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1449 ``v<size>:<abi>:<pref>``
1450 This specifies the alignment for a vector type of a given bit
1452 ``f<size>:<abi>:<pref>``
1453 This specifies the alignment for a floating point type of a given bit
1454 ``<size>``. Only values of ``<size>`` that are supported by the target
1455 will work. 32 (float) and 64 (double) are supported on all targets; 80
1456 or 128 (different flavors of long double) are also supported on some
1459 This specifies the alignment for an object of aggregate type.
1461 If present, specifies that llvm names are mangled in the output. The
1464 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1465 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1466 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1467 symbols get a ``_`` prefix.
1468 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1469 functions also get a suffix based on the frame size.
1470 ``n<size1>:<size2>:<size3>...``
1471 This specifies a set of native integer widths for the target CPU in
1472 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1473 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1474 this set are considered to support most general arithmetic operations
1477 On every specification that takes a ``<abi>:<pref>``, specifying the
1478 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1479 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1481 When constructing the data layout for a given target, LLVM starts with a
1482 default set of specifications which are then (possibly) overridden by
1483 the specifications in the ``datalayout`` keyword. The default
1484 specifications are given in this list:
1486 - ``E`` - big endian
1487 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1488 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1489 same as the default address space.
1490 - ``S0`` - natural stack alignment is unspecified
1491 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1492 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1493 - ``i16:16:16`` - i16 is 16-bit aligned
1494 - ``i32:32:32`` - i32 is 32-bit aligned
1495 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1496 alignment of 64-bits
1497 - ``f16:16:16`` - half is 16-bit aligned
1498 - ``f32:32:32`` - float is 32-bit aligned
1499 - ``f64:64:64`` - double is 64-bit aligned
1500 - ``f128:128:128`` - quad is 128-bit aligned
1501 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1502 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1503 - ``a:0:64`` - aggregates are 64-bit aligned
1505 When LLVM is determining the alignment for a given type, it uses the
1508 #. If the type sought is an exact match for one of the specifications,
1509 that specification is used.
1510 #. If no match is found, and the type sought is an integer type, then
1511 the smallest integer type that is larger than the bitwidth of the
1512 sought type is used. If none of the specifications are larger than
1513 the bitwidth then the largest integer type is used. For example,
1514 given the default specifications above, the i7 type will use the
1515 alignment of i8 (next largest) while both i65 and i256 will use the
1516 alignment of i64 (largest specified).
1517 #. If no match is found, and the type sought is a vector type, then the
1518 largest vector type that is smaller than the sought vector type will
1519 be used as a fall back. This happens because <128 x double> can be
1520 implemented in terms of 64 <2 x double>, for example.
1522 The function of the data layout string may not be what you expect.
1523 Notably, this is not a specification from the frontend of what alignment
1524 the code generator should use.
1526 Instead, if specified, the target data layout is required to match what
1527 the ultimate *code generator* expects. This string is used by the
1528 mid-level optimizers to improve code, and this only works if it matches
1529 what the ultimate code generator uses. There is no way to generate IR
1530 that does not embed this target-specific detail into the IR. If you
1531 don't specify the string, the default specifications will be used to
1532 generate a Data Layout and the optimization phases will operate
1533 accordingly and introduce target specificity into the IR with respect to
1534 these default specifications.
1541 A module may specify a target triple string that describes the target
1542 host. The syntax for the target triple is simply:
1544 .. code-block:: llvm
1546 target triple = "x86_64-apple-macosx10.7.0"
1548 The *target triple* string consists of a series of identifiers delimited
1549 by the minus sign character ('-'). The canonical forms are:
1553 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1554 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1556 This information is passed along to the backend so that it generates
1557 code for the proper architecture. It's possible to override this on the
1558 command line with the ``-mtriple`` command line option.
1560 .. _pointeraliasing:
1562 Pointer Aliasing Rules
1563 ----------------------
1565 Any memory access must be done through a pointer value associated with
1566 an address range of the memory access, otherwise the behavior is
1567 undefined. Pointer values are associated with address ranges according
1568 to the following rules:
1570 - A pointer value is associated with the addresses associated with any
1571 value it is *based* on.
1572 - An address of a global variable is associated with the address range
1573 of the variable's storage.
1574 - The result value of an allocation instruction is associated with the
1575 address range of the allocated storage.
1576 - A null pointer in the default address-space is associated with no
1578 - An integer constant other than zero or a pointer value returned from
1579 a function not defined within LLVM may be associated with address
1580 ranges allocated through mechanisms other than those provided by
1581 LLVM. Such ranges shall not overlap with any ranges of addresses
1582 allocated by mechanisms provided by LLVM.
1584 A pointer value is *based* on another pointer value according to the
1587 - A pointer value formed from a ``getelementptr`` operation is *based*
1588 on the first value operand of the ``getelementptr``.
1589 - The result value of a ``bitcast`` is *based* on the operand of the
1591 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1592 values that contribute (directly or indirectly) to the computation of
1593 the pointer's value.
1594 - The "*based* on" relationship is transitive.
1596 Note that this definition of *"based"* is intentionally similar to the
1597 definition of *"based"* in C99, though it is slightly weaker.
1599 LLVM IR does not associate types with memory. The result type of a
1600 ``load`` merely indicates the size and alignment of the memory from
1601 which to load, as well as the interpretation of the value. The first
1602 operand type of a ``store`` similarly only indicates the size and
1603 alignment of the store.
1605 Consequently, type-based alias analysis, aka TBAA, aka
1606 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1607 :ref:`Metadata <metadata>` may be used to encode additional information
1608 which specialized optimization passes may use to implement type-based
1613 Volatile Memory Accesses
1614 ------------------------
1616 Certain memory accesses, such as :ref:`load <i_load>`'s,
1617 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1618 marked ``volatile``. The optimizers must not change the number of
1619 volatile operations or change their order of execution relative to other
1620 volatile operations. The optimizers *may* change the order of volatile
1621 operations relative to non-volatile operations. This is not Java's
1622 "volatile" and has no cross-thread synchronization behavior.
1624 IR-level volatile loads and stores cannot safely be optimized into
1625 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1626 flagged volatile. Likewise, the backend should never split or merge
1627 target-legal volatile load/store instructions.
1629 .. admonition:: Rationale
1631 Platforms may rely on volatile loads and stores of natively supported
1632 data width to be executed as single instruction. For example, in C
1633 this holds for an l-value of volatile primitive type with native
1634 hardware support, but not necessarily for aggregate types. The
1635 frontend upholds these expectations, which are intentionally
1636 unspecified in the IR. The rules above ensure that IR transformation
1637 do not violate the frontend's contract with the language.
1641 Memory Model for Concurrent Operations
1642 --------------------------------------
1644 The LLVM IR does not define any way to start parallel threads of
1645 execution or to register signal handlers. Nonetheless, there are
1646 platform-specific ways to create them, and we define LLVM IR's behavior
1647 in their presence. This model is inspired by the C++0x memory model.
1649 For a more informal introduction to this model, see the :doc:`Atomics`.
1651 We define a *happens-before* partial order as the least partial order
1654 - Is a superset of single-thread program order, and
1655 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1656 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1657 techniques, like pthread locks, thread creation, thread joining,
1658 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1659 Constraints <ordering>`).
1661 Note that program order does not introduce *happens-before* edges
1662 between a thread and signals executing inside that thread.
1664 Every (defined) read operation (load instructions, memcpy, atomic
1665 loads/read-modify-writes, etc.) R reads a series of bytes written by
1666 (defined) write operations (store instructions, atomic
1667 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1668 section, initialized globals are considered to have a write of the
1669 initializer which is atomic and happens before any other read or write
1670 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1671 may see any write to the same byte, except:
1673 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1674 write\ :sub:`2` happens before R\ :sub:`byte`, then
1675 R\ :sub:`byte` does not see write\ :sub:`1`.
1676 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1677 R\ :sub:`byte` does not see write\ :sub:`3`.
1679 Given that definition, R\ :sub:`byte` is defined as follows:
1681 - If R is volatile, the result is target-dependent. (Volatile is
1682 supposed to give guarantees which can support ``sig_atomic_t`` in
1683 C/C++, and may be used for accesses to addresses that do not behave
1684 like normal memory. It does not generally provide cross-thread
1686 - Otherwise, if there is no write to the same byte that happens before
1687 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1688 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1689 R\ :sub:`byte` returns the value written by that write.
1690 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1691 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1692 Memory Ordering Constraints <ordering>` section for additional
1693 constraints on how the choice is made.
1694 - Otherwise R\ :sub:`byte` returns ``undef``.
1696 R returns the value composed of the series of bytes it read. This
1697 implies that some bytes within the value may be ``undef`` **without**
1698 the entire value being ``undef``. Note that this only defines the
1699 semantics of the operation; it doesn't mean that targets will emit more
1700 than one instruction to read the series of bytes.
1702 Note that in cases where none of the atomic intrinsics are used, this
1703 model places only one restriction on IR transformations on top of what
1704 is required for single-threaded execution: introducing a store to a byte
1705 which might not otherwise be stored is not allowed in general.
1706 (Specifically, in the case where another thread might write to and read
1707 from an address, introducing a store can change a load that may see
1708 exactly one write into a load that may see multiple writes.)
1712 Atomic Memory Ordering Constraints
1713 ----------------------------------
1715 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1716 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1717 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1718 ordering parameters that determine which other atomic instructions on
1719 the same address they *synchronize with*. These semantics are borrowed
1720 from Java and C++0x, but are somewhat more colloquial. If these
1721 descriptions aren't precise enough, check those specs (see spec
1722 references in the :doc:`atomics guide <Atomics>`).
1723 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1724 differently since they don't take an address. See that instruction's
1725 documentation for details.
1727 For a simpler introduction to the ordering constraints, see the
1731 The set of values that can be read is governed by the happens-before
1732 partial order. A value cannot be read unless some operation wrote
1733 it. This is intended to provide a guarantee strong enough to model
1734 Java's non-volatile shared variables. This ordering cannot be
1735 specified for read-modify-write operations; it is not strong enough
1736 to make them atomic in any interesting way.
1738 In addition to the guarantees of ``unordered``, there is a single
1739 total order for modifications by ``monotonic`` operations on each
1740 address. All modification orders must be compatible with the
1741 happens-before order. There is no guarantee that the modification
1742 orders can be combined to a global total order for the whole program
1743 (and this often will not be possible). The read in an atomic
1744 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1745 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1746 order immediately before the value it writes. If one atomic read
1747 happens before another atomic read of the same address, the later
1748 read must see the same value or a later value in the address's
1749 modification order. This disallows reordering of ``monotonic`` (or
1750 stronger) operations on the same address. If an address is written
1751 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1752 read that address repeatedly, the other threads must eventually see
1753 the write. This corresponds to the C++0x/C1x
1754 ``memory_order_relaxed``.
1756 In addition to the guarantees of ``monotonic``, a
1757 *synchronizes-with* edge may be formed with a ``release`` operation.
1758 This is intended to model C++'s ``memory_order_acquire``.
1760 In addition to the guarantees of ``monotonic``, if this operation
1761 writes a value which is subsequently read by an ``acquire``
1762 operation, it *synchronizes-with* that operation. (This isn't a
1763 complete description; see the C++0x definition of a release
1764 sequence.) This corresponds to the C++0x/C1x
1765 ``memory_order_release``.
1766 ``acq_rel`` (acquire+release)
1767 Acts as both an ``acquire`` and ``release`` operation on its
1768 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1769 ``seq_cst`` (sequentially consistent)
1770 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1771 operation that only reads, ``release`` for an operation that only
1772 writes), there is a global total order on all
1773 sequentially-consistent operations on all addresses, which is
1774 consistent with the *happens-before* partial order and with the
1775 modification orders of all the affected addresses. Each
1776 sequentially-consistent read sees the last preceding write to the
1777 same address in this global order. This corresponds to the C++0x/C1x
1778 ``memory_order_seq_cst`` and Java volatile.
1782 If an atomic operation is marked ``singlethread``, it only *synchronizes
1783 with* or participates in modification and seq\_cst total orderings with
1784 other operations running in the same thread (for example, in signal
1792 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1793 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1794 :ref:`frem <i_frem>`) have the following flags that can be set to enable
1795 otherwise unsafe floating point operations
1798 No NaNs - Allow optimizations to assume the arguments and result are not
1799 NaN. Such optimizations are required to retain defined behavior over
1800 NaNs, but the value of the result is undefined.
1803 No Infs - Allow optimizations to assume the arguments and result are not
1804 +/-Inf. Such optimizations are required to retain defined behavior over
1805 +/-Inf, but the value of the result is undefined.
1808 No Signed Zeros - Allow optimizations to treat the sign of a zero
1809 argument or result as insignificant.
1812 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1813 argument rather than perform division.
1816 Fast - Allow algebraically equivalent transformations that may
1817 dramatically change results in floating point (e.g. reassociate). This
1818 flag implies all the others.
1822 Use-list Order Directives
1823 -------------------------
1825 Use-list directives encode the in-memory order of each use-list, allowing the
1826 order to be recreated. ``<order-indexes>`` is a comma-separated list of
1827 indexes that are assigned to the referenced value's uses. The referenced
1828 value's use-list is immediately sorted by these indexes.
1830 Use-list directives may appear at function scope or global scope. They are not
1831 instructions, and have no effect on the semantics of the IR. When they're at
1832 function scope, they must appear after the terminator of the final basic block.
1834 If basic blocks have their address taken via ``blockaddress()`` expressions,
1835 ``uselistorder_bb`` can be used to reorder their use-lists from outside their
1842 uselistorder <ty> <value>, { <order-indexes> }
1843 uselistorder_bb @function, %block { <order-indexes> }
1849 define void @foo(i32 %arg1, i32 %arg2) {
1851 ; ... instructions ...
1853 ; ... instructions ...
1855 ; At function scope.
1856 uselistorder i32 %arg1, { 1, 0, 2 }
1857 uselistorder label %bb, { 1, 0 }
1861 uselistorder i32* @global, { 1, 2, 0 }
1862 uselistorder i32 7, { 1, 0 }
1863 uselistorder i32 (i32) @bar, { 1, 0 }
1864 uselistorder_bb @foo, %bb, { 5, 1, 3, 2, 0, 4 }
1871 The LLVM type system is one of the most important features of the
1872 intermediate representation. Being typed enables a number of
1873 optimizations to be performed on the intermediate representation
1874 directly, without having to do extra analyses on the side before the
1875 transformation. A strong type system makes it easier to read the
1876 generated code and enables novel analyses and transformations that are
1877 not feasible to perform on normal three address code representations.
1887 The void type does not represent any value and has no size.
1905 The function type can be thought of as a function signature. It consists of a
1906 return type and a list of formal parameter types. The return type of a function
1907 type is a void type or first class type --- except for :ref:`label <t_label>`
1908 and :ref:`metadata <t_metadata>` types.
1914 <returntype> (<parameter list>)
1916 ...where '``<parameter list>``' is a comma-separated list of type
1917 specifiers. Optionally, the parameter list may include a type ``...``, which
1918 indicates that the function takes a variable number of arguments. Variable
1919 argument functions can access their arguments with the :ref:`variable argument
1920 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
1921 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
1925 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1926 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1927 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1928 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1929 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1930 | ``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. |
1931 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1932 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1933 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1940 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1941 Values of these types are the only ones which can be produced by
1949 These are the types that are valid in registers from CodeGen's perspective.
1958 The integer type is a very simple type that simply specifies an
1959 arbitrary bit width for the integer type desired. Any bit width from 1
1960 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1968 The number of bits the integer will occupy is specified by the ``N``
1974 +----------------+------------------------------------------------+
1975 | ``i1`` | a single-bit integer. |
1976 +----------------+------------------------------------------------+
1977 | ``i32`` | a 32-bit integer. |
1978 +----------------+------------------------------------------------+
1979 | ``i1942652`` | a really big integer of over 1 million bits. |
1980 +----------------+------------------------------------------------+
1984 Floating Point Types
1985 """"""""""""""""""""
1994 - 16-bit floating point value
1997 - 32-bit floating point value
2000 - 64-bit floating point value
2003 - 128-bit floating point value (112-bit mantissa)
2006 - 80-bit floating point value (X87)
2009 - 128-bit floating point value (two 64-bits)
2016 The x86_mmx type represents a value held in an MMX register on an x86
2017 machine. The operations allowed on it are quite limited: parameters and
2018 return values, load and store, and bitcast. User-specified MMX
2019 instructions are represented as intrinsic or asm calls with arguments
2020 and/or results of this type. There are no arrays, vectors or constants
2037 The pointer type is used to specify memory locations. Pointers are
2038 commonly used to reference objects in memory.
2040 Pointer types may have an optional address space attribute defining the
2041 numbered address space where the pointed-to object resides. The default
2042 address space is number zero. The semantics of non-zero address spaces
2043 are target-specific.
2045 Note that LLVM does not permit pointers to void (``void*``) nor does it
2046 permit pointers to labels (``label*``). Use ``i8*`` instead.
2056 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2057 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
2058 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2059 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
2060 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2061 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
2062 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2071 A vector type is a simple derived type that represents a vector of
2072 elements. Vector types are used when multiple primitive data are
2073 operated in parallel using a single instruction (SIMD). A vector type
2074 requires a size (number of elements) and an underlying primitive data
2075 type. Vector types are considered :ref:`first class <t_firstclass>`.
2081 < <# elements> x <elementtype> >
2083 The number of elements is a constant integer value larger than 0;
2084 elementtype may be any integer, floating point or pointer type. Vectors
2085 of size zero are not allowed.
2089 +-------------------+--------------------------------------------------+
2090 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
2091 +-------------------+--------------------------------------------------+
2092 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
2093 +-------------------+--------------------------------------------------+
2094 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
2095 +-------------------+--------------------------------------------------+
2096 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
2097 +-------------------+--------------------------------------------------+
2106 The label type represents code labels.
2121 The metadata type represents embedded metadata. No derived types may be
2122 created from metadata except for :ref:`function <t_function>` arguments.
2135 Aggregate Types are a subset of derived types that can contain multiple
2136 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
2137 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
2147 The array type is a very simple derived type that arranges elements
2148 sequentially in memory. The array type requires a size (number of
2149 elements) and an underlying data type.
2155 [<# elements> x <elementtype>]
2157 The number of elements is a constant integer value; ``elementtype`` may
2158 be any type with a size.
2162 +------------------+--------------------------------------+
2163 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
2164 +------------------+--------------------------------------+
2165 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
2166 +------------------+--------------------------------------+
2167 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
2168 +------------------+--------------------------------------+
2170 Here are some examples of multidimensional arrays:
2172 +-----------------------------+----------------------------------------------------------+
2173 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
2174 +-----------------------------+----------------------------------------------------------+
2175 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
2176 +-----------------------------+----------------------------------------------------------+
2177 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
2178 +-----------------------------+----------------------------------------------------------+
2180 There is no restriction on indexing beyond the end of the array implied
2181 by a static type (though there are restrictions on indexing beyond the
2182 bounds of an allocated object in some cases). This means that
2183 single-dimension 'variable sized array' addressing can be implemented in
2184 LLVM with a zero length array type. An implementation of 'pascal style
2185 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
2195 The structure type is used to represent a collection of data members
2196 together in memory. The elements of a structure may be any type that has
2199 Structures in memory are accessed using '``load``' and '``store``' by
2200 getting a pointer to a field with the '``getelementptr``' instruction.
2201 Structures in registers are accessed using the '``extractvalue``' and
2202 '``insertvalue``' instructions.
2204 Structures may optionally be "packed" structures, which indicate that
2205 the alignment of the struct is one byte, and that there is no padding
2206 between the elements. In non-packed structs, padding between field types
2207 is inserted as defined by the DataLayout string in the module, which is
2208 required to match what the underlying code generator expects.
2210 Structures can either be "literal" or "identified". A literal structure
2211 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
2212 identified types are always defined at the top level with a name.
2213 Literal types are uniqued by their contents and can never be recursive
2214 or opaque since there is no way to write one. Identified types can be
2215 recursive, can be opaqued, and are never uniqued.
2221 %T1 = type { <type list> } ; Identified normal struct type
2222 %T2 = type <{ <type list> }> ; Identified packed struct type
2226 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2227 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
2228 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2229 | ``{ 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``. |
2230 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2231 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
2232 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2236 Opaque Structure Types
2237 """"""""""""""""""""""
2241 Opaque structure types are used to represent named structure types that
2242 do not have a body specified. This corresponds (for example) to the C
2243 notion of a forward declared structure.
2254 +--------------+-------------------+
2255 | ``opaque`` | An opaque type. |
2256 +--------------+-------------------+
2263 LLVM has several different basic types of constants. This section
2264 describes them all and their syntax.
2269 **Boolean constants**
2270 The two strings '``true``' and '``false``' are both valid constants
2272 **Integer constants**
2273 Standard integers (such as '4') are constants of the
2274 :ref:`integer <t_integer>` type. Negative numbers may be used with
2276 **Floating point constants**
2277 Floating point constants use standard decimal notation (e.g.
2278 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2279 hexadecimal notation (see below). The assembler requires the exact
2280 decimal value of a floating-point constant. For example, the
2281 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2282 decimal in binary. Floating point constants must have a :ref:`floating
2283 point <t_floating>` type.
2284 **Null pointer constants**
2285 The identifier '``null``' is recognized as a null pointer constant
2286 and must be of :ref:`pointer type <t_pointer>`.
2288 The one non-intuitive notation for constants is the hexadecimal form of
2289 floating point constants. For example, the form
2290 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2291 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2292 constants are required (and the only time that they are generated by the
2293 disassembler) is when a floating point constant must be emitted but it
2294 cannot be represented as a decimal floating point number in a reasonable
2295 number of digits. For example, NaN's, infinities, and other special
2296 values are represented in their IEEE hexadecimal format so that assembly
2297 and disassembly do not cause any bits to change in the constants.
2299 When using the hexadecimal form, constants of types half, float, and
2300 double are represented using the 16-digit form shown above (which
2301 matches the IEEE754 representation for double); half and float values
2302 must, however, be exactly representable as IEEE 754 half and single
2303 precision, respectively. Hexadecimal format is always used for long
2304 double, and there are three forms of long double. The 80-bit format used
2305 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2306 128-bit format used by PowerPC (two adjacent doubles) is represented by
2307 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2308 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2309 will only work if they match the long double format on your target.
2310 The IEEE 16-bit format (half precision) is represented by ``0xH``
2311 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2312 (sign bit at the left).
2314 There are no constants of type x86_mmx.
2316 .. _complexconstants:
2321 Complex constants are a (potentially recursive) combination of simple
2322 constants and smaller complex constants.
2324 **Structure constants**
2325 Structure constants are represented with notation similar to
2326 structure type definitions (a comma separated list of elements,
2327 surrounded by braces (``{}``)). For example:
2328 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2329 "``@G = external global i32``". Structure constants must have
2330 :ref:`structure type <t_struct>`, and the number and types of elements
2331 must match those specified by the type.
2333 Array constants are represented with notation similar to array type
2334 definitions (a comma separated list of elements, surrounded by
2335 square brackets (``[]``)). For example:
2336 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2337 :ref:`array type <t_array>`, and the number and types of elements must
2338 match those specified by the type. As a special case, character array
2339 constants may also be represented as a double-quoted string using the ``c``
2340 prefix. For example: "``c"Hello World\0A\00"``".
2341 **Vector constants**
2342 Vector constants are represented with notation similar to vector
2343 type definitions (a comma separated list of elements, surrounded by
2344 less-than/greater-than's (``<>``)). For example:
2345 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2346 must have :ref:`vector type <t_vector>`, and the number and types of
2347 elements must match those specified by the type.
2348 **Zero initialization**
2349 The string '``zeroinitializer``' can be used to zero initialize a
2350 value to zero of *any* type, including scalar and
2351 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2352 having to print large zero initializers (e.g. for large arrays) and
2353 is always exactly equivalent to using explicit zero initializers.
2355 A metadata node is a constant tuple without types. For example:
2356 "``!{!0, !{!2, !0}, !"test"}``". Metadata can reference constant values,
2357 for example: "``!{!0, i32 0, i8* @global, i64 (i64)* @function, !"str"}``".
2358 Unlike other typed constants that are meant to be interpreted as part of
2359 the instruction stream, metadata is a place to attach additional
2360 information such as debug info.
2362 Global Variable and Function Addresses
2363 --------------------------------------
2365 The addresses of :ref:`global variables <globalvars>` and
2366 :ref:`functions <functionstructure>` are always implicitly valid
2367 (link-time) constants. These constants are explicitly referenced when
2368 the :ref:`identifier for the global <identifiers>` is used and always have
2369 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2372 .. code-block:: llvm
2376 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2383 The string '``undef``' can be used anywhere a constant is expected, and
2384 indicates that the user of the value may receive an unspecified
2385 bit-pattern. Undefined values may be of any type (other than '``label``'
2386 or '``void``') and be used anywhere a constant is permitted.
2388 Undefined values are useful because they indicate to the compiler that
2389 the program is well defined no matter what value is used. This gives the
2390 compiler more freedom to optimize. Here are some examples of
2391 (potentially surprising) transformations that are valid (in pseudo IR):
2393 .. code-block:: llvm
2403 This is safe because all of the output bits are affected by the undef
2404 bits. Any output bit can have a zero or one depending on the input bits.
2406 .. code-block:: llvm
2417 These logical operations have bits that are not always affected by the
2418 input. For example, if ``%X`` has a zero bit, then the output of the
2419 '``and``' operation will always be a zero for that bit, no matter what
2420 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2421 optimize or assume that the result of the '``and``' is '``undef``'.
2422 However, it is safe to assume that all bits of the '``undef``' could be
2423 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2424 all the bits of the '``undef``' operand to the '``or``' could be set,
2425 allowing the '``or``' to be folded to -1.
2427 .. code-block:: llvm
2429 %A = select undef, %X, %Y
2430 %B = select undef, 42, %Y
2431 %C = select %X, %Y, undef
2441 This set of examples shows that undefined '``select``' (and conditional
2442 branch) conditions can go *either way*, but they have to come from one
2443 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2444 both known to have a clear low bit, then ``%A`` would have to have a
2445 cleared low bit. However, in the ``%C`` example, the optimizer is
2446 allowed to assume that the '``undef``' operand could be the same as
2447 ``%Y``, allowing the whole '``select``' to be eliminated.
2449 .. code-block:: llvm
2451 %A = xor undef, undef
2468 This example points out that two '``undef``' operands are not
2469 necessarily the same. This can be surprising to people (and also matches
2470 C semantics) where they assume that "``X^X``" is always zero, even if
2471 ``X`` is undefined. This isn't true for a number of reasons, but the
2472 short answer is that an '``undef``' "variable" can arbitrarily change
2473 its value over its "live range". This is true because the variable
2474 doesn't actually *have a live range*. Instead, the value is logically
2475 read from arbitrary registers that happen to be around when needed, so
2476 the value is not necessarily consistent over time. In fact, ``%A`` and
2477 ``%C`` need to have the same semantics or the core LLVM "replace all
2478 uses with" concept would not hold.
2480 .. code-block:: llvm
2488 These examples show the crucial difference between an *undefined value*
2489 and *undefined behavior*. An undefined value (like '``undef``') is
2490 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2491 operation can be constant folded to '``undef``', because the '``undef``'
2492 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2493 However, in the second example, we can make a more aggressive
2494 assumption: because the ``undef`` is allowed to be an arbitrary value,
2495 we are allowed to assume that it could be zero. Since a divide by zero
2496 has *undefined behavior*, we are allowed to assume that the operation
2497 does not execute at all. This allows us to delete the divide and all
2498 code after it. Because the undefined operation "can't happen", the
2499 optimizer can assume that it occurs in dead code.
2501 .. code-block:: llvm
2503 a: store undef -> %X
2504 b: store %X -> undef
2509 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2510 value can be assumed to not have any effect; we can assume that the
2511 value is overwritten with bits that happen to match what was already
2512 there. However, a store *to* an undefined location could clobber
2513 arbitrary memory, therefore, it has undefined behavior.
2520 Poison values are similar to :ref:`undef values <undefvalues>`, however
2521 they also represent the fact that an instruction or constant expression
2522 that cannot evoke side effects has nevertheless detected a condition
2523 that results in undefined behavior.
2525 There is currently no way of representing a poison value in the IR; they
2526 only exist when produced by operations such as :ref:`add <i_add>` with
2529 Poison value behavior is defined in terms of value *dependence*:
2531 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2532 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2533 their dynamic predecessor basic block.
2534 - Function arguments depend on the corresponding actual argument values
2535 in the dynamic callers of their functions.
2536 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2537 instructions that dynamically transfer control back to them.
2538 - :ref:`Invoke <i_invoke>` instructions depend on the
2539 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2540 call instructions that dynamically transfer control back to them.
2541 - Non-volatile loads and stores depend on the most recent stores to all
2542 of the referenced memory addresses, following the order in the IR
2543 (including loads and stores implied by intrinsics such as
2544 :ref:`@llvm.memcpy <int_memcpy>`.)
2545 - An instruction with externally visible side effects depends on the
2546 most recent preceding instruction with externally visible side
2547 effects, following the order in the IR. (This includes :ref:`volatile
2548 operations <volatile>`.)
2549 - An instruction *control-depends* on a :ref:`terminator
2550 instruction <terminators>` if the terminator instruction has
2551 multiple successors and the instruction is always executed when
2552 control transfers to one of the successors, and may not be executed
2553 when control is transferred to another.
2554 - Additionally, an instruction also *control-depends* on a terminator
2555 instruction if the set of instructions it otherwise depends on would
2556 be different if the terminator had transferred control to a different
2558 - Dependence is transitive.
2560 Poison values have the same behavior as :ref:`undef values <undefvalues>`,
2561 with the additional effect that any instruction that has a *dependence*
2562 on a poison value has undefined behavior.
2564 Here are some examples:
2566 .. code-block:: llvm
2569 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2570 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2571 %poison_yet_again = getelementptr i32, i32* @h, i32 %still_poison
2572 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2574 store i32 %poison, i32* @g ; Poison value stored to memory.
2575 %poison2 = load i32, i32* @g ; Poison value loaded back from memory.
2577 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2579 %narrowaddr = bitcast i32* @g to i16*
2580 %wideaddr = bitcast i32* @g to i64*
2581 %poison3 = load i16, i16* %narrowaddr ; Returns a poison value.
2582 %poison4 = load i64, i64* %wideaddr ; Returns a poison value.
2584 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2585 br i1 %cmp, label %true, label %end ; Branch to either destination.
2588 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2589 ; it has undefined behavior.
2593 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2594 ; Both edges into this PHI are
2595 ; control-dependent on %cmp, so this
2596 ; always results in a poison value.
2598 store volatile i32 0, i32* @g ; This would depend on the store in %true
2599 ; if %cmp is true, or the store in %entry
2600 ; otherwise, so this is undefined behavior.
2602 br i1 %cmp, label %second_true, label %second_end
2603 ; The same branch again, but this time the
2604 ; true block doesn't have side effects.
2611 store volatile i32 0, i32* @g ; This time, the instruction always depends
2612 ; on the store in %end. Also, it is
2613 ; control-equivalent to %end, so this is
2614 ; well-defined (ignoring earlier undefined
2615 ; behavior in this example).
2619 Addresses of Basic Blocks
2620 -------------------------
2622 ``blockaddress(@function, %block)``
2624 The '``blockaddress``' constant computes the address of the specified
2625 basic block in the specified function, and always has an ``i8*`` type.
2626 Taking the address of the entry block is illegal.
2628 This value only has defined behavior when used as an operand to the
2629 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2630 against null. Pointer equality tests between labels addresses results in
2631 undefined behavior --- though, again, comparison against null is ok, and
2632 no label is equal to the null pointer. This may be passed around as an
2633 opaque pointer sized value as long as the bits are not inspected. This
2634 allows ``ptrtoint`` and arithmetic to be performed on these values so
2635 long as the original value is reconstituted before the ``indirectbr``
2638 Finally, some targets may provide defined semantics when using the value
2639 as the operand to an inline assembly, but that is target specific.
2643 Constant Expressions
2644 --------------------
2646 Constant expressions are used to allow expressions involving other
2647 constants to be used as constants. Constant expressions may be of any
2648 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2649 that does not have side effects (e.g. load and call are not supported).
2650 The following is the syntax for constant expressions:
2652 ``trunc (CST to TYPE)``
2653 Truncate a constant to another type. The bit size of CST must be
2654 larger than the bit size of TYPE. Both types must be integers.
2655 ``zext (CST to TYPE)``
2656 Zero extend a constant to another type. The bit size of CST must be
2657 smaller than the bit size of TYPE. Both types must be integers.
2658 ``sext (CST to TYPE)``
2659 Sign extend a constant to another type. The bit size of CST must be
2660 smaller than the bit size of TYPE. Both types must be integers.
2661 ``fptrunc (CST to TYPE)``
2662 Truncate a floating point constant to another floating point type.
2663 The size of CST must be larger than the size of TYPE. Both types
2664 must be floating point.
2665 ``fpext (CST to TYPE)``
2666 Floating point extend a constant to another type. The size of CST
2667 must be smaller or equal to the size of TYPE. Both types must be
2669 ``fptoui (CST to TYPE)``
2670 Convert a floating point constant to the corresponding unsigned
2671 integer constant. TYPE must be a scalar or vector integer type. CST
2672 must be of scalar or vector floating point type. Both CST and TYPE
2673 must be scalars, or vectors of the same number of elements. If the
2674 value won't fit in the integer type, the results are undefined.
2675 ``fptosi (CST to TYPE)``
2676 Convert a floating point constant to the corresponding signed
2677 integer constant. TYPE must be a scalar or vector integer type. CST
2678 must be of scalar or vector floating point type. Both CST and TYPE
2679 must be scalars, or vectors of the same number of elements. If the
2680 value won't fit in the integer type, the results are undefined.
2681 ``uitofp (CST to TYPE)``
2682 Convert an unsigned integer constant to the corresponding floating
2683 point constant. TYPE must be a scalar or vector floating point type.
2684 CST must be of scalar or vector integer type. Both CST and TYPE must
2685 be scalars, or vectors of the same number of elements. If the value
2686 won't fit in the floating point type, the results are undefined.
2687 ``sitofp (CST to TYPE)``
2688 Convert a signed integer constant to the corresponding floating
2689 point constant. TYPE must be a scalar or vector floating point type.
2690 CST must be of scalar or vector integer type. Both CST and TYPE must
2691 be scalars, or vectors of the same number of elements. If the value
2692 won't fit in the floating point type, the results are undefined.
2693 ``ptrtoint (CST to TYPE)``
2694 Convert a pointer typed constant to the corresponding integer
2695 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2696 pointer type. The ``CST`` value is zero extended, truncated, or
2697 unchanged to make it fit in ``TYPE``.
2698 ``inttoptr (CST to TYPE)``
2699 Convert an integer constant to a pointer constant. TYPE must be a
2700 pointer type. CST must be of integer type. The CST value is zero
2701 extended, truncated, or unchanged to make it fit in a pointer size.
2702 This one is *really* dangerous!
2703 ``bitcast (CST to TYPE)``
2704 Convert a constant, CST, to another TYPE. The constraints of the
2705 operands are the same as those for the :ref:`bitcast
2706 instruction <i_bitcast>`.
2707 ``addrspacecast (CST to TYPE)``
2708 Convert a constant pointer or constant vector of pointer, CST, to another
2709 TYPE in a different address space. The constraints of the operands are the
2710 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2711 ``getelementptr (TY, CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (TY, CSTPTR, IDX0, IDX1, ...)``
2712 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2713 constants. As with the :ref:`getelementptr <i_getelementptr>`
2714 instruction, the index list may have zero or more indexes, which are
2715 required to make sense for the type of "pointer to TY".
2716 ``select (COND, VAL1, VAL2)``
2717 Perform the :ref:`select operation <i_select>` on constants.
2718 ``icmp COND (VAL1, VAL2)``
2719 Performs the :ref:`icmp operation <i_icmp>` on constants.
2720 ``fcmp COND (VAL1, VAL2)``
2721 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2722 ``extractelement (VAL, IDX)``
2723 Perform the :ref:`extractelement operation <i_extractelement>` on
2725 ``insertelement (VAL, ELT, IDX)``
2726 Perform the :ref:`insertelement operation <i_insertelement>` on
2728 ``shufflevector (VEC1, VEC2, IDXMASK)``
2729 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2731 ``extractvalue (VAL, IDX0, IDX1, ...)``
2732 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2733 constants. The index list is interpreted in a similar manner as
2734 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2735 least one index value must be specified.
2736 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2737 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2738 The index list is interpreted in a similar manner as indices in a
2739 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2740 value must be specified.
2741 ``OPCODE (LHS, RHS)``
2742 Perform the specified operation of the LHS and RHS constants. OPCODE
2743 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2744 binary <bitwiseops>` operations. The constraints on operands are
2745 the same as those for the corresponding instruction (e.g. no bitwise
2746 operations on floating point values are allowed).
2753 Inline Assembler Expressions
2754 ----------------------------
2756 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2757 Inline Assembly <moduleasm>`) through the use of a special value. This
2758 value represents the inline assembler as a string (containing the
2759 instructions to emit), a list of operand constraints (stored as a
2760 string), a flag that indicates whether or not the inline asm expression
2761 has side effects, and a flag indicating whether the function containing
2762 the asm needs to align its stack conservatively. An example inline
2763 assembler expression is:
2765 .. code-block:: llvm
2767 i32 (i32) asm "bswap $0", "=r,r"
2769 Inline assembler expressions may **only** be used as the callee operand
2770 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2771 Thus, typically we have:
2773 .. code-block:: llvm
2775 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2777 Inline asms with side effects not visible in the constraint list must be
2778 marked as having side effects. This is done through the use of the
2779 '``sideeffect``' keyword, like so:
2781 .. code-block:: llvm
2783 call void asm sideeffect "eieio", ""()
2785 In some cases inline asms will contain code that will not work unless
2786 the stack is aligned in some way, such as calls or SSE instructions on
2787 x86, yet will not contain code that does that alignment within the asm.
2788 The compiler should make conservative assumptions about what the asm
2789 might contain and should generate its usual stack alignment code in the
2790 prologue if the '``alignstack``' keyword is present:
2792 .. code-block:: llvm
2794 call void asm alignstack "eieio", ""()
2796 Inline asms also support using non-standard assembly dialects. The
2797 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2798 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2799 the only supported dialects. An example is:
2801 .. code-block:: llvm
2803 call void asm inteldialect "eieio", ""()
2805 If multiple keywords appear the '``sideeffect``' keyword must come
2806 first, the '``alignstack``' keyword second and the '``inteldialect``'
2812 The call instructions that wrap inline asm nodes may have a
2813 "``!srcloc``" MDNode attached to it that contains a list of constant
2814 integers. If present, the code generator will use the integer as the
2815 location cookie value when report errors through the ``LLVMContext``
2816 error reporting mechanisms. This allows a front-end to correlate backend
2817 errors that occur with inline asm back to the source code that produced
2820 .. code-block:: llvm
2822 call void asm sideeffect "something bad", ""(), !srcloc !42
2824 !42 = !{ i32 1234567 }
2826 It is up to the front-end to make sense of the magic numbers it places
2827 in the IR. If the MDNode contains multiple constants, the code generator
2828 will use the one that corresponds to the line of the asm that the error
2836 LLVM IR allows metadata to be attached to instructions in the program
2837 that can convey extra information about the code to the optimizers and
2838 code generator. One example application of metadata is source-level
2839 debug information. There are two metadata primitives: strings and nodes.
2841 Metadata does not have a type, and is not a value. If referenced from a
2842 ``call`` instruction, it uses the ``metadata`` type.
2844 All metadata are identified in syntax by a exclamation point ('``!``').
2846 .. _metadata-string:
2848 Metadata Nodes and Metadata Strings
2849 -----------------------------------
2851 A metadata string is a string surrounded by double quotes. It can
2852 contain any character by escaping non-printable characters with
2853 "``\xx``" where "``xx``" is the two digit hex code. For example:
2856 Metadata nodes are represented with notation similar to structure
2857 constants (a comma separated list of elements, surrounded by braces and
2858 preceded by an exclamation point). Metadata nodes can have any values as
2859 their operand. For example:
2861 .. code-block:: llvm
2863 !{ !"test\00", i32 10}
2865 Metadata nodes that aren't uniqued use the ``distinct`` keyword. For example:
2867 .. code-block:: llvm
2869 !0 = distinct !{!"test\00", i32 10}
2871 ``distinct`` nodes are useful when nodes shouldn't be merged based on their
2872 content. They can also occur when transformations cause uniquing collisions
2873 when metadata operands change.
2875 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2876 metadata nodes, which can be looked up in the module symbol table. For
2879 .. code-block:: llvm
2883 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2884 function is using two metadata arguments:
2886 .. code-block:: llvm
2888 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2890 Metadata can be attached with an instruction. Here metadata ``!21`` is
2891 attached to the ``add`` instruction using the ``!dbg`` identifier:
2893 .. code-block:: llvm
2895 %indvar.next = add i64 %indvar, 1, !dbg !21
2897 More information about specific metadata nodes recognized by the
2898 optimizers and code generator is found below.
2900 .. _specialized-metadata:
2902 Specialized Metadata Nodes
2903 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2905 Specialized metadata nodes are custom data structures in metadata (as opposed
2906 to generic tuples). Their fields are labelled, and can be specified in any
2909 These aren't inherently debug info centric, but currently all the specialized
2910 metadata nodes are related to debug info.
2917 ``MDCompileUnit`` nodes represent a compile unit. The ``enums:``,
2918 ``retainedTypes:``, ``subprograms:``, ``globals:`` and ``imports:`` fields are
2919 tuples containing the debug info to be emitted along with the compile unit,
2920 regardless of code optimizations (some nodes are only emitted if there are
2921 references to them from instructions).
2923 .. code-block:: llvm
2925 !0 = !MDCompileUnit(language: DW_LANG_C99, file: !1, producer: "clang",
2926 isOptimized: true, flags: "-O2", runtimeVersion: 2,
2927 splitDebugFilename: "abc.debug", emissionKind: 1,
2928 enums: !2, retainedTypes: !3, subprograms: !4,
2929 globals: !5, imports: !6)
2931 Compile unit descriptors provide the root scope for objects declared in a
2932 specific compilation unit. File descriptors are defined using this scope.
2933 These descriptors are collected by a named metadata ``!llvm.dbg.cu``. They
2934 keep track of subprograms, global variables, type information, and imported
2935 entities (declarations and namespaces).
2942 ``MDFile`` nodes represent files. The ``filename:`` can include slashes.
2944 .. code-block:: llvm
2946 !0 = !MDFile(filename: "path/to/file", directory: "/path/to/dir")
2948 Files are sometimes used in ``scope:`` fields, and are the only valid target
2949 for ``file:`` fields.
2956 ``MDBasicType`` nodes represent primitive types, such as ``int``, ``bool`` and
2957 ``float``. ``tag:`` defaults to ``DW_TAG_base_type``.
2959 .. code-block:: llvm
2961 !0 = !MDBasicType(name: "unsigned char", size: 8, align: 8,
2962 encoding: DW_ATE_unsigned_char)
2963 !1 = !MDBasicType(tag: DW_TAG_unspecified_type, name: "decltype(nullptr)")
2965 The ``encoding:`` describes the details of the type. Usually it's one of the
2968 .. code-block:: llvm
2974 DW_ATE_signed_char = 6
2976 DW_ATE_unsigned_char = 8
2978 .. _MDSubroutineType:
2983 ``MDSubroutineType`` nodes represent subroutine types. Their ``types:`` field
2984 refers to a tuple; the first operand is the return type, while the rest are the
2985 types of the formal arguments in order. If the first operand is ``null``, that
2986 represents a function with no return value (such as ``void foo() {}`` in C++).
2988 .. code-block:: llvm
2990 !0 = !BasicType(name: "int", size: 32, align: 32, DW_ATE_signed)
2991 !1 = !BasicType(name: "char", size: 8, align: 8, DW_ATE_signed_char)
2992 !2 = !MDSubroutineType(types: !{null, !0, !1}) ; void (int, char)
2999 ``MDDerivedType`` nodes represent types derived from other types, such as
3002 .. code-block:: llvm
3004 !0 = !MDBasicType(name: "unsigned char", size: 8, align: 8,
3005 encoding: DW_ATE_unsigned_char)
3006 !1 = !MDDerivedType(tag: DW_TAG_pointer_type, baseType: !0, size: 32,
3009 The following ``tag:`` values are valid:
3011 .. code-block:: llvm
3013 DW_TAG_formal_parameter = 5
3015 DW_TAG_pointer_type = 15
3016 DW_TAG_reference_type = 16
3018 DW_TAG_ptr_to_member_type = 31
3019 DW_TAG_const_type = 38
3020 DW_TAG_volatile_type = 53
3021 DW_TAG_restrict_type = 55
3023 ``DW_TAG_member`` is used to define a member of a :ref:`composite type
3024 <MDCompositeType>` or :ref:`subprogram <MDSubprogram>`. The type of the member
3025 is the ``baseType:``. The ``offset:`` is the member's bit offset.
3026 ``DW_TAG_formal_parameter`` is used to define a member which is a formal
3027 argument of a subprogram.
3029 ``DW_TAG_typedef`` is used to provide a name for the ``baseType:``.
3031 ``DW_TAG_pointer_type``, ``DW_TAG_reference_type``, ``DW_TAG_const_type``,
3032 ``DW_TAG_volatile_type`` and ``DW_TAG_restrict_type`` are used to qualify the
3035 Note that the ``void *`` type is expressed as a type derived from NULL.
3037 .. _MDCompositeType:
3042 ``MDCompositeType`` nodes represent types composed of other types, like
3043 structures and unions. ``elements:`` points to a tuple of the composed types.
3045 If the source language supports ODR, the ``identifier:`` field gives the unique
3046 identifier used for type merging between modules. When specified, other types
3047 can refer to composite types indirectly via a :ref:`metadata string
3048 <metadata-string>` that matches their identifier.
3050 .. code-block:: llvm
3052 !0 = !MDEnumerator(name: "SixKind", value: 7)
3053 !1 = !MDEnumerator(name: "SevenKind", value: 7)
3054 !2 = !MDEnumerator(name: "NegEightKind", value: -8)
3055 !3 = !MDCompositeType(tag: DW_TAG_enumeration_type, name: "Enum", file: !12,
3056 line: 2, size: 32, align: 32, identifier: "_M4Enum",
3057 elements: !{!0, !1, !2})
3059 The following ``tag:`` values are valid:
3061 .. code-block:: llvm
3063 DW_TAG_array_type = 1
3064 DW_TAG_class_type = 2
3065 DW_TAG_enumeration_type = 4
3066 DW_TAG_structure_type = 19
3067 DW_TAG_union_type = 23
3068 DW_TAG_subroutine_type = 21
3069 DW_TAG_inheritance = 28
3072 For ``DW_TAG_array_type``, the ``elements:`` should be :ref:`subrange
3073 descriptors <MDSubrange>`, each representing the range of subscripts at that
3074 level of indexing. The ``DIFlagVector`` flag to ``flags:`` indicates that an
3075 array type is a native packed vector.
3077 For ``DW_TAG_enumeration_type``, the ``elements:`` should be :ref:`enumerator
3078 descriptors <MDEnumerator>`, each representing the definition of an enumeration
3079 value for the set. All enumeration type descriptors are collected in the
3080 ``enums:`` field of the :ref:`compile unit <MDCompileUnit>`.
3082 For ``DW_TAG_structure_type``, ``DW_TAG_class_type``, and
3083 ``DW_TAG_union_type``, the ``elements:`` should be :ref:`derived types
3084 <MDDerivedType>` with ``tag: DW_TAG_member`` or ``tag: DW_TAG_inheritance``.
3091 ``MDSubrange`` nodes are the elements for ``DW_TAG_array_type`` variants of
3092 :ref:`MDCompositeType`. ``count: -1`` indicates an empty array.
3094 .. code-block:: llvm
3096 !0 = !MDSubrange(count: 5, lowerBound: 0) ; array counting from 0
3097 !1 = !MDSubrange(count: 5, lowerBound: 1) ; array counting from 1
3098 !2 = !MDSubrange(count: -1) ; empty array.
3105 ``MDEnumerator`` nodes are the elements for ``DW_TAG_enumeration_type``
3106 variants of :ref:`MDCompositeType`.
3108 .. code-block:: llvm
3110 !0 = !MDEnumerator(name: "SixKind", value: 7)
3111 !1 = !MDEnumerator(name: "SevenKind", value: 7)
3112 !2 = !MDEnumerator(name: "NegEightKind", value: -8)
3114 MDTemplateTypeParameter
3115 """""""""""""""""""""""
3117 ``MDTemplateTypeParameter`` nodes represent type parameters to generic source
3118 language constructs. They are used (optionally) in :ref:`MDCompositeType` and
3119 :ref:`MDSubprogram` ``templateParams:`` fields.
3121 .. code-block:: llvm
3123 !0 = !MDTemplateTypeParameter(name: "Ty", type: !1)
3125 MDTemplateValueParameter
3126 """"""""""""""""""""""""
3128 ``MDTemplateValueParameter`` nodes represent value parameters to generic source
3129 language constructs. ``tag:`` defaults to ``DW_TAG_template_value_parameter``,
3130 but if specified can also be set to ``DW_TAG_GNU_template_template_param`` or
3131 ``DW_TAG_GNU_template_param_pack``. They are used (optionally) in
3132 :ref:`MDCompositeType` and :ref:`MDSubprogram` ``templateParams:`` fields.
3134 .. code-block:: llvm
3136 !0 = !MDTemplateValueParameter(name: "Ty", type: !1, value: i32 7)
3141 ``MDNamespace`` nodes represent namespaces in the source language.
3143 .. code-block:: llvm
3145 !0 = !MDNamespace(name: "myawesomeproject", scope: !1, file: !2, line: 7)
3150 ``MDGlobalVariable`` nodes represent global variables in the source language.
3152 .. code-block:: llvm
3154 !0 = !MDGlobalVariable(name: "foo", linkageName: "foo", scope: !1,
3155 file: !2, line: 7, type: !3, isLocal: true,
3156 isDefinition: false, variable: i32* @foo,
3159 All global variables should be referenced by the `globals:` field of a
3160 :ref:`compile unit <MDCompileUnit>`.
3167 ``MDSubprogram`` nodes represent functions from the source language. The
3168 ``variables:`` field points at :ref:`variables <MDLocalVariable>` that must be
3169 retained, even if their IR counterparts are optimized out of the IR. The
3170 ``type:`` field must point at an :ref:`MDSubroutineType`.
3172 .. code-block:: llvm
3174 !0 = !MDSubprogram(name: "foo", linkageName: "_Zfoov", scope: !1,
3175 file: !2, line: 7, type: !3, isLocal: true,
3176 isDefinition: false, scopeLine: 8, containingType: !4,
3177 virtuality: DW_VIRTUALITY_pure_virtual, virtualIndex: 10,
3178 flags: DIFlagPrototyped, isOptimized: true,
3179 function: void ()* @_Z3foov,
3180 templateParams: !5, declaration: !6, variables: !7)
3187 ``MDLexicalBlock`` nodes describe nested blocks within a :ref:`subprogram
3188 <MDSubprogram>`. The line number and column numbers are used to dinstinguish
3189 two lexical blocks at same depth. They are valid targets for ``scope:``
3192 .. code-block:: llvm
3194 !0 = distinct !MDLexicalBlock(scope: !1, file: !2, line: 7, column: 35)
3196 Usually lexical blocks are ``distinct`` to prevent node merging based on
3199 .. _MDLexicalBlockFile:
3204 ``MDLexicalBlockFile`` nodes are used to discriminate between sections of a
3205 :ref:`lexical block <MDLexicalBlock>`. The ``file:`` field can be changed to
3206 indicate textual inclusion, or the ``discriminator:`` field can be used to
3207 discriminate between control flow within a single block in the source language.
3209 .. code-block:: llvm
3211 !0 = !MDLexicalBlock(scope: !3, file: !4, line: 7, column: 35)
3212 !1 = !MDLexicalBlockFile(scope: !0, file: !4, discriminator: 0)
3213 !2 = !MDLexicalBlockFile(scope: !0, file: !4, discriminator: 1)
3218 ``MDLocation`` nodes represent source debug locations. The ``scope:`` field is
3219 mandatory, and points at an :ref:`MDLexicalBlockFile`, an
3220 :ref:`MDLexicalBlock`, or an :ref:`MDSubprogram`.
3222 .. code-block:: llvm
3224 !0 = !MDLocation(line: 2900, column: 42, scope: !1, inlinedAt: !2)
3226 .. _MDLocalVariable:
3231 ``MDLocalVariable`` nodes represent local variables in the source language.
3232 Instead of ``DW_TAG_variable``, they use LLVM-specific fake tags to
3233 discriminate between local variables (``DW_TAG_auto_variable``) and subprogram
3234 arguments (``DW_TAG_arg_variable``). In the latter case, the ``arg:`` field
3235 specifies the argument position, and this variable will be included in the
3236 ``variables:`` field of its :ref:`MDSubprogram`.
3238 If set, the ``inlinedAt:`` field points at an :ref:`MDLocation`, and the
3239 variable represents an inlined version of a variable (with all other fields
3240 duplicated from the non-inlined version).
3242 .. code-block:: llvm
3244 !0 = !MDLocalVariable(tag: DW_TAG_arg_variable, name: "this", arg: 0,
3245 scope: !3, file: !2, line: 7, type: !3,
3246 flags: DIFlagArtificial, inlinedAt: !4)
3247 !1 = !MDLocalVariable(tag: DW_TAG_arg_variable, name: "x", arg: 1,
3248 scope: !4, file: !2, line: 7, type: !3,
3250 !1 = !MDLocalVariable(tag: DW_TAG_auto_variable, name: "y",
3251 scope: !5, file: !2, line: 7, type: !3,
3257 ``MDExpression`` nodes represent DWARF expression sequences. They are used in
3258 :ref:`debug intrinsics<dbg_intrinsics>` (such as ``llvm.dbg.declare``) to
3259 describe how the referenced LLVM variable relates to the source language
3262 The current supported vocabulary is limited:
3264 - ``DW_OP_deref`` dereferences the working expression.
3265 - ``DW_OP_plus, 93`` adds ``93`` to the working expression.
3266 - ``DW_OP_bit_piece, 16, 8`` specifies the offset and size (``16`` and ``8``
3267 here, respectively) of the variable piece from the working expression.
3269 .. code-block:: llvm
3271 !0 = !MDExpression(DW_OP_deref)
3272 !1 = !MDExpression(DW_OP_plus, 3)
3273 !2 = !MDExpression(DW_OP_bit_piece, 3, 7)
3274 !3 = !MDExpression(DW_OP_deref, DW_OP_plus, 3, DW_OP_bit_piece, 3, 7)
3279 ``MDObjCProperty`` nodes represent Objective-C property nodes.
3281 .. code-block:: llvm
3283 !3 = !MDObjCProperty(name: "foo", file: !1, line: 7, setter: "setFoo",
3284 getter: "getFoo", attributes: 7, type: !2)
3289 ``MDImportedEntity`` nodes represent entities (such as modules) imported into a
3292 .. code-block:: llvm
3294 !2 = !MDImportedEntity(tag: DW_TAG_imported_module, name: "foo", scope: !0,
3295 entity: !1, line: 7)
3300 In LLVM IR, memory does not have types, so LLVM's own type system is not
3301 suitable for doing TBAA. Instead, metadata is added to the IR to
3302 describe a type system of a higher level language. This can be used to
3303 implement typical C/C++ TBAA, but it can also be used to implement
3304 custom alias analysis behavior for other languages.
3306 The current metadata format is very simple. TBAA metadata nodes have up
3307 to three fields, e.g.:
3309 .. code-block:: llvm
3311 !0 = !{ !"an example type tree" }
3312 !1 = !{ !"int", !0 }
3313 !2 = !{ !"float", !0 }
3314 !3 = !{ !"const float", !2, i64 1 }
3316 The first field is an identity field. It can be any value, usually a
3317 metadata string, which uniquely identifies the type. The most important
3318 name in the tree is the name of the root node. Two trees with different
3319 root node names are entirely disjoint, even if they have leaves with
3322 The second field identifies the type's parent node in the tree, or is
3323 null or omitted for a root node. A type is considered to alias all of
3324 its descendants and all of its ancestors in the tree. Also, a type is
3325 considered to alias all types in other trees, so that bitcode produced
3326 from multiple front-ends is handled conservatively.
3328 If the third field is present, it's an integer which if equal to 1
3329 indicates that the type is "constant" (meaning
3330 ``pointsToConstantMemory`` should return true; see `other useful
3331 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
3333 '``tbaa.struct``' Metadata
3334 ^^^^^^^^^^^^^^^^^^^^^^^^^^
3336 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
3337 aggregate assignment operations in C and similar languages, however it
3338 is defined to copy a contiguous region of memory, which is more than
3339 strictly necessary for aggregate types which contain holes due to
3340 padding. Also, it doesn't contain any TBAA information about the fields
3343 ``!tbaa.struct`` metadata can describe which memory subregions in a
3344 memcpy are padding and what the TBAA tags of the struct are.
3346 The current metadata format is very simple. ``!tbaa.struct`` metadata
3347 nodes are a list of operands which are in conceptual groups of three.
3348 For each group of three, the first operand gives the byte offset of a
3349 field in bytes, the second gives its size in bytes, and the third gives
3352 .. code-block:: llvm
3354 !4 = !{ i64 0, i64 4, !1, i64 8, i64 4, !2 }
3356 This describes a struct with two fields. The first is at offset 0 bytes
3357 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
3358 and has size 4 bytes and has tbaa tag !2.
3360 Note that the fields need not be contiguous. In this example, there is a
3361 4 byte gap between the two fields. This gap represents padding which
3362 does not carry useful data and need not be preserved.
3364 '``noalias``' and '``alias.scope``' Metadata
3365 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3367 ``noalias`` and ``alias.scope`` metadata provide the ability to specify generic
3368 noalias memory-access sets. This means that some collection of memory access
3369 instructions (loads, stores, memory-accessing calls, etc.) that carry
3370 ``noalias`` metadata can specifically be specified not to alias with some other
3371 collection of memory access instructions that carry ``alias.scope`` metadata.
3372 Each type of metadata specifies a list of scopes where each scope has an id and
3373 a domain. When evaluating an aliasing query, if for some some domain, the set
3374 of scopes with that domain in one instruction's ``alias.scope`` list is a
3375 subset of (or equal to) the set of scopes for that domain in another
3376 instruction's ``noalias`` list, then the two memory accesses are assumed not to
3379 The metadata identifying each domain is itself a list containing one or two
3380 entries. The first entry is the name of the domain. Note that if the name is a
3381 string then it can be combined accross functions and translation units. A
3382 self-reference can be used to create globally unique domain names. A
3383 descriptive string may optionally be provided as a second list entry.
3385 The metadata identifying each scope is also itself a list containing two or
3386 three entries. The first entry is the name of the scope. Note that if the name
3387 is a string then it can be combined accross functions and translation units. A
3388 self-reference can be used to create globally unique scope names. A metadata
3389 reference to the scope's domain is the second entry. A descriptive string may
3390 optionally be provided as a third list entry.
3394 .. code-block:: llvm
3396 ; Two scope domains:
3400 ; Some scopes in these domains:
3406 !5 = !{!4} ; A list containing only scope !4
3410 ; These two instructions don't alias:
3411 %0 = load float, float* %c, align 4, !alias.scope !5
3412 store float %0, float* %arrayidx.i, align 4, !noalias !5
3414 ; These two instructions also don't alias (for domain !1, the set of scopes
3415 ; in the !alias.scope equals that in the !noalias list):
3416 %2 = load float, float* %c, align 4, !alias.scope !5
3417 store float %2, float* %arrayidx.i2, align 4, !noalias !6
3419 ; These two instructions don't alias (for domain !0, the set of scopes in
3420 ; the !noalias list is not a superset of, or equal to, the scopes in the
3421 ; !alias.scope list):
3422 %2 = load float, float* %c, align 4, !alias.scope !6
3423 store float %0, float* %arrayidx.i, align 4, !noalias !7
3425 '``fpmath``' Metadata
3426 ^^^^^^^^^^^^^^^^^^^^^
3428 ``fpmath`` metadata may be attached to any instruction of floating point
3429 type. It can be used to express the maximum acceptable error in the
3430 result of that instruction, in ULPs, thus potentially allowing the
3431 compiler to use a more efficient but less accurate method of computing
3432 it. ULP is defined as follows:
3434 If ``x`` is a real number that lies between two finite consecutive
3435 floating-point numbers ``a`` and ``b``, without being equal to one
3436 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
3437 distance between the two non-equal finite floating-point numbers
3438 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
3440 The metadata node shall consist of a single positive floating point
3441 number representing the maximum relative error, for example:
3443 .. code-block:: llvm
3445 !0 = !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
3449 '``range``' Metadata
3450 ^^^^^^^^^^^^^^^^^^^^
3452 ``range`` metadata may be attached only to ``load``, ``call`` and ``invoke`` of
3453 integer types. It expresses the possible ranges the loaded value or the value
3454 returned by the called function at this call site is in. The ranges are
3455 represented with a flattened list of integers. The loaded value or the value
3456 returned is known to be in the union of the ranges defined by each consecutive
3457 pair. Each pair has the following properties:
3459 - The type must match the type loaded by the instruction.
3460 - The pair ``a,b`` represents the range ``[a,b)``.
3461 - Both ``a`` and ``b`` are constants.
3462 - The range is allowed to wrap.
3463 - The range should not represent the full or empty set. That is,
3466 In addition, the pairs must be in signed order of the lower bound and
3467 they must be non-contiguous.
3471 .. code-block:: llvm
3473 %a = load i8, i8* %x, align 1, !range !0 ; Can only be 0 or 1
3474 %b = load i8, i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
3475 %c = call i8 @foo(), !range !2 ; Can only be 0, 1, 3, 4 or 5
3476 %d = invoke i8 @bar() to label %cont
3477 unwind label %lpad, !range !3 ; Can only be -2, -1, 3, 4 or 5
3479 !0 = !{ i8 0, i8 2 }
3480 !1 = !{ i8 255, i8 2 }
3481 !2 = !{ i8 0, i8 2, i8 3, i8 6 }
3482 !3 = !{ i8 -2, i8 0, i8 3, i8 6 }
3487 It is sometimes useful to attach information to loop constructs. Currently,
3488 loop metadata is implemented as metadata attached to the branch instruction
3489 in the loop latch block. This type of metadata refer to a metadata node that is
3490 guaranteed to be separate for each loop. The loop identifier metadata is
3491 specified with the name ``llvm.loop``.
3493 The loop identifier metadata is implemented using a metadata that refers to
3494 itself to avoid merging it with any other identifier metadata, e.g.,
3495 during module linkage or function inlining. That is, each loop should refer
3496 to their own identification metadata even if they reside in separate functions.
3497 The following example contains loop identifier metadata for two separate loop
3500 .. code-block:: llvm
3505 The loop identifier metadata can be used to specify additional
3506 per-loop metadata. Any operands after the first operand can be treated
3507 as user-defined metadata. For example the ``llvm.loop.unroll.count``
3508 suggests an unroll factor to the loop unroller:
3510 .. code-block:: llvm
3512 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
3515 !1 = !{!"llvm.loop.unroll.count", i32 4}
3517 '``llvm.loop.vectorize``' and '``llvm.loop.interleave``'
3518 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3520 Metadata prefixed with ``llvm.loop.vectorize`` or ``llvm.loop.interleave`` are
3521 used to control per-loop vectorization and interleaving parameters such as
3522 vectorization width and interleave count. These metadata should be used in
3523 conjunction with ``llvm.loop`` loop identification metadata. The
3524 ``llvm.loop.vectorize`` and ``llvm.loop.interleave`` metadata are only
3525 optimization hints and the optimizer will only interleave and vectorize loops if
3526 it believes it is safe to do so. The ``llvm.mem.parallel_loop_access`` metadata
3527 which contains information about loop-carried memory dependencies can be helpful
3528 in determining the safety of these transformations.
3530 '``llvm.loop.interleave.count``' Metadata
3531 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3533 This metadata suggests an interleave count to the loop interleaver.
3534 The first operand is the string ``llvm.loop.interleave.count`` and the
3535 second operand is an integer specifying the interleave count. For
3538 .. code-block:: llvm
3540 !0 = !{!"llvm.loop.interleave.count", i32 4}
3542 Note that setting ``llvm.loop.interleave.count`` to 1 disables interleaving
3543 multiple iterations of the loop. If ``llvm.loop.interleave.count`` is set to 0
3544 then the interleave count will be determined automatically.
3546 '``llvm.loop.vectorize.enable``' Metadata
3547 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3549 This metadata selectively enables or disables vectorization for the loop. The
3550 first operand is the string ``llvm.loop.vectorize.enable`` and the second operand
3551 is a bit. If the bit operand value is 1 vectorization is enabled. A value of
3552 0 disables vectorization:
3554 .. code-block:: llvm
3556 !0 = !{!"llvm.loop.vectorize.enable", i1 0}
3557 !1 = !{!"llvm.loop.vectorize.enable", i1 1}
3559 '``llvm.loop.vectorize.width``' Metadata
3560 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3562 This metadata sets the target width of the vectorizer. The first
3563 operand is the string ``llvm.loop.vectorize.width`` and the second
3564 operand is an integer specifying the width. For example:
3566 .. code-block:: llvm
3568 !0 = !{!"llvm.loop.vectorize.width", i32 4}
3570 Note that setting ``llvm.loop.vectorize.width`` to 1 disables
3571 vectorization of the loop. If ``llvm.loop.vectorize.width`` is set to
3572 0 or if the loop does not have this metadata the width will be
3573 determined automatically.
3575 '``llvm.loop.unroll``'
3576 ^^^^^^^^^^^^^^^^^^^^^^
3578 Metadata prefixed with ``llvm.loop.unroll`` are loop unrolling
3579 optimization hints such as the unroll factor. ``llvm.loop.unroll``
3580 metadata should be used in conjunction with ``llvm.loop`` loop
3581 identification metadata. The ``llvm.loop.unroll`` metadata are only
3582 optimization hints and the unrolling will only be performed if the
3583 optimizer believes it is safe to do so.
3585 '``llvm.loop.unroll.count``' Metadata
3586 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3588 This metadata suggests an unroll factor to the loop unroller. The
3589 first operand is the string ``llvm.loop.unroll.count`` and the second
3590 operand is a positive integer specifying the unroll factor. For
3593 .. code-block:: llvm
3595 !0 = !{!"llvm.loop.unroll.count", i32 4}
3597 If the trip count of the loop is less than the unroll count the loop
3598 will be partially unrolled.
3600 '``llvm.loop.unroll.disable``' Metadata
3601 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3603 This metadata either disables loop unrolling. The metadata has a single operand
3604 which is the string ``llvm.loop.unroll.disable``. For example:
3606 .. code-block:: llvm
3608 !0 = !{!"llvm.loop.unroll.disable"}
3610 '``llvm.loop.unroll.runtime.disable``' Metadata
3611 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3613 This metadata either disables runtime loop unrolling. The metadata has a single
3614 operand which is the string ``llvm.loop.unroll.runtime.disable``. For example:
3616 .. code-block:: llvm
3618 !0 = !{!"llvm.loop.unroll.runtime.disable"}
3620 '``llvm.loop.unroll.full``' Metadata
3621 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3623 This metadata either suggests that the loop should be unrolled fully. The
3624 metadata has a single operand which is the string ``llvm.loop.unroll.disable``.
3627 .. code-block:: llvm
3629 !0 = !{!"llvm.loop.unroll.full"}
3634 Metadata types used to annotate memory accesses with information helpful
3635 for optimizations are prefixed with ``llvm.mem``.
3637 '``llvm.mem.parallel_loop_access``' Metadata
3638 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3640 The ``llvm.mem.parallel_loop_access`` metadata refers to a loop identifier,
3641 or metadata containing a list of loop identifiers for nested loops.
3642 The metadata is attached to memory accessing instructions and denotes that
3643 no loop carried memory dependence exist between it and other instructions denoted
3644 with the same loop identifier.
3646 Precisely, given two instructions ``m1`` and ``m2`` that both have the
3647 ``llvm.mem.parallel_loop_access`` metadata, with ``L1`` and ``L2`` being the
3648 set of loops associated with that metadata, respectively, then there is no loop
3649 carried dependence between ``m1`` and ``m2`` for loops in both ``L1`` and
3652 As a special case, if all memory accessing instructions in a loop have
3653 ``llvm.mem.parallel_loop_access`` metadata that refers to that loop, then the
3654 loop has no loop carried memory dependences and is considered to be a parallel
3657 Note that if not all memory access instructions have such metadata referring to
3658 the loop, then the loop is considered not being trivially parallel. Additional
3659 memory dependence analysis is required to make that determination. As a fail
3660 safe mechanism, this causes loops that were originally parallel to be considered
3661 sequential (if optimization passes that are unaware of the parallel semantics
3662 insert new memory instructions into the loop body).
3664 Example of a loop that is considered parallel due to its correct use of
3665 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
3666 metadata types that refer to the same loop identifier metadata.
3668 .. code-block:: llvm
3672 %val0 = load i32, i32* %arrayidx, !llvm.mem.parallel_loop_access !0
3674 store i32 %val0, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3676 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
3682 It is also possible to have nested parallel loops. In that case the
3683 memory accesses refer to a list of loop identifier metadata nodes instead of
3684 the loop identifier metadata node directly:
3686 .. code-block:: llvm
3690 %val1 = load i32, i32* %arrayidx3, !llvm.mem.parallel_loop_access !2
3692 br label %inner.for.body
3696 %val0 = load i32, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3698 store i32 %val0, i32* %arrayidx2, !llvm.mem.parallel_loop_access !0
3700 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
3704 store i32 %val1, i32* %arrayidx4, !llvm.mem.parallel_loop_access !2
3706 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
3708 outer.for.end: ; preds = %for.body
3710 !0 = !{!1, !2} ; a list of loop identifiers
3711 !1 = !{!1} ; an identifier for the inner loop
3712 !2 = !{!2} ; an identifier for the outer loop
3717 The ``llvm.bitsets`` global metadata is used to implement
3718 :doc:`bitsets <BitSets>`.
3720 Module Flags Metadata
3721 =====================
3723 Information about the module as a whole is difficult to convey to LLVM's
3724 subsystems. The LLVM IR isn't sufficient to transmit this information.
3725 The ``llvm.module.flags`` named metadata exists in order to facilitate
3726 this. These flags are in the form of key / value pairs --- much like a
3727 dictionary --- making it easy for any subsystem who cares about a flag to
3730 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
3731 Each triplet has the following form:
3733 - The first element is a *behavior* flag, which specifies the behavior
3734 when two (or more) modules are merged together, and it encounters two
3735 (or more) metadata with the same ID. The supported behaviors are
3737 - The second element is a metadata string that is a unique ID for the
3738 metadata. Each module may only have one flag entry for each unique ID (not
3739 including entries with the **Require** behavior).
3740 - The third element is the value of the flag.
3742 When two (or more) modules are merged together, the resulting
3743 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
3744 each unique metadata ID string, there will be exactly one entry in the merged
3745 modules ``llvm.module.flags`` metadata table, and the value for that entry will
3746 be determined by the merge behavior flag, as described below. The only exception
3747 is that entries with the *Require* behavior are always preserved.
3749 The following behaviors are supported:
3760 Emits an error if two values disagree, otherwise the resulting value
3761 is that of the operands.
3765 Emits a warning if two values disagree. The result value will be the
3766 operand for the flag from the first module being linked.
3770 Adds a requirement that another module flag be present and have a
3771 specified value after linking is performed. The value must be a
3772 metadata pair, where the first element of the pair is the ID of the
3773 module flag to be restricted, and the second element of the pair is
3774 the value the module flag should be restricted to. This behavior can
3775 be used to restrict the allowable results (via triggering of an
3776 error) of linking IDs with the **Override** behavior.
3780 Uses the specified value, regardless of the behavior or value of the
3781 other module. If both modules specify **Override**, but the values
3782 differ, an error will be emitted.
3786 Appends the two values, which are required to be metadata nodes.
3790 Appends the two values, which are required to be metadata
3791 nodes. However, duplicate entries in the second list are dropped
3792 during the append operation.
3794 It is an error for a particular unique flag ID to have multiple behaviors,
3795 except in the case of **Require** (which adds restrictions on another metadata
3796 value) or **Override**.
3798 An example of module flags:
3800 .. code-block:: llvm
3802 !0 = !{ i32 1, !"foo", i32 1 }
3803 !1 = !{ i32 4, !"bar", i32 37 }
3804 !2 = !{ i32 2, !"qux", i32 42 }
3805 !3 = !{ i32 3, !"qux",
3810 !llvm.module.flags = !{ !0, !1, !2, !3 }
3812 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
3813 if two or more ``!"foo"`` flags are seen is to emit an error if their
3814 values are not equal.
3816 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
3817 behavior if two or more ``!"bar"`` flags are seen is to use the value
3820 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
3821 behavior if two or more ``!"qux"`` flags are seen is to emit a
3822 warning if their values are not equal.
3824 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
3830 The behavior is to emit an error if the ``llvm.module.flags`` does not
3831 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
3834 Objective-C Garbage Collection Module Flags Metadata
3835 ----------------------------------------------------
3837 On the Mach-O platform, Objective-C stores metadata about garbage
3838 collection in a special section called "image info". The metadata
3839 consists of a version number and a bitmask specifying what types of
3840 garbage collection are supported (if any) by the file. If two or more
3841 modules are linked together their garbage collection metadata needs to
3842 be merged rather than appended together.
3844 The Objective-C garbage collection module flags metadata consists of the
3845 following key-value pairs:
3854 * - ``Objective-C Version``
3855 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
3857 * - ``Objective-C Image Info Version``
3858 - **[Required]** --- The version of the image info section. Currently
3861 * - ``Objective-C Image Info Section``
3862 - **[Required]** --- The section to place the metadata. Valid values are
3863 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
3864 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
3865 Objective-C ABI version 2.
3867 * - ``Objective-C Garbage Collection``
3868 - **[Required]** --- Specifies whether garbage collection is supported or
3869 not. Valid values are 0, for no garbage collection, and 2, for garbage
3870 collection supported.
3872 * - ``Objective-C GC Only``
3873 - **[Optional]** --- Specifies that only garbage collection is supported.
3874 If present, its value must be 6. This flag requires that the
3875 ``Objective-C Garbage Collection`` flag have the value 2.
3877 Some important flag interactions:
3879 - If a module with ``Objective-C Garbage Collection`` set to 0 is
3880 merged with a module with ``Objective-C Garbage Collection`` set to
3881 2, then the resulting module has the
3882 ``Objective-C Garbage Collection`` flag set to 0.
3883 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
3884 merged with a module with ``Objective-C GC Only`` set to 6.
3886 Automatic Linker Flags Module Flags Metadata
3887 --------------------------------------------
3889 Some targets support embedding flags to the linker inside individual object
3890 files. Typically this is used in conjunction with language extensions which
3891 allow source files to explicitly declare the libraries they depend on, and have
3892 these automatically be transmitted to the linker via object files.
3894 These flags are encoded in the IR using metadata in the module flags section,
3895 using the ``Linker Options`` key. The merge behavior for this flag is required
3896 to be ``AppendUnique``, and the value for the key is expected to be a metadata
3897 node which should be a list of other metadata nodes, each of which should be a
3898 list of metadata strings defining linker options.
3900 For example, the following metadata section specifies two separate sets of
3901 linker options, presumably to link against ``libz`` and the ``Cocoa``
3904 !0 = !{ i32 6, !"Linker Options",
3907 !{ !"-framework", !"Cocoa" } } }
3908 !llvm.module.flags = !{ !0 }
3910 The metadata encoding as lists of lists of options, as opposed to a collapsed
3911 list of options, is chosen so that the IR encoding can use multiple option
3912 strings to specify e.g., a single library, while still having that specifier be
3913 preserved as an atomic element that can be recognized by a target specific
3914 assembly writer or object file emitter.
3916 Each individual option is required to be either a valid option for the target's
3917 linker, or an option that is reserved by the target specific assembly writer or
3918 object file emitter. No other aspect of these options is defined by the IR.
3920 C type width Module Flags Metadata
3921 ----------------------------------
3923 The ARM backend emits a section into each generated object file describing the
3924 options that it was compiled with (in a compiler-independent way) to prevent
3925 linking incompatible objects, and to allow automatic library selection. Some
3926 of these options are not visible at the IR level, namely wchar_t width and enum
3929 To pass this information to the backend, these options are encoded in module
3930 flags metadata, using the following key-value pairs:
3940 - * 0 --- sizeof(wchar_t) == 4
3941 * 1 --- sizeof(wchar_t) == 2
3944 - * 0 --- Enums are at least as large as an ``int``.
3945 * 1 --- Enums are stored in the smallest integer type which can
3946 represent all of its values.
3948 For example, the following metadata section specifies that the module was
3949 compiled with a ``wchar_t`` width of 4 bytes, and the underlying type of an
3950 enum is the smallest type which can represent all of its values::
3952 !llvm.module.flags = !{!0, !1}
3953 !0 = !{i32 1, !"short_wchar", i32 1}
3954 !1 = !{i32 1, !"short_enum", i32 0}
3956 .. _intrinsicglobalvariables:
3958 Intrinsic Global Variables
3959 ==========================
3961 LLVM has a number of "magic" global variables that contain data that
3962 affect code generation or other IR semantics. These are documented here.
3963 All globals of this sort should have a section specified as
3964 "``llvm.metadata``". This section and all globals that start with
3965 "``llvm.``" are reserved for use by LLVM.
3969 The '``llvm.used``' Global Variable
3970 -----------------------------------
3972 The ``@llvm.used`` global is an array which has
3973 :ref:`appending linkage <linkage_appending>`. This array contains a list of
3974 pointers to named global variables, functions and aliases which may optionally
3975 have a pointer cast formed of bitcast or getelementptr. For example, a legal
3978 .. code-block:: llvm
3983 @llvm.used = appending global [2 x i8*] [
3985 i8* bitcast (i32* @Y to i8*)
3986 ], section "llvm.metadata"
3988 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
3989 and linker are required to treat the symbol as if there is a reference to the
3990 symbol that it cannot see (which is why they have to be named). For example, if
3991 a variable has internal linkage and no references other than that from the
3992 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
3993 references from inline asms and other things the compiler cannot "see", and
3994 corresponds to "``attribute((used))``" in GNU C.
3996 On some targets, the code generator must emit a directive to the
3997 assembler or object file to prevent the assembler and linker from
3998 molesting the symbol.
4000 .. _gv_llvmcompilerused:
4002 The '``llvm.compiler.used``' Global Variable
4003 --------------------------------------------
4005 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
4006 directive, except that it only prevents the compiler from touching the
4007 symbol. On targets that support it, this allows an intelligent linker to
4008 optimize references to the symbol without being impeded as it would be
4011 This is a rare construct that should only be used in rare circumstances,
4012 and should not be exposed to source languages.
4014 .. _gv_llvmglobalctors:
4016 The '``llvm.global_ctors``' Global Variable
4017 -------------------------------------------
4019 .. code-block:: llvm
4021 %0 = type { i32, void ()*, i8* }
4022 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor, i8* @data }]
4024 The ``@llvm.global_ctors`` array contains a list of constructor
4025 functions, priorities, and an optional associated global or function.
4026 The functions referenced by this array will be called in ascending order
4027 of priority (i.e. lowest first) when the module is loaded. The order of
4028 functions with the same priority is not defined.
4030 If the third field is present, non-null, and points to a global variable
4031 or function, the initializer function will only run if the associated
4032 data from the current module is not discarded.
4034 .. _llvmglobaldtors:
4036 The '``llvm.global_dtors``' Global Variable
4037 -------------------------------------------
4039 .. code-block:: llvm
4041 %0 = type { i32, void ()*, i8* }
4042 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor, i8* @data }]
4044 The ``@llvm.global_dtors`` array contains a list of destructor
4045 functions, priorities, and an optional associated global or function.
4046 The functions referenced by this array will be called in descending
4047 order of priority (i.e. highest first) when the module is unloaded. The
4048 order of functions with the same priority is not defined.
4050 If the third field is present, non-null, and points to a global variable
4051 or function, the destructor function will only run if the associated
4052 data from the current module is not discarded.
4054 Instruction Reference
4055 =====================
4057 The LLVM instruction set consists of several different classifications
4058 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
4059 instructions <binaryops>`, :ref:`bitwise binary
4060 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
4061 :ref:`other instructions <otherops>`.
4065 Terminator Instructions
4066 -----------------------
4068 As mentioned :ref:`previously <functionstructure>`, every basic block in a
4069 program ends with a "Terminator" instruction, which indicates which
4070 block should be executed after the current block is finished. These
4071 terminator instructions typically yield a '``void``' value: they produce
4072 control flow, not values (the one exception being the
4073 ':ref:`invoke <i_invoke>`' instruction).
4075 The terminator instructions are: ':ref:`ret <i_ret>`',
4076 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
4077 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
4078 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
4082 '``ret``' Instruction
4083 ^^^^^^^^^^^^^^^^^^^^^
4090 ret <type> <value> ; Return a value from a non-void function
4091 ret void ; Return from void function
4096 The '``ret``' instruction is used to return control flow (and optionally
4097 a value) from a function back to the caller.
4099 There are two forms of the '``ret``' instruction: one that returns a
4100 value and then causes control flow, and one that just causes control
4106 The '``ret``' instruction optionally accepts a single argument, the
4107 return value. The type of the return value must be a ':ref:`first
4108 class <t_firstclass>`' type.
4110 A function is not :ref:`well formed <wellformed>` if it it has a non-void
4111 return type and contains a '``ret``' instruction with no return value or
4112 a return value with a type that does not match its type, or if it has a
4113 void return type and contains a '``ret``' instruction with a return
4119 When the '``ret``' instruction is executed, control flow returns back to
4120 the calling function's context. If the caller is a
4121 ":ref:`call <i_call>`" instruction, execution continues at the
4122 instruction after the call. If the caller was an
4123 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
4124 beginning of the "normal" destination block. If the instruction returns
4125 a value, that value shall set the call or invoke instruction's return
4131 .. code-block:: llvm
4133 ret i32 5 ; Return an integer value of 5
4134 ret void ; Return from a void function
4135 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
4139 '``br``' Instruction
4140 ^^^^^^^^^^^^^^^^^^^^
4147 br i1 <cond>, label <iftrue>, label <iffalse>
4148 br label <dest> ; Unconditional branch
4153 The '``br``' instruction is used to cause control flow to transfer to a
4154 different basic block in the current function. There are two forms of
4155 this instruction, corresponding to a conditional branch and an
4156 unconditional branch.
4161 The conditional branch form of the '``br``' instruction takes a single
4162 '``i1``' value and two '``label``' values. The unconditional form of the
4163 '``br``' instruction takes a single '``label``' value as a target.
4168 Upon execution of a conditional '``br``' instruction, the '``i1``'
4169 argument is evaluated. If the value is ``true``, control flows to the
4170 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
4171 to the '``iffalse``' ``label`` argument.
4176 .. code-block:: llvm
4179 %cond = icmp eq i32 %a, %b
4180 br i1 %cond, label %IfEqual, label %IfUnequal
4188 '``switch``' Instruction
4189 ^^^^^^^^^^^^^^^^^^^^^^^^
4196 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
4201 The '``switch``' instruction is used to transfer control flow to one of
4202 several different places. It is a generalization of the '``br``'
4203 instruction, allowing a branch to occur to one of many possible
4209 The '``switch``' instruction uses three parameters: an integer
4210 comparison value '``value``', a default '``label``' destination, and an
4211 array of pairs of comparison value constants and '``label``'s. The table
4212 is not allowed to contain duplicate constant entries.
4217 The ``switch`` instruction specifies a table of values and destinations.
4218 When the '``switch``' instruction is executed, this table is searched
4219 for the given value. If the value is found, control flow is transferred
4220 to the corresponding destination; otherwise, control flow is transferred
4221 to the default destination.
4226 Depending on properties of the target machine and the particular
4227 ``switch`` instruction, this instruction may be code generated in
4228 different ways. For example, it could be generated as a series of
4229 chained conditional branches or with a lookup table.
4234 .. code-block:: llvm
4236 ; Emulate a conditional br instruction
4237 %Val = zext i1 %value to i32
4238 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
4240 ; Emulate an unconditional br instruction
4241 switch i32 0, label %dest [ ]
4243 ; Implement a jump table:
4244 switch i32 %val, label %otherwise [ i32 0, label %onzero
4246 i32 2, label %ontwo ]
4250 '``indirectbr``' Instruction
4251 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4258 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
4263 The '``indirectbr``' instruction implements an indirect branch to a
4264 label within the current function, whose address is specified by
4265 "``address``". Address must be derived from a
4266 :ref:`blockaddress <blockaddress>` constant.
4271 The '``address``' argument is the address of the label to jump to. The
4272 rest of the arguments indicate the full set of possible destinations
4273 that the address may point to. Blocks are allowed to occur multiple
4274 times in the destination list, though this isn't particularly useful.
4276 This destination list is required so that dataflow analysis has an
4277 accurate understanding of the CFG.
4282 Control transfers to the block specified in the address argument. All
4283 possible destination blocks must be listed in the label list, otherwise
4284 this instruction has undefined behavior. This implies that jumps to
4285 labels defined in other functions have undefined behavior as well.
4290 This is typically implemented with a jump through a register.
4295 .. code-block:: llvm
4297 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
4301 '``invoke``' Instruction
4302 ^^^^^^^^^^^^^^^^^^^^^^^^
4309 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
4310 to label <normal label> unwind label <exception label>
4315 The '``invoke``' instruction causes control to transfer to a specified
4316 function, with the possibility of control flow transfer to either the
4317 '``normal``' label or the '``exception``' label. If the callee function
4318 returns with the "``ret``" instruction, control flow will return to the
4319 "normal" label. If the callee (or any indirect callees) returns via the
4320 ":ref:`resume <i_resume>`" instruction or other exception handling
4321 mechanism, control is interrupted and continued at the dynamically
4322 nearest "exception" label.
4324 The '``exception``' label is a `landing
4325 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
4326 '``exception``' label is required to have the
4327 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
4328 information about the behavior of the program after unwinding happens,
4329 as its first non-PHI instruction. The restrictions on the
4330 "``landingpad``" instruction's tightly couples it to the "``invoke``"
4331 instruction, so that the important information contained within the
4332 "``landingpad``" instruction can't be lost through normal code motion.
4337 This instruction requires several arguments:
4339 #. The optional "cconv" marker indicates which :ref:`calling
4340 convention <callingconv>` the call should use. If none is
4341 specified, the call defaults to using C calling conventions.
4342 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
4343 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
4345 #. '``ptr to function ty``': shall be the signature of the pointer to
4346 function value being invoked. In most cases, this is a direct
4347 function invocation, but indirect ``invoke``'s are just as possible,
4348 branching off an arbitrary pointer to function value.
4349 #. '``function ptr val``': An LLVM value containing a pointer to a
4350 function to be invoked.
4351 #. '``function args``': argument list whose types match the function
4352 signature argument types and parameter attributes. All arguments must
4353 be of :ref:`first class <t_firstclass>` type. If the function signature
4354 indicates the function accepts a variable number of arguments, the
4355 extra arguments can be specified.
4356 #. '``normal label``': the label reached when the called function
4357 executes a '``ret``' instruction.
4358 #. '``exception label``': the label reached when a callee returns via
4359 the :ref:`resume <i_resume>` instruction or other exception handling
4361 #. The optional :ref:`function attributes <fnattrs>` list. Only
4362 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
4363 attributes are valid here.
4368 This instruction is designed to operate as a standard '``call``'
4369 instruction in most regards. The primary difference is that it
4370 establishes an association with a label, which is used by the runtime
4371 library to unwind the stack.
4373 This instruction is used in languages with destructors to ensure that
4374 proper cleanup is performed in the case of either a ``longjmp`` or a
4375 thrown exception. Additionally, this is important for implementation of
4376 '``catch``' clauses in high-level languages that support them.
4378 For the purposes of the SSA form, the definition of the value returned
4379 by the '``invoke``' instruction is deemed to occur on the edge from the
4380 current block to the "normal" label. If the callee unwinds then no
4381 return value is available.
4386 .. code-block:: llvm
4388 %retval = invoke i32 @Test(i32 15) to label %Continue
4389 unwind label %TestCleanup ; i32:retval set
4390 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
4391 unwind label %TestCleanup ; i32:retval set
4395 '``resume``' Instruction
4396 ^^^^^^^^^^^^^^^^^^^^^^^^
4403 resume <type> <value>
4408 The '``resume``' instruction is a terminator instruction that has no
4414 The '``resume``' instruction requires one argument, which must have the
4415 same type as the result of any '``landingpad``' instruction in the same
4421 The '``resume``' instruction resumes propagation of an existing
4422 (in-flight) exception whose unwinding was interrupted with a
4423 :ref:`landingpad <i_landingpad>` instruction.
4428 .. code-block:: llvm
4430 resume { i8*, i32 } %exn
4434 '``unreachable``' Instruction
4435 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4447 The '``unreachable``' instruction has no defined semantics. This
4448 instruction is used to inform the optimizer that a particular portion of
4449 the code is not reachable. This can be used to indicate that the code
4450 after a no-return function cannot be reached, and other facts.
4455 The '``unreachable``' instruction has no defined semantics.
4462 Binary operators are used to do most of the computation in a program.
4463 They require two operands of the same type, execute an operation on
4464 them, and produce a single value. The operands might represent multiple
4465 data, as is the case with the :ref:`vector <t_vector>` data type. The
4466 result value has the same type as its operands.
4468 There are several different binary operators:
4472 '``add``' Instruction
4473 ^^^^^^^^^^^^^^^^^^^^^
4480 <result> = add <ty> <op1>, <op2> ; yields ty:result
4481 <result> = add nuw <ty> <op1>, <op2> ; yields ty:result
4482 <result> = add nsw <ty> <op1>, <op2> ; yields ty:result
4483 <result> = add nuw nsw <ty> <op1>, <op2> ; yields ty:result
4488 The '``add``' instruction returns the sum of its two operands.
4493 The two arguments to the '``add``' instruction must be
4494 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4495 arguments must have identical types.
4500 The value produced is the integer sum of the two operands.
4502 If the sum has unsigned overflow, the result returned is the
4503 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4506 Because LLVM integers use a two's complement representation, this
4507 instruction is appropriate for both signed and unsigned integers.
4509 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4510 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4511 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
4512 unsigned and/or signed overflow, respectively, occurs.
4517 .. code-block:: llvm
4519 <result> = add i32 4, %var ; yields i32:result = 4 + %var
4523 '``fadd``' Instruction
4524 ^^^^^^^^^^^^^^^^^^^^^^
4531 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4536 The '``fadd``' instruction returns the sum of its two operands.
4541 The two arguments to the '``fadd``' instruction must be :ref:`floating
4542 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4543 Both arguments must have identical types.
4548 The value produced is the floating point sum of the two operands. This
4549 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
4550 which are optimization hints to enable otherwise unsafe floating point
4556 .. code-block:: llvm
4558 <result> = fadd float 4.0, %var ; yields float:result = 4.0 + %var
4560 '``sub``' Instruction
4561 ^^^^^^^^^^^^^^^^^^^^^
4568 <result> = sub <ty> <op1>, <op2> ; yields ty:result
4569 <result> = sub nuw <ty> <op1>, <op2> ; yields ty:result
4570 <result> = sub nsw <ty> <op1>, <op2> ; yields ty:result
4571 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields ty:result
4576 The '``sub``' instruction returns the difference of its two operands.
4578 Note that the '``sub``' instruction is used to represent the '``neg``'
4579 instruction present in most other intermediate representations.
4584 The two arguments to the '``sub``' instruction must be
4585 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4586 arguments must have identical types.
4591 The value produced is the integer difference of the two operands.
4593 If the difference has unsigned overflow, the result returned is the
4594 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4597 Because LLVM integers use a two's complement representation, this
4598 instruction is appropriate for both signed and unsigned integers.
4600 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4601 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4602 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
4603 unsigned and/or signed overflow, respectively, occurs.
4608 .. code-block:: llvm
4610 <result> = sub i32 4, %var ; yields i32:result = 4 - %var
4611 <result> = sub i32 0, %val ; yields i32:result = -%var
4615 '``fsub``' Instruction
4616 ^^^^^^^^^^^^^^^^^^^^^^
4623 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4628 The '``fsub``' instruction returns the difference of its two operands.
4630 Note that the '``fsub``' instruction is used to represent the '``fneg``'
4631 instruction present in most other intermediate representations.
4636 The two arguments to the '``fsub``' instruction must be :ref:`floating
4637 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4638 Both arguments must have identical types.
4643 The value produced is the floating point difference of the two operands.
4644 This instruction can also take any number of :ref:`fast-math
4645 flags <fastmath>`, which are optimization hints to enable otherwise
4646 unsafe floating point optimizations:
4651 .. code-block:: llvm
4653 <result> = fsub float 4.0, %var ; yields float:result = 4.0 - %var
4654 <result> = fsub float -0.0, %val ; yields float:result = -%var
4656 '``mul``' Instruction
4657 ^^^^^^^^^^^^^^^^^^^^^
4664 <result> = mul <ty> <op1>, <op2> ; yields ty:result
4665 <result> = mul nuw <ty> <op1>, <op2> ; yields ty:result
4666 <result> = mul nsw <ty> <op1>, <op2> ; yields ty:result
4667 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields ty:result
4672 The '``mul``' instruction returns the product of its two operands.
4677 The two arguments to the '``mul``' instruction must be
4678 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4679 arguments must have identical types.
4684 The value produced is the integer product of the two operands.
4686 If the result of the multiplication has unsigned overflow, the result
4687 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
4688 bit width of the result.
4690 Because LLVM integers use a two's complement representation, and the
4691 result is the same width as the operands, this instruction returns the
4692 correct result for both signed and unsigned integers. If a full product
4693 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
4694 sign-extended or zero-extended as appropriate to the width of the full
4697 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4698 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4699 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
4700 unsigned and/or signed overflow, respectively, occurs.
4705 .. code-block:: llvm
4707 <result> = mul i32 4, %var ; yields i32:result = 4 * %var
4711 '``fmul``' Instruction
4712 ^^^^^^^^^^^^^^^^^^^^^^
4719 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4724 The '``fmul``' instruction returns the product of its two operands.
4729 The two arguments to the '``fmul``' instruction must be :ref:`floating
4730 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4731 Both arguments must have identical types.
4736 The value produced is the floating point product of the two operands.
4737 This instruction can also take any number of :ref:`fast-math
4738 flags <fastmath>`, which are optimization hints to enable otherwise
4739 unsafe floating point optimizations:
4744 .. code-block:: llvm
4746 <result> = fmul float 4.0, %var ; yields float:result = 4.0 * %var
4748 '``udiv``' Instruction
4749 ^^^^^^^^^^^^^^^^^^^^^^
4756 <result> = udiv <ty> <op1>, <op2> ; yields ty:result
4757 <result> = udiv exact <ty> <op1>, <op2> ; yields ty:result
4762 The '``udiv``' instruction returns the quotient of its two operands.
4767 The two arguments to the '``udiv``' instruction must be
4768 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4769 arguments must have identical types.
4774 The value produced is the unsigned integer quotient of the two operands.
4776 Note that unsigned integer division and signed integer division are
4777 distinct operations; for signed integer division, use '``sdiv``'.
4779 Division by zero leads to undefined behavior.
4781 If the ``exact`` keyword is present, the result value of the ``udiv`` is
4782 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
4783 such, "((a udiv exact b) mul b) == a").
4788 .. code-block:: llvm
4790 <result> = udiv i32 4, %var ; yields i32:result = 4 / %var
4792 '``sdiv``' Instruction
4793 ^^^^^^^^^^^^^^^^^^^^^^
4800 <result> = sdiv <ty> <op1>, <op2> ; yields ty:result
4801 <result> = sdiv exact <ty> <op1>, <op2> ; yields ty:result
4806 The '``sdiv``' instruction returns the quotient of its two operands.
4811 The two arguments to the '``sdiv``' instruction must be
4812 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4813 arguments must have identical types.
4818 The value produced is the signed integer quotient of the two operands
4819 rounded towards zero.
4821 Note that signed integer division and unsigned integer division are
4822 distinct operations; for unsigned integer division, use '``udiv``'.
4824 Division by zero leads to undefined behavior. Overflow also leads to
4825 undefined behavior; this is a rare case, but can occur, for example, by
4826 doing a 32-bit division of -2147483648 by -1.
4828 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
4829 a :ref:`poison value <poisonvalues>` if the result would be rounded.
4834 .. code-block:: llvm
4836 <result> = sdiv i32 4, %var ; yields i32:result = 4 / %var
4840 '``fdiv``' Instruction
4841 ^^^^^^^^^^^^^^^^^^^^^^
4848 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4853 The '``fdiv``' instruction returns the quotient of its two operands.
4858 The two arguments to the '``fdiv``' instruction must be :ref:`floating
4859 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4860 Both arguments must have identical types.
4865 The value produced is the floating point quotient of the two operands.
4866 This instruction can also take any number of :ref:`fast-math
4867 flags <fastmath>`, which are optimization hints to enable otherwise
4868 unsafe floating point optimizations:
4873 .. code-block:: llvm
4875 <result> = fdiv float 4.0, %var ; yields float:result = 4.0 / %var
4877 '``urem``' Instruction
4878 ^^^^^^^^^^^^^^^^^^^^^^
4885 <result> = urem <ty> <op1>, <op2> ; yields ty:result
4890 The '``urem``' instruction returns the remainder from the unsigned
4891 division of its two arguments.
4896 The two arguments to the '``urem``' instruction must be
4897 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4898 arguments must have identical types.
4903 This instruction returns the unsigned integer *remainder* of a division.
4904 This instruction always performs an unsigned division to get the
4907 Note that unsigned integer remainder and signed integer remainder are
4908 distinct operations; for signed integer remainder, use '``srem``'.
4910 Taking the remainder of a division by zero leads to undefined behavior.
4915 .. code-block:: llvm
4917 <result> = urem i32 4, %var ; yields i32:result = 4 % %var
4919 '``srem``' Instruction
4920 ^^^^^^^^^^^^^^^^^^^^^^
4927 <result> = srem <ty> <op1>, <op2> ; yields ty:result
4932 The '``srem``' instruction returns the remainder from the signed
4933 division of its two operands. This instruction can also take
4934 :ref:`vector <t_vector>` versions of the values in which case the elements
4940 The two arguments to the '``srem``' instruction must be
4941 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4942 arguments must have identical types.
4947 This instruction returns the *remainder* of a division (where the result
4948 is either zero or has the same sign as the dividend, ``op1``), not the
4949 *modulo* operator (where the result is either zero or has the same sign
4950 as the divisor, ``op2``) of a value. For more information about the
4951 difference, see `The Math
4952 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
4953 table of how this is implemented in various languages, please see
4955 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
4957 Note that signed integer remainder and unsigned integer remainder are
4958 distinct operations; for unsigned integer remainder, use '``urem``'.
4960 Taking the remainder of a division by zero leads to undefined behavior.
4961 Overflow also leads to undefined behavior; this is a rare case, but can
4962 occur, for example, by taking the remainder of a 32-bit division of
4963 -2147483648 by -1. (The remainder doesn't actually overflow, but this
4964 rule lets srem be implemented using instructions that return both the
4965 result of the division and the remainder.)
4970 .. code-block:: llvm
4972 <result> = srem i32 4, %var ; yields i32:result = 4 % %var
4976 '``frem``' Instruction
4977 ^^^^^^^^^^^^^^^^^^^^^^
4984 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4989 The '``frem``' instruction returns the remainder from the division of
4995 The two arguments to the '``frem``' instruction must be :ref:`floating
4996 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4997 Both arguments must have identical types.
5002 This instruction returns the *remainder* of a division. The remainder
5003 has the same sign as the dividend. This instruction can also take any
5004 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
5005 to enable otherwise unsafe floating point optimizations:
5010 .. code-block:: llvm
5012 <result> = frem float 4.0, %var ; yields float:result = 4.0 % %var
5016 Bitwise Binary Operations
5017 -------------------------
5019 Bitwise binary operators are used to do various forms of bit-twiddling
5020 in a program. They are generally very efficient instructions and can
5021 commonly be strength reduced from other instructions. They require two
5022 operands of the same type, execute an operation on them, and produce a
5023 single value. The resulting value is the same type as its operands.
5025 '``shl``' Instruction
5026 ^^^^^^^^^^^^^^^^^^^^^
5033 <result> = shl <ty> <op1>, <op2> ; yields ty:result
5034 <result> = shl nuw <ty> <op1>, <op2> ; yields ty:result
5035 <result> = shl nsw <ty> <op1>, <op2> ; yields ty:result
5036 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields ty:result
5041 The '``shl``' instruction returns the first operand shifted to the left
5042 a specified number of bits.
5047 Both arguments to the '``shl``' instruction must be the same
5048 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
5049 '``op2``' is treated as an unsigned value.
5054 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
5055 where ``n`` is the width of the result. If ``op2`` is (statically or
5056 dynamically) negative or equal to or larger than the number of bits in
5057 ``op1``, the result is undefined. If the arguments are vectors, each
5058 vector element of ``op1`` is shifted by the corresponding shift amount
5061 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
5062 value <poisonvalues>` if it shifts out any non-zero bits. If the
5063 ``nsw`` keyword is present, then the shift produces a :ref:`poison
5064 value <poisonvalues>` if it shifts out any bits that disagree with the
5065 resultant sign bit. As such, NUW/NSW have the same semantics as they
5066 would if the shift were expressed as a mul instruction with the same
5067 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
5072 .. code-block:: llvm
5074 <result> = shl i32 4, %var ; yields i32: 4 << %var
5075 <result> = shl i32 4, 2 ; yields i32: 16
5076 <result> = shl i32 1, 10 ; yields i32: 1024
5077 <result> = shl i32 1, 32 ; undefined
5078 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
5080 '``lshr``' Instruction
5081 ^^^^^^^^^^^^^^^^^^^^^^
5088 <result> = lshr <ty> <op1>, <op2> ; yields ty:result
5089 <result> = lshr exact <ty> <op1>, <op2> ; yields ty:result
5094 The '``lshr``' instruction (logical shift right) returns the first
5095 operand shifted to the right a specified number of bits with zero fill.
5100 Both arguments to the '``lshr``' instruction must be the same
5101 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
5102 '``op2``' is treated as an unsigned value.
5107 This instruction always performs a logical shift right operation. The
5108 most significant bits of the result will be filled with zero bits after
5109 the shift. If ``op2`` is (statically or dynamically) equal to or larger
5110 than the number of bits in ``op1``, the result is undefined. If the
5111 arguments are vectors, each vector element of ``op1`` is shifted by the
5112 corresponding shift amount in ``op2``.
5114 If the ``exact`` keyword is present, the result value of the ``lshr`` is
5115 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
5121 .. code-block:: llvm
5123 <result> = lshr i32 4, 1 ; yields i32:result = 2
5124 <result> = lshr i32 4, 2 ; yields i32:result = 1
5125 <result> = lshr i8 4, 3 ; yields i8:result = 0
5126 <result> = lshr i8 -2, 1 ; yields i8:result = 0x7F
5127 <result> = lshr i32 1, 32 ; undefined
5128 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
5130 '``ashr``' Instruction
5131 ^^^^^^^^^^^^^^^^^^^^^^
5138 <result> = ashr <ty> <op1>, <op2> ; yields ty:result
5139 <result> = ashr exact <ty> <op1>, <op2> ; yields ty:result
5144 The '``ashr``' instruction (arithmetic shift right) returns the first
5145 operand shifted to the right a specified number of bits with sign
5151 Both arguments to the '``ashr``' instruction must be the same
5152 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
5153 '``op2``' is treated as an unsigned value.
5158 This instruction always performs an arithmetic shift right operation,
5159 The most significant bits of the result will be filled with the sign bit
5160 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
5161 than the number of bits in ``op1``, the result is undefined. If the
5162 arguments are vectors, each vector element of ``op1`` is shifted by the
5163 corresponding shift amount in ``op2``.
5165 If the ``exact`` keyword is present, the result value of the ``ashr`` is
5166 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
5172 .. code-block:: llvm
5174 <result> = ashr i32 4, 1 ; yields i32:result = 2
5175 <result> = ashr i32 4, 2 ; yields i32:result = 1
5176 <result> = ashr i8 4, 3 ; yields i8:result = 0
5177 <result> = ashr i8 -2, 1 ; yields i8:result = -1
5178 <result> = ashr i32 1, 32 ; undefined
5179 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
5181 '``and``' Instruction
5182 ^^^^^^^^^^^^^^^^^^^^^
5189 <result> = and <ty> <op1>, <op2> ; yields ty:result
5194 The '``and``' instruction returns the bitwise logical and of its two
5200 The two arguments to the '``and``' instruction must be
5201 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5202 arguments must have identical types.
5207 The truth table used for the '``and``' instruction is:
5224 .. code-block:: llvm
5226 <result> = and i32 4, %var ; yields i32:result = 4 & %var
5227 <result> = and i32 15, 40 ; yields i32:result = 8
5228 <result> = and i32 4, 8 ; yields i32:result = 0
5230 '``or``' Instruction
5231 ^^^^^^^^^^^^^^^^^^^^
5238 <result> = or <ty> <op1>, <op2> ; yields ty:result
5243 The '``or``' instruction returns the bitwise logical inclusive or of its
5249 The two arguments to the '``or``' instruction must be
5250 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5251 arguments must have identical types.
5256 The truth table used for the '``or``' instruction is:
5275 <result> = or i32 4, %var ; yields i32:result = 4 | %var
5276 <result> = or i32 15, 40 ; yields i32:result = 47
5277 <result> = or i32 4, 8 ; yields i32:result = 12
5279 '``xor``' Instruction
5280 ^^^^^^^^^^^^^^^^^^^^^
5287 <result> = xor <ty> <op1>, <op2> ; yields ty:result
5292 The '``xor``' instruction returns the bitwise logical exclusive or of
5293 its two operands. The ``xor`` is used to implement the "one's
5294 complement" operation, which is the "~" operator in C.
5299 The two arguments to the '``xor``' instruction must be
5300 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5301 arguments must have identical types.
5306 The truth table used for the '``xor``' instruction is:
5323 .. code-block:: llvm
5325 <result> = xor i32 4, %var ; yields i32:result = 4 ^ %var
5326 <result> = xor i32 15, 40 ; yields i32:result = 39
5327 <result> = xor i32 4, 8 ; yields i32:result = 12
5328 <result> = xor i32 %V, -1 ; yields i32:result = ~%V
5333 LLVM supports several instructions to represent vector operations in a
5334 target-independent manner. These instructions cover the element-access
5335 and vector-specific operations needed to process vectors effectively.
5336 While LLVM does directly support these vector operations, many
5337 sophisticated algorithms will want to use target-specific intrinsics to
5338 take full advantage of a specific target.
5340 .. _i_extractelement:
5342 '``extractelement``' Instruction
5343 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5350 <result> = extractelement <n x <ty>> <val>, <ty2> <idx> ; yields <ty>
5355 The '``extractelement``' instruction extracts a single scalar element
5356 from a vector at a specified index.
5361 The first operand of an '``extractelement``' instruction is a value of
5362 :ref:`vector <t_vector>` type. The second operand is an index indicating
5363 the position from which to extract the element. The index may be a
5364 variable of any integer type.
5369 The result is a scalar of the same type as the element type of ``val``.
5370 Its value is the value at position ``idx`` of ``val``. If ``idx``
5371 exceeds the length of ``val``, the results are undefined.
5376 .. code-block:: llvm
5378 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
5380 .. _i_insertelement:
5382 '``insertelement``' Instruction
5383 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5390 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, <ty2> <idx> ; yields <n x <ty>>
5395 The '``insertelement``' instruction inserts a scalar element into a
5396 vector at a specified index.
5401 The first operand of an '``insertelement``' instruction is a value of
5402 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
5403 type must equal the element type of the first operand. The third operand
5404 is an index indicating the position at which to insert the value. The
5405 index may be a variable of any integer type.
5410 The result is a vector of the same type as ``val``. Its element values
5411 are those of ``val`` except at position ``idx``, where it gets the value
5412 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
5418 .. code-block:: llvm
5420 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
5422 .. _i_shufflevector:
5424 '``shufflevector``' Instruction
5425 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5432 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
5437 The '``shufflevector``' instruction constructs a permutation of elements
5438 from two input vectors, returning a vector with the same element type as
5439 the input and length that is the same as the shuffle mask.
5444 The first two operands of a '``shufflevector``' instruction are vectors
5445 with the same type. The third argument is a shuffle mask whose element
5446 type is always 'i32'. The result of the instruction is a vector whose
5447 length is the same as the shuffle mask and whose element type is the
5448 same as the element type of the first two operands.
5450 The shuffle mask operand is required to be a constant vector with either
5451 constant integer or undef values.
5456 The elements of the two input vectors are numbered from left to right
5457 across both of the vectors. The shuffle mask operand specifies, for each
5458 element of the result vector, which element of the two input vectors the
5459 result element gets. The element selector may be undef (meaning "don't
5460 care") and the second operand may be undef if performing a shuffle from
5466 .. code-block:: llvm
5468 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
5469 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
5470 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
5471 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
5472 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
5473 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
5474 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
5475 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
5477 Aggregate Operations
5478 --------------------
5480 LLVM supports several instructions for working with
5481 :ref:`aggregate <t_aggregate>` values.
5485 '``extractvalue``' Instruction
5486 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5493 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
5498 The '``extractvalue``' instruction extracts the value of a member field
5499 from an :ref:`aggregate <t_aggregate>` value.
5504 The first operand of an '``extractvalue``' instruction is a value of
5505 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
5506 constant indices to specify which value to extract in a similar manner
5507 as indices in a '``getelementptr``' instruction.
5509 The major differences to ``getelementptr`` indexing are:
5511 - Since the value being indexed is not a pointer, the first index is
5512 omitted and assumed to be zero.
5513 - At least one index must be specified.
5514 - Not only struct indices but also array indices must be in bounds.
5519 The result is the value at the position in the aggregate specified by
5525 .. code-block:: llvm
5527 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
5531 '``insertvalue``' Instruction
5532 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5539 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
5544 The '``insertvalue``' instruction inserts a value into a member field in
5545 an :ref:`aggregate <t_aggregate>` value.
5550 The first operand of an '``insertvalue``' instruction is a value of
5551 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
5552 a first-class value to insert. The following operands are constant
5553 indices indicating the position at which to insert the value in a
5554 similar manner as indices in a '``extractvalue``' instruction. The value
5555 to insert must have the same type as the value identified by the
5561 The result is an aggregate of the same type as ``val``. Its value is
5562 that of ``val`` except that the value at the position specified by the
5563 indices is that of ``elt``.
5568 .. code-block:: llvm
5570 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
5571 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
5572 %agg3 = insertvalue {i32, {float}} undef, float %val, 1, 0 ; yields {i32 undef, {float %val}}
5576 Memory Access and Addressing Operations
5577 ---------------------------------------
5579 A key design point of an SSA-based representation is how it represents
5580 memory. In LLVM, no memory locations are in SSA form, which makes things
5581 very simple. This section describes how to read, write, and allocate
5586 '``alloca``' Instruction
5587 ^^^^^^^^^^^^^^^^^^^^^^^^
5594 <result> = alloca [inalloca] <type> [, <ty> <NumElements>] [, align <alignment>] ; yields type*:result
5599 The '``alloca``' instruction allocates memory on the stack frame of the
5600 currently executing function, to be automatically released when this
5601 function returns to its caller. The object is always allocated in the
5602 generic address space (address space zero).
5607 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
5608 bytes of memory on the runtime stack, returning a pointer of the
5609 appropriate type to the program. If "NumElements" is specified, it is
5610 the number of elements allocated, otherwise "NumElements" is defaulted
5611 to be one. If a constant alignment is specified, the value result of the
5612 allocation is guaranteed to be aligned to at least that boundary. The
5613 alignment may not be greater than ``1 << 29``. If not specified, or if
5614 zero, the target can choose to align the allocation on any convenient
5615 boundary compatible with the type.
5617 '``type``' may be any sized type.
5622 Memory is allocated; a pointer is returned. The operation is undefined
5623 if there is insufficient stack space for the allocation. '``alloca``'d
5624 memory is automatically released when the function returns. The
5625 '``alloca``' instruction is commonly used to represent automatic
5626 variables that must have an address available. When the function returns
5627 (either with the ``ret`` or ``resume`` instructions), the memory is
5628 reclaimed. Allocating zero bytes is legal, but the result is undefined.
5629 The order in which memory is allocated (ie., which way the stack grows)
5635 .. code-block:: llvm
5637 %ptr = alloca i32 ; yields i32*:ptr
5638 %ptr = alloca i32, i32 4 ; yields i32*:ptr
5639 %ptr = alloca i32, i32 4, align 1024 ; yields i32*:ptr
5640 %ptr = alloca i32, align 1024 ; yields i32*:ptr
5644 '``load``' Instruction
5645 ^^^^^^^^^^^^^^^^^^^^^^
5652 <result> = load [volatile] <ty>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>][, !nonnull !<index>]
5653 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
5654 !<index> = !{ i32 1 }
5659 The '``load``' instruction is used to read from memory.
5664 The argument to the ``load`` instruction specifies the memory address
5665 from which to load. The type specified must be a :ref:`first
5666 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
5667 then the optimizer is not allowed to modify the number or order of
5668 execution of this ``load`` with other :ref:`volatile
5669 operations <volatile>`.
5671 If the ``load`` is marked as ``atomic``, it takes an extra
5672 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5673 ``release`` and ``acq_rel`` orderings are not valid on ``load``
5674 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5675 when they may see multiple atomic stores. The type of the pointee must
5676 be an integer type whose bit width is a power of two greater than or
5677 equal to eight and less than or equal to a target-specific size limit.
5678 ``align`` must be explicitly specified on atomic loads, and the load has
5679 undefined behavior if the alignment is not set to a value which is at
5680 least the size in bytes of the pointee. ``!nontemporal`` does not have
5681 any defined semantics for atomic loads.
5683 The optional constant ``align`` argument specifies the alignment of the
5684 operation (that is, the alignment of the memory address). A value of 0
5685 or an omitted ``align`` argument means that the operation has the ABI
5686 alignment for the target. It is the responsibility of the code emitter
5687 to ensure that the alignment information is correct. Overestimating the
5688 alignment results in undefined behavior. Underestimating the alignment
5689 may produce less efficient code. An alignment of 1 is always safe. The
5690 maximum possible alignment is ``1 << 29``.
5692 The optional ``!nontemporal`` metadata must reference a single
5693 metadata name ``<index>`` corresponding to a metadata node with one
5694 ``i32`` entry of value 1. The existence of the ``!nontemporal``
5695 metadata on the instruction tells the optimizer and code generator
5696 that this load is not expected to be reused in the cache. The code
5697 generator may select special instructions to save cache bandwidth, such
5698 as the ``MOVNT`` instruction on x86.
5700 The optional ``!invariant.load`` metadata must reference a single
5701 metadata name ``<index>`` corresponding to a metadata node with no
5702 entries. The existence of the ``!invariant.load`` metadata on the
5703 instruction tells the optimizer and code generator that the address
5704 operand to this load points to memory which can be assumed unchanged.
5705 Being invariant does not imply that a location is dereferenceable,
5706 but it does imply that once the location is known dereferenceable
5707 its value is henceforth unchanging.
5709 The optional ``!nonnull`` metadata must reference a single
5710 metadata name ``<index>`` corresponding to a metadata node with no
5711 entries. The existence of the ``!nonnull`` metadata on the
5712 instruction tells the optimizer that the value loaded is known to
5713 never be null. This is analogous to the ''nonnull'' attribute
5714 on parameters and return values. This metadata can only be applied
5715 to loads of a pointer type.
5720 The location of memory pointed to is loaded. If the value being loaded
5721 is of scalar type then the number of bytes read does not exceed the
5722 minimum number of bytes needed to hold all bits of the type. For
5723 example, loading an ``i24`` reads at most three bytes. When loading a
5724 value of a type like ``i20`` with a size that is not an integral number
5725 of bytes, the result is undefined if the value was not originally
5726 written using a store of the same type.
5731 .. code-block:: llvm
5733 %ptr = alloca i32 ; yields i32*:ptr
5734 store i32 3, i32* %ptr ; yields void
5735 %val = load i32, i32* %ptr ; yields i32:val = i32 3
5739 '``store``' Instruction
5740 ^^^^^^^^^^^^^^^^^^^^^^^
5747 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields void
5748 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields void
5753 The '``store``' instruction is used to write to memory.
5758 There are two arguments to the ``store`` instruction: a value to store
5759 and an address at which to store it. The type of the ``<pointer>``
5760 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
5761 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
5762 then the optimizer is not allowed to modify the number or order of
5763 execution of this ``store`` with other :ref:`volatile
5764 operations <volatile>`.
5766 If the ``store`` is marked as ``atomic``, it takes an extra
5767 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5768 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
5769 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5770 when they may see multiple atomic stores. The type of the pointee must
5771 be an integer type whose bit width is a power of two greater than or
5772 equal to eight and less than or equal to a target-specific size limit.
5773 ``align`` must be explicitly specified on atomic stores, and the store
5774 has undefined behavior if the alignment is not set to a value which is
5775 at least the size in bytes of the pointee. ``!nontemporal`` does not
5776 have any defined semantics for atomic stores.
5778 The optional constant ``align`` argument specifies the alignment of the
5779 operation (that is, the alignment of the memory address). A value of 0
5780 or an omitted ``align`` argument means that the operation has the ABI
5781 alignment for the target. It is the responsibility of the code emitter
5782 to ensure that the alignment information is correct. Overestimating the
5783 alignment results in undefined behavior. Underestimating the
5784 alignment may produce less efficient code. An alignment of 1 is always
5785 safe. The maximum possible alignment is ``1 << 29``.
5787 The optional ``!nontemporal`` metadata must reference a single metadata
5788 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
5789 value 1. The existence of the ``!nontemporal`` metadata on the instruction
5790 tells the optimizer and code generator that this load is not expected to
5791 be reused in the cache. The code generator may select special
5792 instructions to save cache bandwidth, such as the MOVNT instruction on
5798 The contents of memory are updated to contain ``<value>`` at the
5799 location specified by the ``<pointer>`` operand. If ``<value>`` is
5800 of scalar type then the number of bytes written does not exceed the
5801 minimum number of bytes needed to hold all bits of the type. For
5802 example, storing an ``i24`` writes at most three bytes. When writing a
5803 value of a type like ``i20`` with a size that is not an integral number
5804 of bytes, it is unspecified what happens to the extra bits that do not
5805 belong to the type, but they will typically be overwritten.
5810 .. code-block:: llvm
5812 %ptr = alloca i32 ; yields i32*:ptr
5813 store i32 3, i32* %ptr ; yields void
5814 %val = load i32* %ptr ; yields i32:val = i32 3
5818 '``fence``' Instruction
5819 ^^^^^^^^^^^^^^^^^^^^^^^
5826 fence [singlethread] <ordering> ; yields void
5831 The '``fence``' instruction is used to introduce happens-before edges
5837 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
5838 defines what *synchronizes-with* edges they add. They can only be given
5839 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
5844 A fence A which has (at least) ``release`` ordering semantics
5845 *synchronizes with* a fence B with (at least) ``acquire`` ordering
5846 semantics if and only if there exist atomic operations X and Y, both
5847 operating on some atomic object M, such that A is sequenced before X, X
5848 modifies M (either directly or through some side effect of a sequence
5849 headed by X), Y is sequenced before B, and Y observes M. This provides a
5850 *happens-before* dependency between A and B. Rather than an explicit
5851 ``fence``, one (but not both) of the atomic operations X or Y might
5852 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
5853 still *synchronize-with* the explicit ``fence`` and establish the
5854 *happens-before* edge.
5856 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
5857 ``acquire`` and ``release`` semantics specified above, participates in
5858 the global program order of other ``seq_cst`` operations and/or fences.
5860 The optional ":ref:`singlethread <singlethread>`" argument specifies
5861 that the fence only synchronizes with other fences in the same thread.
5862 (This is useful for interacting with signal handlers.)
5867 .. code-block:: llvm
5869 fence acquire ; yields void
5870 fence singlethread seq_cst ; yields void
5874 '``cmpxchg``' Instruction
5875 ^^^^^^^^^^^^^^^^^^^^^^^^^
5882 cmpxchg [weak] [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <success ordering> <failure ordering> ; yields { ty, i1 }
5887 The '``cmpxchg``' instruction is used to atomically modify memory. It
5888 loads a value in memory and compares it to a given value. If they are
5889 equal, it tries to store a new value into the memory.
5894 There are three arguments to the '``cmpxchg``' instruction: an address
5895 to operate on, a value to compare to the value currently be at that
5896 address, and a new value to place at that address if the compared values
5897 are equal. The type of '<cmp>' must be an integer type whose bit width
5898 is a power of two greater than or equal to eight and less than or equal
5899 to a target-specific size limit. '<cmp>' and '<new>' must have the same
5900 type, and the type of '<pointer>' must be a pointer to that type. If the
5901 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
5902 to modify the number or order of execution of this ``cmpxchg`` with
5903 other :ref:`volatile operations <volatile>`.
5905 The success and failure :ref:`ordering <ordering>` arguments specify how this
5906 ``cmpxchg`` synchronizes with other atomic operations. Both ordering parameters
5907 must be at least ``monotonic``, the ordering constraint on failure must be no
5908 stronger than that on success, and the failure ordering cannot be either
5909 ``release`` or ``acq_rel``.
5911 The optional "``singlethread``" argument declares that the ``cmpxchg``
5912 is only atomic with respect to code (usually signal handlers) running in
5913 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
5914 respect to all other code in the system.
5916 The pointer passed into cmpxchg must have alignment greater than or
5917 equal to the size in memory of the operand.
5922 The contents of memory at the location specified by the '``<pointer>``' operand
5923 is read and compared to '``<cmp>``'; if the read value is the equal, the
5924 '``<new>``' is written. The original value at the location is returned, together
5925 with a flag indicating success (true) or failure (false).
5927 If the cmpxchg operation is marked as ``weak`` then a spurious failure is
5928 permitted: the operation may not write ``<new>`` even if the comparison
5931 If the cmpxchg operation is strong (the default), the i1 value is 1 if and only
5932 if the value loaded equals ``cmp``.
5934 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose of
5935 identifying release sequences. A failed ``cmpxchg`` is equivalent to an atomic
5936 load with an ordering parameter determined the second ordering parameter.
5941 .. code-block:: llvm
5944 %orig = atomic load i32, i32* %ptr unordered ; yields i32
5948 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
5949 %squared = mul i32 %cmp, %cmp
5950 %val_success = cmpxchg i32* %ptr, i32 %cmp, i32 %squared acq_rel monotonic ; yields { i32, i1 }
5951 %value_loaded = extractvalue { i32, i1 } %val_success, 0
5952 %success = extractvalue { i32, i1 } %val_success, 1
5953 br i1 %success, label %done, label %loop
5960 '``atomicrmw``' Instruction
5961 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
5968 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields ty
5973 The '``atomicrmw``' instruction is used to atomically modify memory.
5978 There are three arguments to the '``atomicrmw``' instruction: an
5979 operation to apply, an address whose value to modify, an argument to the
5980 operation. The operation must be one of the following keywords:
5994 The type of '<value>' must be an integer type whose bit width is a power
5995 of two greater than or equal to eight and less than or equal to a
5996 target-specific size limit. The type of the '``<pointer>``' operand must
5997 be a pointer to that type. If the ``atomicrmw`` is marked as
5998 ``volatile``, then the optimizer is not allowed to modify the number or
5999 order of execution of this ``atomicrmw`` with other :ref:`volatile
6000 operations <volatile>`.
6005 The contents of memory at the location specified by the '``<pointer>``'
6006 operand are atomically read, modified, and written back. The original
6007 value at the location is returned. The modification is specified by the
6010 - xchg: ``*ptr = val``
6011 - add: ``*ptr = *ptr + val``
6012 - sub: ``*ptr = *ptr - val``
6013 - and: ``*ptr = *ptr & val``
6014 - nand: ``*ptr = ~(*ptr & val)``
6015 - or: ``*ptr = *ptr | val``
6016 - xor: ``*ptr = *ptr ^ val``
6017 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
6018 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
6019 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
6021 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
6027 .. code-block:: llvm
6029 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields i32
6031 .. _i_getelementptr:
6033 '``getelementptr``' Instruction
6034 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6041 <result> = getelementptr <ty>, <ty>* <ptrval>{, <ty> <idx>}*
6042 <result> = getelementptr inbounds <ty>, <ty>* <ptrval>{, <ty> <idx>}*
6043 <result> = getelementptr <ty>, <ptr vector> <ptrval>, <vector index type> <idx>
6048 The '``getelementptr``' instruction is used to get the address of a
6049 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
6050 address calculation only and does not access memory.
6055 The first argument is always a type used as the basis for the calculations.
6056 The second argument is always a pointer or a vector of pointers, and is the
6057 base address to start from. The remaining arguments are indices
6058 that indicate which of the elements of the aggregate object are indexed.
6059 The interpretation of each index is dependent on the type being indexed
6060 into. The first index always indexes the pointer value given as the
6061 first argument, the second index indexes a value of the type pointed to
6062 (not necessarily the value directly pointed to, since the first index
6063 can be non-zero), etc. The first type indexed into must be a pointer
6064 value, subsequent types can be arrays, vectors, and structs. Note that
6065 subsequent types being indexed into can never be pointers, since that
6066 would require loading the pointer before continuing calculation.
6068 The type of each index argument depends on the type it is indexing into.
6069 When indexing into a (optionally packed) structure, only ``i32`` integer
6070 **constants** are allowed (when using a vector of indices they must all
6071 be the **same** ``i32`` integer constant). When indexing into an array,
6072 pointer or vector, integers of any width are allowed, and they are not
6073 required to be constant. These integers are treated as signed values
6076 For example, let's consider a C code fragment and how it gets compiled
6092 int *foo(struct ST *s) {
6093 return &s[1].Z.B[5][13];
6096 The LLVM code generated by Clang is:
6098 .. code-block:: llvm
6100 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
6101 %struct.ST = type { i32, double, %struct.RT }
6103 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
6105 %arrayidx = getelementptr inbounds %struct.ST, %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
6112 In the example above, the first index is indexing into the
6113 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
6114 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
6115 indexes into the third element of the structure, yielding a
6116 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
6117 structure. The third index indexes into the second element of the
6118 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
6119 dimensions of the array are subscripted into, yielding an '``i32``'
6120 type. The '``getelementptr``' instruction returns a pointer to this
6121 element, thus computing a value of '``i32*``' type.
6123 Note that it is perfectly legal to index partially through a structure,
6124 returning a pointer to an inner element. Because of this, the LLVM code
6125 for the given testcase is equivalent to:
6127 .. code-block:: llvm
6129 define i32* @foo(%struct.ST* %s) {
6130 %t1 = getelementptr %struct.ST, %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
6131 %t2 = getelementptr %struct.ST, %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
6132 %t3 = getelementptr %struct.RT, %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
6133 %t4 = getelementptr [10 x [20 x i32]], [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
6134 %t5 = getelementptr [20 x i32], [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
6138 If the ``inbounds`` keyword is present, the result value of the
6139 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
6140 pointer is not an *in bounds* address of an allocated object, or if any
6141 of the addresses that would be formed by successive addition of the
6142 offsets implied by the indices to the base address with infinitely
6143 precise signed arithmetic are not an *in bounds* address of that
6144 allocated object. The *in bounds* addresses for an allocated object are
6145 all the addresses that point into the object, plus the address one byte
6146 past the end. In cases where the base is a vector of pointers the
6147 ``inbounds`` keyword applies to each of the computations element-wise.
6149 If the ``inbounds`` keyword is not present, the offsets are added to the
6150 base address with silently-wrapping two's complement arithmetic. If the
6151 offsets have a different width from the pointer, they are sign-extended
6152 or truncated to the width of the pointer. The result value of the
6153 ``getelementptr`` may be outside the object pointed to by the base
6154 pointer. The result value may not necessarily be used to access memory
6155 though, even if it happens to point into allocated storage. See the
6156 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
6159 The getelementptr instruction is often confusing. For some more insight
6160 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
6165 .. code-block:: llvm
6167 ; yields [12 x i8]*:aptr
6168 %aptr = getelementptr {i32, [12 x i8]}, {i32, [12 x i8]}* %saptr, i64 0, i32 1
6170 %vptr = getelementptr {i32, <2 x i8>}, {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
6172 %eptr = getelementptr [12 x i8], [12 x i8]* %aptr, i64 0, i32 1
6174 %iptr = getelementptr [10 x i32], [10 x i32]* @arr, i16 0, i16 0
6176 In cases where the pointer argument is a vector of pointers, each index
6177 must be a vector with the same number of elements. For example:
6179 .. code-block:: llvm
6181 %A = getelementptr i8, <4 x i8*> %ptrs, <4 x i64> %offsets,
6183 Conversion Operations
6184 ---------------------
6186 The instructions in this category are the conversion instructions
6187 (casting) which all take a single operand and a type. They perform
6188 various bit conversions on the operand.
6190 '``trunc .. to``' Instruction
6191 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6198 <result> = trunc <ty> <value> to <ty2> ; yields ty2
6203 The '``trunc``' instruction truncates its operand to the type ``ty2``.
6208 The '``trunc``' instruction takes a value to trunc, and a type to trunc
6209 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
6210 of the same number of integers. The bit size of the ``value`` must be
6211 larger than the bit size of the destination type, ``ty2``. Equal sized
6212 types are not allowed.
6217 The '``trunc``' instruction truncates the high order bits in ``value``
6218 and converts the remaining bits to ``ty2``. Since the source size must
6219 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
6220 It will always truncate bits.
6225 .. code-block:: llvm
6227 %X = trunc i32 257 to i8 ; yields i8:1
6228 %Y = trunc i32 123 to i1 ; yields i1:true
6229 %Z = trunc i32 122 to i1 ; yields i1:false
6230 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
6232 '``zext .. to``' Instruction
6233 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6240 <result> = zext <ty> <value> to <ty2> ; yields ty2
6245 The '``zext``' instruction zero extends its operand to type ``ty2``.
6250 The '``zext``' instruction takes a value to cast, and a type to cast it
6251 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
6252 the same number of integers. The bit size of the ``value`` must be
6253 smaller than the bit size of the destination type, ``ty2``.
6258 The ``zext`` fills the high order bits of the ``value`` with zero bits
6259 until it reaches the size of the destination type, ``ty2``.
6261 When zero extending from i1, the result will always be either 0 or 1.
6266 .. code-block:: llvm
6268 %X = zext i32 257 to i64 ; yields i64:257
6269 %Y = zext i1 true to i32 ; yields i32:1
6270 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
6272 '``sext .. to``' Instruction
6273 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6280 <result> = sext <ty> <value> to <ty2> ; yields ty2
6285 The '``sext``' sign extends ``value`` to the type ``ty2``.
6290 The '``sext``' instruction takes a value to cast, and a type to cast it
6291 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
6292 the same number of integers. The bit size of the ``value`` must be
6293 smaller than the bit size of the destination type, ``ty2``.
6298 The '``sext``' instruction performs a sign extension by copying the sign
6299 bit (highest order bit) of the ``value`` until it reaches the bit size
6300 of the type ``ty2``.
6302 When sign extending from i1, the extension always results in -1 or 0.
6307 .. code-block:: llvm
6309 %X = sext i8 -1 to i16 ; yields i16 :65535
6310 %Y = sext i1 true to i32 ; yields i32:-1
6311 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
6313 '``fptrunc .. to``' Instruction
6314 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6321 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
6326 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
6331 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
6332 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
6333 The size of ``value`` must be larger than the size of ``ty2``. This
6334 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
6339 The '``fptrunc``' instruction truncates a ``value`` from a larger
6340 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
6341 point <t_floating>` type. If the value cannot fit within the
6342 destination type, ``ty2``, then the results are undefined.
6347 .. code-block:: llvm
6349 %X = fptrunc double 123.0 to float ; yields float:123.0
6350 %Y = fptrunc double 1.0E+300 to float ; yields undefined
6352 '``fpext .. to``' Instruction
6353 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6360 <result> = fpext <ty> <value> to <ty2> ; yields ty2
6365 The '``fpext``' extends a floating point ``value`` to a larger floating
6371 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
6372 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
6373 to. The source type must be smaller than the destination type.
6378 The '``fpext``' instruction extends the ``value`` from a smaller
6379 :ref:`floating point <t_floating>` type to a larger :ref:`floating
6380 point <t_floating>` type. The ``fpext`` cannot be used to make a
6381 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
6382 *no-op cast* for a floating point cast.
6387 .. code-block:: llvm
6389 %X = fpext float 3.125 to double ; yields double:3.125000e+00
6390 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
6392 '``fptoui .. to``' Instruction
6393 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6400 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
6405 The '``fptoui``' converts a floating point ``value`` to its unsigned
6406 integer equivalent of type ``ty2``.
6411 The '``fptoui``' instruction takes a value to cast, which must be a
6412 scalar or vector :ref:`floating point <t_floating>` value, and a type to
6413 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
6414 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
6415 type with the same number of elements as ``ty``
6420 The '``fptoui``' instruction converts its :ref:`floating
6421 point <t_floating>` operand into the nearest (rounding towards zero)
6422 unsigned integer value. If the value cannot fit in ``ty2``, the results
6428 .. code-block:: llvm
6430 %X = fptoui double 123.0 to i32 ; yields i32:123
6431 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
6432 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
6434 '``fptosi .. to``' Instruction
6435 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6442 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
6447 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
6448 ``value`` to type ``ty2``.
6453 The '``fptosi``' instruction takes a value to cast, which must be a
6454 scalar or vector :ref:`floating point <t_floating>` value, and a type to
6455 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
6456 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
6457 type with the same number of elements as ``ty``
6462 The '``fptosi``' instruction converts its :ref:`floating
6463 point <t_floating>` operand into the nearest (rounding towards zero)
6464 signed integer value. If the value cannot fit in ``ty2``, the results
6470 .. code-block:: llvm
6472 %X = fptosi double -123.0 to i32 ; yields i32:-123
6473 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
6474 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
6476 '``uitofp .. to``' Instruction
6477 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6484 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
6489 The '``uitofp``' instruction regards ``value`` as an unsigned integer
6490 and converts that value to the ``ty2`` type.
6495 The '``uitofp``' instruction takes a value to cast, which must be a
6496 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
6497 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
6498 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
6499 type with the same number of elements as ``ty``
6504 The '``uitofp``' instruction interprets its operand as an unsigned
6505 integer quantity and converts it to the corresponding floating point
6506 value. If the value cannot fit in the floating point value, the results
6512 .. code-block:: llvm
6514 %X = uitofp i32 257 to float ; yields float:257.0
6515 %Y = uitofp i8 -1 to double ; yields double:255.0
6517 '``sitofp .. to``' Instruction
6518 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6525 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
6530 The '``sitofp``' instruction regards ``value`` as a signed integer and
6531 converts that value to the ``ty2`` type.
6536 The '``sitofp``' instruction takes a value to cast, which must be a
6537 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
6538 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
6539 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
6540 type with the same number of elements as ``ty``
6545 The '``sitofp``' instruction interprets its operand as a signed integer
6546 quantity and converts it to the corresponding floating point value. If
6547 the value cannot fit in the floating point value, the results are
6553 .. code-block:: llvm
6555 %X = sitofp i32 257 to float ; yields float:257.0
6556 %Y = sitofp i8 -1 to double ; yields double:-1.0
6560 '``ptrtoint .. to``' Instruction
6561 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6568 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
6573 The '``ptrtoint``' instruction converts the pointer or a vector of
6574 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
6579 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
6580 a a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
6581 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
6582 a vector of integers type.
6587 The '``ptrtoint``' instruction converts ``value`` to integer type
6588 ``ty2`` by interpreting the pointer value as an integer and either
6589 truncating or zero extending that value to the size of the integer type.
6590 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
6591 ``value`` is larger than ``ty2`` then a truncation is done. If they are
6592 the same size, then nothing is done (*no-op cast*) other than a type
6598 .. code-block:: llvm
6600 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
6601 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
6602 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
6606 '``inttoptr .. to``' Instruction
6607 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6614 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
6619 The '``inttoptr``' instruction converts an integer ``value`` to a
6620 pointer type, ``ty2``.
6625 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
6626 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
6632 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
6633 applying either a zero extension or a truncation depending on the size
6634 of the integer ``value``. If ``value`` is larger than the size of a
6635 pointer then a truncation is done. If ``value`` is smaller than the size
6636 of a pointer then a zero extension is done. If they are the same size,
6637 nothing is done (*no-op cast*).
6642 .. code-block:: llvm
6644 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
6645 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
6646 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
6647 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
6651 '``bitcast .. to``' Instruction
6652 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6659 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
6664 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
6670 The '``bitcast``' instruction takes a value to cast, which must be a
6671 non-aggregate first class value, and a type to cast it to, which must
6672 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
6673 bit sizes of ``value`` and the destination type, ``ty2``, must be
6674 identical. If the source type is a pointer, the destination type must
6675 also be a pointer of the same size. This instruction supports bitwise
6676 conversion of vectors to integers and to vectors of other types (as
6677 long as they have the same size).
6682 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
6683 is always a *no-op cast* because no bits change with this
6684 conversion. The conversion is done as if the ``value`` had been stored
6685 to memory and read back as type ``ty2``. Pointer (or vector of
6686 pointers) types may only be converted to other pointer (or vector of
6687 pointers) types with the same address space through this instruction.
6688 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
6689 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
6694 .. code-block:: llvm
6696 %X = bitcast i8 255 to i8 ; yields i8 :-1
6697 %Y = bitcast i32* %x to sint* ; yields sint*:%x
6698 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
6699 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
6701 .. _i_addrspacecast:
6703 '``addrspacecast .. to``' Instruction
6704 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6711 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
6716 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
6717 address space ``n`` to type ``pty2`` in address space ``m``.
6722 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
6723 to cast and a pointer type to cast it to, which must have a different
6729 The '``addrspacecast``' instruction converts the pointer value
6730 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
6731 value modification, depending on the target and the address space
6732 pair. Pointer conversions within the same address space must be
6733 performed with the ``bitcast`` instruction. Note that if the address space
6734 conversion is legal then both result and operand refer to the same memory
6740 .. code-block:: llvm
6742 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
6743 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
6744 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
6751 The instructions in this category are the "miscellaneous" instructions,
6752 which defy better classification.
6756 '``icmp``' Instruction
6757 ^^^^^^^^^^^^^^^^^^^^^^
6764 <result> = icmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6769 The '``icmp``' instruction returns a boolean value or a vector of
6770 boolean values based on comparison of its two integer, integer vector,
6771 pointer, or pointer vector operands.
6776 The '``icmp``' instruction takes three operands. The first operand is
6777 the condition code indicating the kind of comparison to perform. It is
6778 not a value, just a keyword. The possible condition code are:
6781 #. ``ne``: not equal
6782 #. ``ugt``: unsigned greater than
6783 #. ``uge``: unsigned greater or equal
6784 #. ``ult``: unsigned less than
6785 #. ``ule``: unsigned less or equal
6786 #. ``sgt``: signed greater than
6787 #. ``sge``: signed greater or equal
6788 #. ``slt``: signed less than
6789 #. ``sle``: signed less or equal
6791 The remaining two arguments must be :ref:`integer <t_integer>` or
6792 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
6793 must also be identical types.
6798 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
6799 code given as ``cond``. The comparison performed always yields either an
6800 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
6802 #. ``eq``: yields ``true`` if the operands are equal, ``false``
6803 otherwise. No sign interpretation is necessary or performed.
6804 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
6805 otherwise. No sign interpretation is necessary or performed.
6806 #. ``ugt``: interprets the operands as unsigned values and yields
6807 ``true`` if ``op1`` is greater than ``op2``.
6808 #. ``uge``: interprets the operands as unsigned values and yields
6809 ``true`` if ``op1`` is greater than or equal to ``op2``.
6810 #. ``ult``: interprets the operands as unsigned values and yields
6811 ``true`` if ``op1`` is less than ``op2``.
6812 #. ``ule``: interprets the operands as unsigned values and yields
6813 ``true`` if ``op1`` is less than or equal to ``op2``.
6814 #. ``sgt``: interprets the operands as signed values and yields ``true``
6815 if ``op1`` is greater than ``op2``.
6816 #. ``sge``: interprets the operands as signed values and yields ``true``
6817 if ``op1`` is greater than or equal to ``op2``.
6818 #. ``slt``: interprets the operands as signed values and yields ``true``
6819 if ``op1`` is less than ``op2``.
6820 #. ``sle``: interprets the operands as signed values and yields ``true``
6821 if ``op1`` is less than or equal to ``op2``.
6823 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
6824 are compared as if they were integers.
6826 If the operands are integer vectors, then they are compared element by
6827 element. The result is an ``i1`` vector with the same number of elements
6828 as the values being compared. Otherwise, the result is an ``i1``.
6833 .. code-block:: llvm
6835 <result> = icmp eq i32 4, 5 ; yields: result=false
6836 <result> = icmp ne float* %X, %X ; yields: result=false
6837 <result> = icmp ult i16 4, 5 ; yields: result=true
6838 <result> = icmp sgt i16 4, 5 ; yields: result=false
6839 <result> = icmp ule i16 -4, 5 ; yields: result=false
6840 <result> = icmp sge i16 4, 5 ; yields: result=false
6842 Note that the code generator does not yet support vector types with the
6843 ``icmp`` instruction.
6847 '``fcmp``' Instruction
6848 ^^^^^^^^^^^^^^^^^^^^^^
6855 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6860 The '``fcmp``' instruction returns a boolean value or vector of boolean
6861 values based on comparison of its operands.
6863 If the operands are floating point scalars, then the result type is a
6864 boolean (:ref:`i1 <t_integer>`).
6866 If the operands are floating point vectors, then the result type is a
6867 vector of boolean with the same number of elements as the operands being
6873 The '``fcmp``' instruction takes three operands. The first operand is
6874 the condition code indicating the kind of comparison to perform. It is
6875 not a value, just a keyword. The possible condition code are:
6877 #. ``false``: no comparison, always returns false
6878 #. ``oeq``: ordered and equal
6879 #. ``ogt``: ordered and greater than
6880 #. ``oge``: ordered and greater than or equal
6881 #. ``olt``: ordered and less than
6882 #. ``ole``: ordered and less than or equal
6883 #. ``one``: ordered and not equal
6884 #. ``ord``: ordered (no nans)
6885 #. ``ueq``: unordered or equal
6886 #. ``ugt``: unordered or greater than
6887 #. ``uge``: unordered or greater than or equal
6888 #. ``ult``: unordered or less than
6889 #. ``ule``: unordered or less than or equal
6890 #. ``une``: unordered or not equal
6891 #. ``uno``: unordered (either nans)
6892 #. ``true``: no comparison, always returns true
6894 *Ordered* means that neither operand is a QNAN while *unordered* means
6895 that either operand may be a QNAN.
6897 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
6898 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
6899 type. They must have identical types.
6904 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
6905 condition code given as ``cond``. If the operands are vectors, then the
6906 vectors are compared element by element. Each comparison performed
6907 always yields an :ref:`i1 <t_integer>` result, as follows:
6909 #. ``false``: always yields ``false``, regardless of operands.
6910 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
6911 is equal to ``op2``.
6912 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
6913 is greater than ``op2``.
6914 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
6915 is greater than or equal to ``op2``.
6916 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
6917 is less than ``op2``.
6918 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
6919 is less than or equal to ``op2``.
6920 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
6921 is not equal to ``op2``.
6922 #. ``ord``: yields ``true`` if both operands are not a QNAN.
6923 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
6925 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
6926 greater than ``op2``.
6927 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
6928 greater than or equal to ``op2``.
6929 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
6931 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
6932 less than or equal to ``op2``.
6933 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
6934 not equal to ``op2``.
6935 #. ``uno``: yields ``true`` if either operand is a QNAN.
6936 #. ``true``: always yields ``true``, regardless of operands.
6941 .. code-block:: llvm
6943 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
6944 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
6945 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
6946 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
6948 Note that the code generator does not yet support vector types with the
6949 ``fcmp`` instruction.
6953 '``phi``' Instruction
6954 ^^^^^^^^^^^^^^^^^^^^^
6961 <result> = phi <ty> [ <val0>, <label0>], ...
6966 The '``phi``' instruction is used to implement the φ node in the SSA
6967 graph representing the function.
6972 The type of the incoming values is specified with the first type field.
6973 After this, the '``phi``' instruction takes a list of pairs as
6974 arguments, with one pair for each predecessor basic block of the current
6975 block. Only values of :ref:`first class <t_firstclass>` type may be used as
6976 the value arguments to the PHI node. Only labels may be used as the
6979 There must be no non-phi instructions between the start of a basic block
6980 and the PHI instructions: i.e. PHI instructions must be first in a basic
6983 For the purposes of the SSA form, the use of each incoming value is
6984 deemed to occur on the edge from the corresponding predecessor block to
6985 the current block (but after any definition of an '``invoke``'
6986 instruction's return value on the same edge).
6991 At runtime, the '``phi``' instruction logically takes on the value
6992 specified by the pair corresponding to the predecessor basic block that
6993 executed just prior to the current block.
6998 .. code-block:: llvm
7000 Loop: ; Infinite loop that counts from 0 on up...
7001 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
7002 %nextindvar = add i32 %indvar, 1
7007 '``select``' Instruction
7008 ^^^^^^^^^^^^^^^^^^^^^^^^
7015 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
7017 selty is either i1 or {<N x i1>}
7022 The '``select``' instruction is used to choose one value based on a
7023 condition, without IR-level branching.
7028 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
7029 values indicating the condition, and two values of the same :ref:`first
7030 class <t_firstclass>` type.
7035 If the condition is an i1 and it evaluates to 1, the instruction returns
7036 the first value argument; otherwise, it returns the second value
7039 If the condition is a vector of i1, then the value arguments must be
7040 vectors of the same size, and the selection is done element by element.
7042 If the condition is an i1 and the value arguments are vectors of the
7043 same size, then an entire vector is selected.
7048 .. code-block:: llvm
7050 %X = select i1 true, i8 17, i8 42 ; yields i8:17
7054 '``call``' Instruction
7055 ^^^^^^^^^^^^^^^^^^^^^^
7062 <result> = [tail | musttail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
7067 The '``call``' instruction represents a simple function call.
7072 This instruction requires several arguments:
7074 #. The optional ``tail`` and ``musttail`` markers indicate that the optimizers
7075 should perform tail call optimization. The ``tail`` marker is a hint that
7076 `can be ignored <CodeGenerator.html#sibcallopt>`_. The ``musttail`` marker
7077 means that the call must be tail call optimized in order for the program to
7078 be correct. The ``musttail`` marker provides these guarantees:
7080 #. The call will not cause unbounded stack growth if it is part of a
7081 recursive cycle in the call graph.
7082 #. Arguments with the :ref:`inalloca <attr_inalloca>` attribute are
7085 Both markers imply that the callee does not access allocas or varargs from
7086 the caller. Calls marked ``musttail`` must obey the following additional
7089 - The call must immediately precede a :ref:`ret <i_ret>` instruction,
7090 or a pointer bitcast followed by a ret instruction.
7091 - The ret instruction must return the (possibly bitcasted) value
7092 produced by the call or void.
7093 - The caller and callee prototypes must match. Pointer types of
7094 parameters or return types may differ in pointee type, but not
7096 - The calling conventions of the caller and callee must match.
7097 - All ABI-impacting function attributes, such as sret, byval, inreg,
7098 returned, and inalloca, must match.
7099 - The callee must be varargs iff the caller is varargs. Bitcasting a
7100 non-varargs function to the appropriate varargs type is legal so
7101 long as the non-varargs prefixes obey the other rules.
7103 Tail call optimization for calls marked ``tail`` is guaranteed to occur if
7104 the following conditions are met:
7106 - Caller and callee both have the calling convention ``fastcc``.
7107 - The call is in tail position (ret immediately follows call and ret
7108 uses value of call or is void).
7109 - Option ``-tailcallopt`` is enabled, or
7110 ``llvm::GuaranteedTailCallOpt`` is ``true``.
7111 - `Platform-specific constraints are
7112 met. <CodeGenerator.html#tailcallopt>`_
7114 #. The optional "cconv" marker indicates which :ref:`calling
7115 convention <callingconv>` the call should use. If none is
7116 specified, the call defaults to using C calling conventions. The
7117 calling convention of the call must match the calling convention of
7118 the target function, or else the behavior is undefined.
7119 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
7120 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
7122 #. '``ty``': the type of the call instruction itself which is also the
7123 type of the return value. Functions that return no value are marked
7125 #. '``fnty``': shall be the signature of the pointer to function value
7126 being invoked. The argument types must match the types implied by
7127 this signature. This type can be omitted if the function is not
7128 varargs and if the function type does not return a pointer to a
7130 #. '``fnptrval``': An LLVM value containing a pointer to a function to
7131 be invoked. In most cases, this is a direct function invocation, but
7132 indirect ``call``'s are just as possible, calling an arbitrary pointer
7134 #. '``function args``': argument list whose types match the function
7135 signature argument types and parameter attributes. All arguments must
7136 be of :ref:`first class <t_firstclass>` type. If the function signature
7137 indicates the function accepts a variable number of arguments, the
7138 extra arguments can be specified.
7139 #. The optional :ref:`function attributes <fnattrs>` list. Only
7140 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
7141 attributes are valid here.
7146 The '``call``' instruction is used to cause control flow to transfer to
7147 a specified function, with its incoming arguments bound to the specified
7148 values. Upon a '``ret``' instruction in the called function, control
7149 flow continues with the instruction after the function call, and the
7150 return value of the function is bound to the result argument.
7155 .. code-block:: llvm
7157 %retval = call i32 @test(i32 %argc)
7158 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
7159 %X = tail call i32 @foo() ; yields i32
7160 %Y = tail call fastcc i32 @foo() ; yields i32
7161 call void %foo(i8 97 signext)
7163 %struct.A = type { i32, i8 }
7164 %r = call %struct.A @foo() ; yields { i32, i8 }
7165 %gr = extractvalue %struct.A %r, 0 ; yields i32
7166 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
7167 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
7168 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
7170 llvm treats calls to some functions with names and arguments that match
7171 the standard C99 library as being the C99 library functions, and may
7172 perform optimizations or generate code for them under that assumption.
7173 This is something we'd like to change in the future to provide better
7174 support for freestanding environments and non-C-based languages.
7178 '``va_arg``' Instruction
7179 ^^^^^^^^^^^^^^^^^^^^^^^^
7186 <resultval> = va_arg <va_list*> <arglist>, <argty>
7191 The '``va_arg``' instruction is used to access arguments passed through
7192 the "variable argument" area of a function call. It is used to implement
7193 the ``va_arg`` macro in C.
7198 This instruction takes a ``va_list*`` value and the type of the
7199 argument. It returns a value of the specified argument type and
7200 increments the ``va_list`` to point to the next argument. The actual
7201 type of ``va_list`` is target specific.
7206 The '``va_arg``' instruction loads an argument of the specified type
7207 from the specified ``va_list`` and causes the ``va_list`` to point to
7208 the next argument. For more information, see the variable argument
7209 handling :ref:`Intrinsic Functions <int_varargs>`.
7211 It is legal for this instruction to be called in a function which does
7212 not take a variable number of arguments, for example, the ``vfprintf``
7215 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
7216 function <intrinsics>` because it takes a type as an argument.
7221 See the :ref:`variable argument processing <int_varargs>` section.
7223 Note that the code generator does not yet fully support va\_arg on many
7224 targets. Also, it does not currently support va\_arg with aggregate
7225 types on any target.
7229 '``landingpad``' Instruction
7230 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7237 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
7238 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
7240 <clause> := catch <type> <value>
7241 <clause> := filter <array constant type> <array constant>
7246 The '``landingpad``' instruction is used by `LLVM's exception handling
7247 system <ExceptionHandling.html#overview>`_ to specify that a basic block
7248 is a landing pad --- one where the exception lands, and corresponds to the
7249 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
7250 defines values supplied by the personality function (``pers_fn``) upon
7251 re-entry to the function. The ``resultval`` has the type ``resultty``.
7256 This instruction takes a ``pers_fn`` value. This is the personality
7257 function associated with the unwinding mechanism. The optional
7258 ``cleanup`` flag indicates that the landing pad block is a cleanup.
7260 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
7261 contains the global variable representing the "type" that may be caught
7262 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
7263 clause takes an array constant as its argument. Use
7264 "``[0 x i8**] undef``" for a filter which cannot throw. The
7265 '``landingpad``' instruction must contain *at least* one ``clause`` or
7266 the ``cleanup`` flag.
7271 The '``landingpad``' instruction defines the values which are set by the
7272 personality function (``pers_fn``) upon re-entry to the function, and
7273 therefore the "result type" of the ``landingpad`` instruction. As with
7274 calling conventions, how the personality function results are
7275 represented in LLVM IR is target specific.
7277 The clauses are applied in order from top to bottom. If two
7278 ``landingpad`` instructions are merged together through inlining, the
7279 clauses from the calling function are appended to the list of clauses.
7280 When the call stack is being unwound due to an exception being thrown,
7281 the exception is compared against each ``clause`` in turn. If it doesn't
7282 match any of the clauses, and the ``cleanup`` flag is not set, then
7283 unwinding continues further up the call stack.
7285 The ``landingpad`` instruction has several restrictions:
7287 - A landing pad block is a basic block which is the unwind destination
7288 of an '``invoke``' instruction.
7289 - A landing pad block must have a '``landingpad``' instruction as its
7290 first non-PHI instruction.
7291 - There can be only one '``landingpad``' instruction within the landing
7293 - A basic block that is not a landing pad block may not include a
7294 '``landingpad``' instruction.
7295 - All '``landingpad``' instructions in a function must have the same
7296 personality function.
7301 .. code-block:: llvm
7303 ;; A landing pad which can catch an integer.
7304 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
7306 ;; A landing pad that is a cleanup.
7307 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
7309 ;; A landing pad which can catch an integer and can only throw a double.
7310 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
7312 filter [1 x i8**] [@_ZTId]
7319 LLVM supports the notion of an "intrinsic function". These functions
7320 have well known names and semantics and are required to follow certain
7321 restrictions. Overall, these intrinsics represent an extension mechanism
7322 for the LLVM language that does not require changing all of the
7323 transformations in LLVM when adding to the language (or the bitcode
7324 reader/writer, the parser, etc...).
7326 Intrinsic function names must all start with an "``llvm.``" prefix. This
7327 prefix is reserved in LLVM for intrinsic names; thus, function names may
7328 not begin with this prefix. Intrinsic functions must always be external
7329 functions: you cannot define the body of intrinsic functions. Intrinsic
7330 functions may only be used in call or invoke instructions: it is illegal
7331 to take the address of an intrinsic function. Additionally, because
7332 intrinsic functions are part of the LLVM language, it is required if any
7333 are added that they be documented here.
7335 Some intrinsic functions can be overloaded, i.e., the intrinsic
7336 represents a family of functions that perform the same operation but on
7337 different data types. Because LLVM can represent over 8 million
7338 different integer types, overloading is used commonly to allow an
7339 intrinsic function to operate on any integer type. One or more of the
7340 argument types or the result type can be overloaded to accept any
7341 integer type. Argument types may also be defined as exactly matching a
7342 previous argument's type or the result type. This allows an intrinsic
7343 function which accepts multiple arguments, but needs all of them to be
7344 of the same type, to only be overloaded with respect to a single
7345 argument or the result.
7347 Overloaded intrinsics will have the names of its overloaded argument
7348 types encoded into its function name, each preceded by a period. Only
7349 those types which are overloaded result in a name suffix. Arguments
7350 whose type is matched against another type do not. For example, the
7351 ``llvm.ctpop`` function can take an integer of any width and returns an
7352 integer of exactly the same integer width. This leads to a family of
7353 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
7354 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
7355 overloaded, and only one type suffix is required. Because the argument's
7356 type is matched against the return type, it does not require its own
7359 To learn how to add an intrinsic function, please see the `Extending
7360 LLVM Guide <ExtendingLLVM.html>`_.
7364 Variable Argument Handling Intrinsics
7365 -------------------------------------
7367 Variable argument support is defined in LLVM with the
7368 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
7369 functions. These functions are related to the similarly named macros
7370 defined in the ``<stdarg.h>`` header file.
7372 All of these functions operate on arguments that use a target-specific
7373 value type "``va_list``". The LLVM assembly language reference manual
7374 does not define what this type is, so all transformations should be
7375 prepared to handle these functions regardless of the type used.
7377 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
7378 variable argument handling intrinsic functions are used.
7380 .. code-block:: llvm
7382 ; This struct is different for every platform. For most platforms,
7383 ; it is merely an i8*.
7384 %struct.va_list = type { i8* }
7386 ; For Unix x86_64 platforms, va_list is the following struct:
7387 ; %struct.va_list = type { i32, i32, i8*, i8* }
7389 define i32 @test(i32 %X, ...) {
7390 ; Initialize variable argument processing
7391 %ap = alloca %struct.va_list
7392 %ap2 = bitcast %struct.va_list* %ap to i8*
7393 call void @llvm.va_start(i8* %ap2)
7395 ; Read a single integer argument
7396 %tmp = va_arg i8* %ap2, i32
7398 ; Demonstrate usage of llvm.va_copy and llvm.va_end
7400 %aq2 = bitcast i8** %aq to i8*
7401 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
7402 call void @llvm.va_end(i8* %aq2)
7404 ; Stop processing of arguments.
7405 call void @llvm.va_end(i8* %ap2)
7409 declare void @llvm.va_start(i8*)
7410 declare void @llvm.va_copy(i8*, i8*)
7411 declare void @llvm.va_end(i8*)
7415 '``llvm.va_start``' Intrinsic
7416 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7423 declare void @llvm.va_start(i8* <arglist>)
7428 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
7429 subsequent use by ``va_arg``.
7434 The argument is a pointer to a ``va_list`` element to initialize.
7439 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
7440 available in C. In a target-dependent way, it initializes the
7441 ``va_list`` element to which the argument points, so that the next call
7442 to ``va_arg`` will produce the first variable argument passed to the
7443 function. Unlike the C ``va_start`` macro, this intrinsic does not need
7444 to know the last argument of the function as the compiler can figure
7447 '``llvm.va_end``' Intrinsic
7448 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7455 declare void @llvm.va_end(i8* <arglist>)
7460 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
7461 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
7466 The argument is a pointer to a ``va_list`` to destroy.
7471 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
7472 available in C. In a target-dependent way, it destroys the ``va_list``
7473 element to which the argument points. Calls to
7474 :ref:`llvm.va_start <int_va_start>` and
7475 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
7480 '``llvm.va_copy``' Intrinsic
7481 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7488 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
7493 The '``llvm.va_copy``' intrinsic copies the current argument position
7494 from the source argument list to the destination argument list.
7499 The first argument is a pointer to a ``va_list`` element to initialize.
7500 The second argument is a pointer to a ``va_list`` element to copy from.
7505 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
7506 available in C. In a target-dependent way, it copies the source
7507 ``va_list`` element into the destination ``va_list`` element. This
7508 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
7509 arbitrarily complex and require, for example, memory allocation.
7511 Accurate Garbage Collection Intrinsics
7512 --------------------------------------
7514 LLVM's support for `Accurate Garbage Collection <GarbageCollection.html>`_
7515 (GC) requires the frontend to generate code containing appropriate intrinsic
7516 calls and select an appropriate GC strategy which knows how to lower these
7517 intrinsics in a manner which is appropriate for the target collector.
7519 These intrinsics allow identification of :ref:`GC roots on the
7520 stack <int_gcroot>`, as well as garbage collector implementations that
7521 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
7522 Frontends for type-safe garbage collected languages should generate
7523 these intrinsics to make use of the LLVM garbage collectors. For more
7524 details, see `Garbage Collection with LLVM <GarbageCollection.html>`_.
7526 Experimental Statepoint Intrinsics
7527 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7529 LLVM provides an second experimental set of intrinsics for describing garbage
7530 collection safepoints in compiled code. These intrinsics are an alternative
7531 to the ``llvm.gcroot`` intrinsics, but are compatible with the ones for
7532 :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers. The
7533 differences in approach are covered in the `Garbage Collection with LLVM
7534 <GarbageCollection.html>`_ documentation. The intrinsics themselves are
7535 described in :doc:`Statepoints`.
7539 '``llvm.gcroot``' Intrinsic
7540 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7547 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
7552 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
7553 the code generator, and allows some metadata to be associated with it.
7558 The first argument specifies the address of a stack object that contains
7559 the root pointer. The second pointer (which must be either a constant or
7560 a global value address) contains the meta-data to be associated with the
7566 At runtime, a call to this intrinsic stores a null pointer into the
7567 "ptrloc" location. At compile-time, the code generator generates
7568 information to allow the runtime to find the pointer at GC safe points.
7569 The '``llvm.gcroot``' intrinsic may only be used in a function which
7570 :ref:`specifies a GC algorithm <gc>`.
7574 '``llvm.gcread``' Intrinsic
7575 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7582 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
7587 The '``llvm.gcread``' intrinsic identifies reads of references from heap
7588 locations, allowing garbage collector implementations that require read
7594 The second argument is the address to read from, which should be an
7595 address allocated from the garbage collector. The first object is a
7596 pointer to the start of the referenced object, if needed by the language
7597 runtime (otherwise null).
7602 The '``llvm.gcread``' intrinsic has the same semantics as a load
7603 instruction, but may be replaced with substantially more complex code by
7604 the garbage collector runtime, as needed. The '``llvm.gcread``'
7605 intrinsic may only be used in a function which :ref:`specifies a GC
7610 '``llvm.gcwrite``' Intrinsic
7611 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7618 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
7623 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
7624 locations, allowing garbage collector implementations that require write
7625 barriers (such as generational or reference counting collectors).
7630 The first argument is the reference to store, the second is the start of
7631 the object to store it to, and the third is the address of the field of
7632 Obj to store to. If the runtime does not require a pointer to the
7633 object, Obj may be null.
7638 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
7639 instruction, but may be replaced with substantially more complex code by
7640 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
7641 intrinsic may only be used in a function which :ref:`specifies a GC
7644 Code Generator Intrinsics
7645 -------------------------
7647 These intrinsics are provided by LLVM to expose special features that
7648 may only be implemented with code generator support.
7650 '``llvm.returnaddress``' Intrinsic
7651 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7658 declare i8 *@llvm.returnaddress(i32 <level>)
7663 The '``llvm.returnaddress``' intrinsic attempts to compute a
7664 target-specific value indicating the return address of the current
7665 function or one of its callers.
7670 The argument to this intrinsic indicates which function to return the
7671 address for. Zero indicates the calling function, one indicates its
7672 caller, etc. The argument is **required** to be a constant integer
7678 The '``llvm.returnaddress``' intrinsic either returns a pointer
7679 indicating the return address of the specified call frame, or zero if it
7680 cannot be identified. The value returned by this intrinsic is likely to
7681 be incorrect or 0 for arguments other than zero, so it should only be
7682 used for debugging purposes.
7684 Note that calling this intrinsic does not prevent function inlining or
7685 other aggressive transformations, so the value returned may not be that
7686 of the obvious source-language caller.
7688 '``llvm.frameaddress``' Intrinsic
7689 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7696 declare i8* @llvm.frameaddress(i32 <level>)
7701 The '``llvm.frameaddress``' intrinsic attempts to return the
7702 target-specific frame pointer value for the specified stack frame.
7707 The argument to this intrinsic indicates which function to return the
7708 frame pointer for. Zero indicates the calling function, one indicates
7709 its caller, etc. The argument is **required** to be a constant integer
7715 The '``llvm.frameaddress``' intrinsic either returns a pointer
7716 indicating the frame address of the specified call frame, or zero if it
7717 cannot be identified. The value returned by this intrinsic is likely to
7718 be incorrect or 0 for arguments other than zero, so it should only be
7719 used for debugging purposes.
7721 Note that calling this intrinsic does not prevent function inlining or
7722 other aggressive transformations, so the value returned may not be that
7723 of the obvious source-language caller.
7725 '``llvm.frameescape``' and '``llvm.framerecover``' Intrinsics
7726 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7733 declare void @llvm.frameescape(...)
7734 declare i8* @llvm.framerecover(i8* %func, i8* %fp, i32 %idx)
7739 The '``llvm.frameescape``' intrinsic escapes offsets of a collection of static
7740 allocas, and the '``llvm.framerecover``' intrinsic applies those offsets to a
7741 live frame pointer to recover the address of the allocation. The offset is
7742 computed during frame layout of the caller of ``llvm.frameescape``.
7747 All arguments to '``llvm.frameescape``' must be pointers to static allocas or
7748 casts of static allocas. Each function can only call '``llvm.frameescape``'
7749 once, and it can only do so from the entry block.
7751 The ``func`` argument to '``llvm.framerecover``' must be a constant
7752 bitcasted pointer to a function defined in the current module. The code
7753 generator cannot determine the frame allocation offset of functions defined in
7756 The ``fp`` argument to '``llvm.framerecover``' must be a frame
7757 pointer of a call frame that is currently live. The return value of
7758 '``llvm.frameaddress``' is one way to produce such a value, but most platforms
7759 also expose the frame pointer through stack unwinding mechanisms.
7761 The ``idx`` argument to '``llvm.framerecover``' indicates which alloca passed to
7762 '``llvm.frameescape``' to recover. It is zero-indexed.
7767 These intrinsics allow a group of functions to access one stack memory
7768 allocation in an ancestor stack frame. The memory returned from
7769 '``llvm.frameallocate``' may be allocated prior to stack realignment, so the
7770 memory is only aligned to the ABI-required stack alignment. Each function may
7771 only call '``llvm.frameallocate``' one or zero times from the function entry
7772 block. The frame allocation intrinsic inhibits inlining, as any frame
7773 allocations in the inlined function frame are likely to be at a different
7774 offset from the one used by '``llvm.framerecover``' called with the
7777 .. _int_read_register:
7778 .. _int_write_register:
7780 '``llvm.read_register``' and '``llvm.write_register``' Intrinsics
7781 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7788 declare i32 @llvm.read_register.i32(metadata)
7789 declare i64 @llvm.read_register.i64(metadata)
7790 declare void @llvm.write_register.i32(metadata, i32 @value)
7791 declare void @llvm.write_register.i64(metadata, i64 @value)
7797 The '``llvm.read_register``' and '``llvm.write_register``' intrinsics
7798 provides access to the named register. The register must be valid on
7799 the architecture being compiled to. The type needs to be compatible
7800 with the register being read.
7805 The '``llvm.read_register``' intrinsic returns the current value of the
7806 register, where possible. The '``llvm.write_register``' intrinsic sets
7807 the current value of the register, where possible.
7809 This is useful to implement named register global variables that need
7810 to always be mapped to a specific register, as is common practice on
7811 bare-metal programs including OS kernels.
7813 The compiler doesn't check for register availability or use of the used
7814 register in surrounding code, including inline assembly. Because of that,
7815 allocatable registers are not supported.
7817 Warning: So far it only works with the stack pointer on selected
7818 architectures (ARM, AArch64, PowerPC and x86_64). Significant amount of
7819 work is needed to support other registers and even more so, allocatable
7824 '``llvm.stacksave``' Intrinsic
7825 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7832 declare i8* @llvm.stacksave()
7837 The '``llvm.stacksave``' intrinsic is used to remember the current state
7838 of the function stack, for use with
7839 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
7840 implementing language features like scoped automatic variable sized
7846 This intrinsic returns a opaque pointer value that can be passed to
7847 :ref:`llvm.stackrestore <int_stackrestore>`. When an
7848 ``llvm.stackrestore`` intrinsic is executed with a value saved from
7849 ``llvm.stacksave``, it effectively restores the state of the stack to
7850 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
7851 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
7852 were allocated after the ``llvm.stacksave`` was executed.
7854 .. _int_stackrestore:
7856 '``llvm.stackrestore``' Intrinsic
7857 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7864 declare void @llvm.stackrestore(i8* %ptr)
7869 The '``llvm.stackrestore``' intrinsic is used to restore the state of
7870 the function stack to the state it was in when the corresponding
7871 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
7872 useful for implementing language features like scoped automatic variable
7873 sized arrays in C99.
7878 See the description for :ref:`llvm.stacksave <int_stacksave>`.
7880 '``llvm.prefetch``' Intrinsic
7881 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7888 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
7893 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
7894 insert a prefetch instruction if supported; otherwise, it is a noop.
7895 Prefetches have no effect on the behavior of the program but can change
7896 its performance characteristics.
7901 ``address`` is the address to be prefetched, ``rw`` is the specifier
7902 determining if the fetch should be for a read (0) or write (1), and
7903 ``locality`` is a temporal locality specifier ranging from (0) - no
7904 locality, to (3) - extremely local keep in cache. The ``cache type``
7905 specifies whether the prefetch is performed on the data (1) or
7906 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
7907 arguments must be constant integers.
7912 This intrinsic does not modify the behavior of the program. In
7913 particular, prefetches cannot trap and do not produce a value. On
7914 targets that support this intrinsic, the prefetch can provide hints to
7915 the processor cache for better performance.
7917 '``llvm.pcmarker``' Intrinsic
7918 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7925 declare void @llvm.pcmarker(i32 <id>)
7930 The '``llvm.pcmarker``' intrinsic is a method to export a Program
7931 Counter (PC) in a region of code to simulators and other tools. The
7932 method is target specific, but it is expected that the marker will use
7933 exported symbols to transmit the PC of the marker. The marker makes no
7934 guarantees that it will remain with any specific instruction after
7935 optimizations. It is possible that the presence of a marker will inhibit
7936 optimizations. The intended use is to be inserted after optimizations to
7937 allow correlations of simulation runs.
7942 ``id`` is a numerical id identifying the marker.
7947 This intrinsic does not modify the behavior of the program. Backends
7948 that do not support this intrinsic may ignore it.
7950 '``llvm.readcyclecounter``' Intrinsic
7951 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7958 declare i64 @llvm.readcyclecounter()
7963 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
7964 counter register (or similar low latency, high accuracy clocks) on those
7965 targets that support it. On X86, it should map to RDTSC. On Alpha, it
7966 should map to RPCC. As the backing counters overflow quickly (on the
7967 order of 9 seconds on alpha), this should only be used for small
7973 When directly supported, reading the cycle counter should not modify any
7974 memory. Implementations are allowed to either return a application
7975 specific value or a system wide value. On backends without support, this
7976 is lowered to a constant 0.
7978 Note that runtime support may be conditional on the privilege-level code is
7979 running at and the host platform.
7981 '``llvm.clear_cache``' Intrinsic
7982 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7989 declare void @llvm.clear_cache(i8*, i8*)
7994 The '``llvm.clear_cache``' intrinsic ensures visibility of modifications
7995 in the specified range to the execution unit of the processor. On
7996 targets with non-unified instruction and data cache, the implementation
7997 flushes the instruction cache.
8002 On platforms with coherent instruction and data caches (e.g. x86), this
8003 intrinsic is a nop. On platforms with non-coherent instruction and data
8004 cache (e.g. ARM, MIPS), the intrinsic is lowered either to appropriate
8005 instructions or a system call, if cache flushing requires special
8008 The default behavior is to emit a call to ``__clear_cache`` from the run
8011 This instrinsic does *not* empty the instruction pipeline. Modifications
8012 of the current function are outside the scope of the intrinsic.
8014 '``llvm.instrprof_increment``' Intrinsic
8015 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8022 declare void @llvm.instrprof_increment(i8* <name>, i64 <hash>,
8023 i32 <num-counters>, i32 <index>)
8028 The '``llvm.instrprof_increment``' intrinsic can be emitted by a
8029 frontend for use with instrumentation based profiling. These will be
8030 lowered by the ``-instrprof`` pass to generate execution counts of a
8036 The first argument is a pointer to a global variable containing the
8037 name of the entity being instrumented. This should generally be the
8038 (mangled) function name for a set of counters.
8040 The second argument is a hash value that can be used by the consumer
8041 of the profile data to detect changes to the instrumented source, and
8042 the third is the number of counters associated with ``name``. It is an
8043 error if ``hash`` or ``num-counters`` differ between two instances of
8044 ``instrprof_increment`` that refer to the same name.
8046 The last argument refers to which of the counters for ``name`` should
8047 be incremented. It should be a value between 0 and ``num-counters``.
8052 This intrinsic represents an increment of a profiling counter. It will
8053 cause the ``-instrprof`` pass to generate the appropriate data
8054 structures and the code to increment the appropriate value, in a
8055 format that can be written out by a compiler runtime and consumed via
8056 the ``llvm-profdata`` tool.
8058 Standard C Library Intrinsics
8059 -----------------------------
8061 LLVM provides intrinsics for a few important standard C library
8062 functions. These intrinsics allow source-language front-ends to pass
8063 information about the alignment of the pointer arguments to the code
8064 generator, providing opportunity for more efficient code generation.
8068 '``llvm.memcpy``' Intrinsic
8069 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8074 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
8075 integer bit width and for different address spaces. Not all targets
8076 support all bit widths however.
8080 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
8081 i32 <len>, i32 <align>, i1 <isvolatile>)
8082 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
8083 i64 <len>, i32 <align>, i1 <isvolatile>)
8088 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
8089 source location to the destination location.
8091 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
8092 intrinsics do not return a value, takes extra alignment/isvolatile
8093 arguments and the pointers can be in specified address spaces.
8098 The first argument is a pointer to the destination, the second is a
8099 pointer to the source. The third argument is an integer argument
8100 specifying the number of bytes to copy, the fourth argument is the
8101 alignment of the source and destination locations, and the fifth is a
8102 boolean indicating a volatile access.
8104 If the call to this intrinsic has an alignment value that is not 0 or 1,
8105 then the caller guarantees that both the source and destination pointers
8106 are aligned to that boundary.
8108 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
8109 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
8110 very cleanly specified and it is unwise to depend on it.
8115 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
8116 source location to the destination location, which are not allowed to
8117 overlap. It copies "len" bytes of memory over. If the argument is known
8118 to be aligned to some boundary, this can be specified as the fourth
8119 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
8121 '``llvm.memmove``' Intrinsic
8122 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8127 This is an overloaded intrinsic. You can use llvm.memmove on any integer
8128 bit width and for different address space. Not all targets support all
8133 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
8134 i32 <len>, i32 <align>, i1 <isvolatile>)
8135 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
8136 i64 <len>, i32 <align>, i1 <isvolatile>)
8141 The '``llvm.memmove.*``' intrinsics move a block of memory from the
8142 source location to the destination location. It is similar to the
8143 '``llvm.memcpy``' intrinsic but allows the two memory locations to
8146 Note that, unlike the standard libc function, the ``llvm.memmove.*``
8147 intrinsics do not return a value, takes extra alignment/isvolatile
8148 arguments and the pointers can be in specified address spaces.
8153 The first argument is a pointer to the destination, the second is a
8154 pointer to the source. The third argument is an integer argument
8155 specifying the number of bytes to copy, the fourth argument is the
8156 alignment of the source and destination locations, and the fifth is a
8157 boolean indicating a volatile access.
8159 If the call to this intrinsic has an alignment value that is not 0 or 1,
8160 then the caller guarantees that the source and destination pointers are
8161 aligned to that boundary.
8163 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
8164 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
8165 not very cleanly specified and it is unwise to depend on it.
8170 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
8171 source location to the destination location, which may overlap. It
8172 copies "len" bytes of memory over. If the argument is known to be
8173 aligned to some boundary, this can be specified as the fourth argument,
8174 otherwise it should be set to 0 or 1 (both meaning no alignment).
8176 '``llvm.memset.*``' Intrinsics
8177 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8182 This is an overloaded intrinsic. You can use llvm.memset on any integer
8183 bit width and for different address spaces. However, not all targets
8184 support all bit widths.
8188 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
8189 i32 <len>, i32 <align>, i1 <isvolatile>)
8190 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
8191 i64 <len>, i32 <align>, i1 <isvolatile>)
8196 The '``llvm.memset.*``' intrinsics fill a block of memory with a
8197 particular byte value.
8199 Note that, unlike the standard libc function, the ``llvm.memset``
8200 intrinsic does not return a value and takes extra alignment/volatile
8201 arguments. Also, the destination can be in an arbitrary address space.
8206 The first argument is a pointer to the destination to fill, the second
8207 is the byte value with which to fill it, the third argument is an
8208 integer argument specifying the number of bytes to fill, and the fourth
8209 argument is the known alignment of the destination location.
8211 If the call to this intrinsic has an alignment value that is not 0 or 1,
8212 then the caller guarantees that the destination pointer is aligned to
8215 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
8216 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
8217 very cleanly specified and it is unwise to depend on it.
8222 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
8223 at the destination location. If the argument is known to be aligned to
8224 some boundary, this can be specified as the fourth argument, otherwise
8225 it should be set to 0 or 1 (both meaning no alignment).
8227 '``llvm.sqrt.*``' Intrinsic
8228 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8233 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
8234 floating point or vector of floating point type. Not all targets support
8239 declare float @llvm.sqrt.f32(float %Val)
8240 declare double @llvm.sqrt.f64(double %Val)
8241 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
8242 declare fp128 @llvm.sqrt.f128(fp128 %Val)
8243 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
8248 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
8249 returning the same value as the libm '``sqrt``' functions would. Unlike
8250 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
8251 negative numbers other than -0.0 (which allows for better optimization,
8252 because there is no need to worry about errno being set).
8253 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
8258 The argument and return value are floating point numbers of the same
8264 This function returns the sqrt of the specified operand if it is a
8265 nonnegative floating point number.
8267 '``llvm.powi.*``' Intrinsic
8268 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8273 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
8274 floating point or vector of floating point type. Not all targets support
8279 declare float @llvm.powi.f32(float %Val, i32 %power)
8280 declare double @llvm.powi.f64(double %Val, i32 %power)
8281 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
8282 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
8283 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
8288 The '``llvm.powi.*``' intrinsics return the first operand raised to the
8289 specified (positive or negative) power. The order of evaluation of
8290 multiplications is not defined. When a vector of floating point type is
8291 used, the second argument remains a scalar integer value.
8296 The second argument is an integer power, and the first is a value to
8297 raise to that power.
8302 This function returns the first value raised to the second power with an
8303 unspecified sequence of rounding operations.
8305 '``llvm.sin.*``' Intrinsic
8306 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8311 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
8312 floating point or vector of floating point type. Not all targets support
8317 declare float @llvm.sin.f32(float %Val)
8318 declare double @llvm.sin.f64(double %Val)
8319 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
8320 declare fp128 @llvm.sin.f128(fp128 %Val)
8321 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
8326 The '``llvm.sin.*``' intrinsics return the sine of the operand.
8331 The argument and return value are floating point numbers of the same
8337 This function returns the sine of the specified operand, returning the
8338 same values as the libm ``sin`` functions would, and handles error
8339 conditions in the same way.
8341 '``llvm.cos.*``' Intrinsic
8342 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8347 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
8348 floating point or vector of floating point type. Not all targets support
8353 declare float @llvm.cos.f32(float %Val)
8354 declare double @llvm.cos.f64(double %Val)
8355 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
8356 declare fp128 @llvm.cos.f128(fp128 %Val)
8357 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
8362 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
8367 The argument and return value are floating point numbers of the same
8373 This function returns the cosine of the specified operand, returning the
8374 same values as the libm ``cos`` functions would, and handles error
8375 conditions in the same way.
8377 '``llvm.pow.*``' Intrinsic
8378 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8383 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
8384 floating point or vector of floating point type. Not all targets support
8389 declare float @llvm.pow.f32(float %Val, float %Power)
8390 declare double @llvm.pow.f64(double %Val, double %Power)
8391 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
8392 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
8393 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
8398 The '``llvm.pow.*``' intrinsics return the first operand raised to the
8399 specified (positive or negative) power.
8404 The second argument is a floating point power, and the first is a value
8405 to raise to that power.
8410 This function returns the first value raised to the second power,
8411 returning the same values as the libm ``pow`` functions would, and
8412 handles error conditions in the same way.
8414 '``llvm.exp.*``' Intrinsic
8415 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8420 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
8421 floating point or vector of floating point type. Not all targets support
8426 declare float @llvm.exp.f32(float %Val)
8427 declare double @llvm.exp.f64(double %Val)
8428 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
8429 declare fp128 @llvm.exp.f128(fp128 %Val)
8430 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
8435 The '``llvm.exp.*``' intrinsics perform the exp function.
8440 The argument and return value are floating point numbers of the same
8446 This function returns the same values as the libm ``exp`` functions
8447 would, and handles error conditions in the same way.
8449 '``llvm.exp2.*``' Intrinsic
8450 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8455 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
8456 floating point or vector of floating point type. Not all targets support
8461 declare float @llvm.exp2.f32(float %Val)
8462 declare double @llvm.exp2.f64(double %Val)
8463 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
8464 declare fp128 @llvm.exp2.f128(fp128 %Val)
8465 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
8470 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
8475 The argument and return value are floating point numbers of the same
8481 This function returns the same values as the libm ``exp2`` functions
8482 would, and handles error conditions in the same way.
8484 '``llvm.log.*``' Intrinsic
8485 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8490 This is an overloaded intrinsic. You can use ``llvm.log`` on any
8491 floating point or vector of floating point type. Not all targets support
8496 declare float @llvm.log.f32(float %Val)
8497 declare double @llvm.log.f64(double %Val)
8498 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
8499 declare fp128 @llvm.log.f128(fp128 %Val)
8500 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
8505 The '``llvm.log.*``' intrinsics perform the log function.
8510 The argument and return value are floating point numbers of the same
8516 This function returns the same values as the libm ``log`` functions
8517 would, and handles error conditions in the same way.
8519 '``llvm.log10.*``' Intrinsic
8520 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8525 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
8526 floating point or vector of floating point type. Not all targets support
8531 declare float @llvm.log10.f32(float %Val)
8532 declare double @llvm.log10.f64(double %Val)
8533 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
8534 declare fp128 @llvm.log10.f128(fp128 %Val)
8535 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
8540 The '``llvm.log10.*``' intrinsics perform the log10 function.
8545 The argument and return value are floating point numbers of the same
8551 This function returns the same values as the libm ``log10`` functions
8552 would, and handles error conditions in the same way.
8554 '``llvm.log2.*``' Intrinsic
8555 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8560 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
8561 floating point or vector of floating point type. Not all targets support
8566 declare float @llvm.log2.f32(float %Val)
8567 declare double @llvm.log2.f64(double %Val)
8568 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
8569 declare fp128 @llvm.log2.f128(fp128 %Val)
8570 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
8575 The '``llvm.log2.*``' intrinsics perform the log2 function.
8580 The argument and return value are floating point numbers of the same
8586 This function returns the same values as the libm ``log2`` functions
8587 would, and handles error conditions in the same way.
8589 '``llvm.fma.*``' Intrinsic
8590 ^^^^^^^^^^^^^^^^^^^^^^^^^^
8595 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
8596 floating point or vector of floating point type. Not all targets support
8601 declare float @llvm.fma.f32(float %a, float %b, float %c)
8602 declare double @llvm.fma.f64(double %a, double %b, double %c)
8603 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
8604 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
8605 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
8610 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
8616 The argument and return value are floating point numbers of the same
8622 This function returns the same values as the libm ``fma`` functions
8623 would, and does not set errno.
8625 '``llvm.fabs.*``' Intrinsic
8626 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8631 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
8632 floating point or vector of floating point type. Not all targets support
8637 declare float @llvm.fabs.f32(float %Val)
8638 declare double @llvm.fabs.f64(double %Val)
8639 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
8640 declare fp128 @llvm.fabs.f128(fp128 %Val)
8641 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
8646 The '``llvm.fabs.*``' intrinsics return the absolute value of the
8652 The argument and return value are floating point numbers of the same
8658 This function returns the same values as the libm ``fabs`` functions
8659 would, and handles error conditions in the same way.
8661 '``llvm.minnum.*``' Intrinsic
8662 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8667 This is an overloaded intrinsic. You can use ``llvm.minnum`` on any
8668 floating point or vector of floating point type. Not all targets support
8673 declare float @llvm.minnum.f32(float %Val0, float %Val1)
8674 declare double @llvm.minnum.f64(double %Val0, double %Val1)
8675 declare x86_fp80 @llvm.minnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
8676 declare fp128 @llvm.minnum.f128(fp128 %Val0, fp128 %Val1)
8677 declare ppc_fp128 @llvm.minnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
8682 The '``llvm.minnum.*``' intrinsics return the minimum of the two
8689 The arguments and return value are floating point numbers of the same
8695 Follows the IEEE-754 semantics for minNum, which also match for libm's
8698 If either operand is a NaN, returns the other non-NaN operand. Returns
8699 NaN only if both operands are NaN. If the operands compare equal,
8700 returns a value that compares equal to both operands. This means that
8701 fmin(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
8703 '``llvm.maxnum.*``' Intrinsic
8704 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8709 This is an overloaded intrinsic. You can use ``llvm.maxnum`` on any
8710 floating point or vector of floating point type. Not all targets support
8715 declare float @llvm.maxnum.f32(float %Val0, float %Val1l)
8716 declare double @llvm.maxnum.f64(double %Val0, double %Val1)
8717 declare x86_fp80 @llvm.maxnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
8718 declare fp128 @llvm.maxnum.f128(fp128 %Val0, fp128 %Val1)
8719 declare ppc_fp128 @llvm.maxnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
8724 The '``llvm.maxnum.*``' intrinsics return the maximum of the two
8731 The arguments and return value are floating point numbers of the same
8736 Follows the IEEE-754 semantics for maxNum, which also match for libm's
8739 If either operand is a NaN, returns the other non-NaN operand. Returns
8740 NaN only if both operands are NaN. If the operands compare equal,
8741 returns a value that compares equal to both operands. This means that
8742 fmax(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
8744 '``llvm.copysign.*``' Intrinsic
8745 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8750 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
8751 floating point or vector of floating point type. Not all targets support
8756 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
8757 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
8758 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
8759 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
8760 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
8765 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
8766 first operand and the sign of the second operand.
8771 The arguments and return value are floating point numbers of the same
8777 This function returns the same values as the libm ``copysign``
8778 functions would, and handles error conditions in the same way.
8780 '``llvm.floor.*``' Intrinsic
8781 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8786 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
8787 floating point or vector of floating point type. Not all targets support
8792 declare float @llvm.floor.f32(float %Val)
8793 declare double @llvm.floor.f64(double %Val)
8794 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
8795 declare fp128 @llvm.floor.f128(fp128 %Val)
8796 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
8801 The '``llvm.floor.*``' intrinsics return the floor of the operand.
8806 The argument and return value are floating point numbers of the same
8812 This function returns the same values as the libm ``floor`` functions
8813 would, and handles error conditions in the same way.
8815 '``llvm.ceil.*``' Intrinsic
8816 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8821 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
8822 floating point or vector of floating point type. Not all targets support
8827 declare float @llvm.ceil.f32(float %Val)
8828 declare double @llvm.ceil.f64(double %Val)
8829 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
8830 declare fp128 @llvm.ceil.f128(fp128 %Val)
8831 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
8836 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
8841 The argument and return value are floating point numbers of the same
8847 This function returns the same values as the libm ``ceil`` functions
8848 would, and handles error conditions in the same way.
8850 '``llvm.trunc.*``' Intrinsic
8851 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8856 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
8857 floating point or vector of floating point type. Not all targets support
8862 declare float @llvm.trunc.f32(float %Val)
8863 declare double @llvm.trunc.f64(double %Val)
8864 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
8865 declare fp128 @llvm.trunc.f128(fp128 %Val)
8866 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
8871 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
8872 nearest integer not larger in magnitude than the operand.
8877 The argument and return value are floating point numbers of the same
8883 This function returns the same values as the libm ``trunc`` functions
8884 would, and handles error conditions in the same way.
8886 '``llvm.rint.*``' Intrinsic
8887 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8892 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
8893 floating point or vector of floating point type. Not all targets support
8898 declare float @llvm.rint.f32(float %Val)
8899 declare double @llvm.rint.f64(double %Val)
8900 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
8901 declare fp128 @llvm.rint.f128(fp128 %Val)
8902 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
8907 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
8908 nearest integer. It may raise an inexact floating-point exception if the
8909 operand isn't an integer.
8914 The argument and return value are floating point numbers of the same
8920 This function returns the same values as the libm ``rint`` functions
8921 would, and handles error conditions in the same way.
8923 '``llvm.nearbyint.*``' Intrinsic
8924 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8929 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
8930 floating point or vector of floating point type. Not all targets support
8935 declare float @llvm.nearbyint.f32(float %Val)
8936 declare double @llvm.nearbyint.f64(double %Val)
8937 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
8938 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
8939 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
8944 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
8950 The argument and return value are floating point numbers of the same
8956 This function returns the same values as the libm ``nearbyint``
8957 functions would, and handles error conditions in the same way.
8959 '``llvm.round.*``' Intrinsic
8960 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8965 This is an overloaded intrinsic. You can use ``llvm.round`` on any
8966 floating point or vector of floating point type. Not all targets support
8971 declare float @llvm.round.f32(float %Val)
8972 declare double @llvm.round.f64(double %Val)
8973 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
8974 declare fp128 @llvm.round.f128(fp128 %Val)
8975 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
8980 The '``llvm.round.*``' intrinsics returns the operand rounded to the
8986 The argument and return value are floating point numbers of the same
8992 This function returns the same values as the libm ``round``
8993 functions would, and handles error conditions in the same way.
8995 Bit Manipulation Intrinsics
8996 ---------------------------
8998 LLVM provides intrinsics for a few important bit manipulation
8999 operations. These allow efficient code generation for some algorithms.
9001 '``llvm.bswap.*``' Intrinsics
9002 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9007 This is an overloaded intrinsic function. You can use bswap on any
9008 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
9012 declare i16 @llvm.bswap.i16(i16 <id>)
9013 declare i32 @llvm.bswap.i32(i32 <id>)
9014 declare i64 @llvm.bswap.i64(i64 <id>)
9019 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
9020 values with an even number of bytes (positive multiple of 16 bits).
9021 These are useful for performing operations on data that is not in the
9022 target's native byte order.
9027 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
9028 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
9029 intrinsic returns an i32 value that has the four bytes of the input i32
9030 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
9031 returned i32 will have its bytes in 3, 2, 1, 0 order. The
9032 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
9033 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
9036 '``llvm.ctpop.*``' Intrinsic
9037 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9042 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
9043 bit width, or on any vector with integer elements. Not all targets
9044 support all bit widths or vector types, however.
9048 declare i8 @llvm.ctpop.i8(i8 <src>)
9049 declare i16 @llvm.ctpop.i16(i16 <src>)
9050 declare i32 @llvm.ctpop.i32(i32 <src>)
9051 declare i64 @llvm.ctpop.i64(i64 <src>)
9052 declare i256 @llvm.ctpop.i256(i256 <src>)
9053 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
9058 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
9064 The only argument is the value to be counted. The argument may be of any
9065 integer type, or a vector with integer elements. The return type must
9066 match the argument type.
9071 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
9072 each element of a vector.
9074 '``llvm.ctlz.*``' Intrinsic
9075 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9080 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
9081 integer bit width, or any vector whose elements are integers. Not all
9082 targets support all bit widths or vector types, however.
9086 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
9087 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
9088 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
9089 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
9090 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
9091 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
9096 The '``llvm.ctlz``' family of intrinsic functions counts the number of
9097 leading zeros in a variable.
9102 The first argument is the value to be counted. This argument may be of
9103 any integer type, or a vector with integer element type. The return
9104 type must match the first argument type.
9106 The second argument must be a constant and is a flag to indicate whether
9107 the intrinsic should ensure that a zero as the first argument produces a
9108 defined result. Historically some architectures did not provide a
9109 defined result for zero values as efficiently, and many algorithms are
9110 now predicated on avoiding zero-value inputs.
9115 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
9116 zeros in a variable, or within each element of the vector. If
9117 ``src == 0`` then the result is the size in bits of the type of ``src``
9118 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
9119 ``llvm.ctlz(i32 2) = 30``.
9121 '``llvm.cttz.*``' Intrinsic
9122 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9127 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
9128 integer bit width, or any vector of integer elements. Not all targets
9129 support all bit widths or vector types, however.
9133 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
9134 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
9135 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
9136 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
9137 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
9138 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
9143 The '``llvm.cttz``' family of intrinsic functions counts the number of
9149 The first argument is the value to be counted. This argument may be of
9150 any integer type, or a vector with integer element type. The return
9151 type must match the first argument type.
9153 The second argument must be a constant and is a flag to indicate whether
9154 the intrinsic should ensure that a zero as the first argument produces a
9155 defined result. Historically some architectures did not provide a
9156 defined result for zero values as efficiently, and many algorithms are
9157 now predicated on avoiding zero-value inputs.
9162 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
9163 zeros in a variable, or within each element of a vector. If ``src == 0``
9164 then the result is the size in bits of the type of ``src`` if
9165 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
9166 ``llvm.cttz(2) = 1``.
9170 Arithmetic with Overflow Intrinsics
9171 -----------------------------------
9173 LLVM provides intrinsics for some arithmetic with overflow operations.
9175 '``llvm.sadd.with.overflow.*``' Intrinsics
9176 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9181 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
9182 on any integer bit width.
9186 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
9187 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
9188 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
9193 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
9194 a signed addition of the two arguments, and indicate whether an overflow
9195 occurred during the signed summation.
9200 The arguments (%a and %b) and the first element of the result structure
9201 may be of integer types of any bit width, but they must have the same
9202 bit width. The second element of the result structure must be of type
9203 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
9209 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
9210 a signed addition of the two variables. They return a structure --- the
9211 first element of which is the signed summation, and the second element
9212 of which is a bit specifying if the signed summation resulted in an
9218 .. code-block:: llvm
9220 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
9221 %sum = extractvalue {i32, i1} %res, 0
9222 %obit = extractvalue {i32, i1} %res, 1
9223 br i1 %obit, label %overflow, label %normal
9225 '``llvm.uadd.with.overflow.*``' Intrinsics
9226 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9231 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
9232 on any integer bit width.
9236 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
9237 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
9238 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
9243 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
9244 an unsigned addition of the two arguments, and indicate whether a carry
9245 occurred during the unsigned summation.
9250 The arguments (%a and %b) and the first element of the result structure
9251 may be of integer types of any bit width, but they must have the same
9252 bit width. The second element of the result structure must be of type
9253 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
9259 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
9260 an unsigned addition of the two arguments. They return a structure --- the
9261 first element of which is the sum, and the second element of which is a
9262 bit specifying if the unsigned summation resulted in a carry.
9267 .. code-block:: llvm
9269 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
9270 %sum = extractvalue {i32, i1} %res, 0
9271 %obit = extractvalue {i32, i1} %res, 1
9272 br i1 %obit, label %carry, label %normal
9274 '``llvm.ssub.with.overflow.*``' Intrinsics
9275 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9280 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
9281 on any integer bit width.
9285 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
9286 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
9287 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
9292 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
9293 a signed subtraction of the two arguments, and indicate whether an
9294 overflow occurred during the signed subtraction.
9299 The arguments (%a and %b) and the first element of the result structure
9300 may be of integer types of any bit width, but they must have the same
9301 bit width. The second element of the result structure must be of type
9302 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
9308 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
9309 a signed subtraction of the two arguments. They return a structure --- the
9310 first element of which is the subtraction, and the second element of
9311 which is a bit specifying if the signed subtraction resulted in an
9317 .. code-block:: llvm
9319 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
9320 %sum = extractvalue {i32, i1} %res, 0
9321 %obit = extractvalue {i32, i1} %res, 1
9322 br i1 %obit, label %overflow, label %normal
9324 '``llvm.usub.with.overflow.*``' Intrinsics
9325 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9330 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
9331 on any integer bit width.
9335 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
9336 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
9337 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
9342 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
9343 an unsigned subtraction of the two arguments, and indicate whether an
9344 overflow occurred during the unsigned subtraction.
9349 The arguments (%a and %b) and the first element of the result structure
9350 may be of integer types of any bit width, but they must have the same
9351 bit width. The second element of the result structure must be of type
9352 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
9358 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
9359 an unsigned subtraction of the two arguments. They return a structure ---
9360 the first element of which is the subtraction, and the second element of
9361 which is a bit specifying if the unsigned subtraction resulted in an
9367 .. code-block:: llvm
9369 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
9370 %sum = extractvalue {i32, i1} %res, 0
9371 %obit = extractvalue {i32, i1} %res, 1
9372 br i1 %obit, label %overflow, label %normal
9374 '``llvm.smul.with.overflow.*``' Intrinsics
9375 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9380 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
9381 on any integer bit width.
9385 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
9386 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
9387 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
9392 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
9393 a signed multiplication of the two arguments, and indicate whether an
9394 overflow occurred during the signed multiplication.
9399 The arguments (%a and %b) and the first element of the result structure
9400 may be of integer types of any bit width, but they must have the same
9401 bit width. The second element of the result structure must be of type
9402 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
9408 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
9409 a signed multiplication of the two arguments. They return a structure ---
9410 the first element of which is the multiplication, and the second element
9411 of which is a bit specifying if the signed multiplication resulted in an
9417 .. code-block:: llvm
9419 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
9420 %sum = extractvalue {i32, i1} %res, 0
9421 %obit = extractvalue {i32, i1} %res, 1
9422 br i1 %obit, label %overflow, label %normal
9424 '``llvm.umul.with.overflow.*``' Intrinsics
9425 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9430 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
9431 on any integer bit width.
9435 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
9436 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
9437 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
9442 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
9443 a unsigned multiplication of the two arguments, and indicate whether an
9444 overflow occurred during the unsigned multiplication.
9449 The arguments (%a and %b) and the first element of the result structure
9450 may be of integer types of any bit width, but they must have the same
9451 bit width. The second element of the result structure must be of type
9452 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
9458 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
9459 an unsigned multiplication of the two arguments. They return a structure ---
9460 the first element of which is the multiplication, and the second
9461 element of which is a bit specifying if the unsigned multiplication
9462 resulted in an overflow.
9467 .. code-block:: llvm
9469 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
9470 %sum = extractvalue {i32, i1} %res, 0
9471 %obit = extractvalue {i32, i1} %res, 1
9472 br i1 %obit, label %overflow, label %normal
9474 Specialised Arithmetic Intrinsics
9475 ---------------------------------
9477 '``llvm.fmuladd.*``' Intrinsic
9478 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9485 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
9486 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
9491 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
9492 expressions that can be fused if the code generator determines that (a) the
9493 target instruction set has support for a fused operation, and (b) that the
9494 fused operation is more efficient than the equivalent, separate pair of mul
9495 and add instructions.
9500 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
9501 multiplicands, a and b, and an addend c.
9510 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
9512 is equivalent to the expression a \* b + c, except that rounding will
9513 not be performed between the multiplication and addition steps if the
9514 code generator fuses the operations. Fusion is not guaranteed, even if
9515 the target platform supports it. If a fused multiply-add is required the
9516 corresponding llvm.fma.\* intrinsic function should be used
9517 instead. This never sets errno, just as '``llvm.fma.*``'.
9522 .. code-block:: llvm
9524 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields float:r2 = (a * b) + c
9526 Half Precision Floating Point Intrinsics
9527 ----------------------------------------
9529 For most target platforms, half precision floating point is a
9530 storage-only format. This means that it is a dense encoding (in memory)
9531 but does not support computation in the format.
9533 This means that code must first load the half-precision floating point
9534 value as an i16, then convert it to float with
9535 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
9536 then be performed on the float value (including extending to double
9537 etc). To store the value back to memory, it is first converted to float
9538 if needed, then converted to i16 with
9539 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
9542 .. _int_convert_to_fp16:
9544 '``llvm.convert.to.fp16``' Intrinsic
9545 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9552 declare i16 @llvm.convert.to.fp16.f32(float %a)
9553 declare i16 @llvm.convert.to.fp16.f64(double %a)
9558 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
9559 conventional floating point type to half precision floating point format.
9564 The intrinsic function contains single argument - the value to be
9570 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
9571 conventional floating point format to half precision floating point format. The
9572 return value is an ``i16`` which contains the converted number.
9577 .. code-block:: llvm
9579 %res = call i16 @llvm.convert.to.fp16.f32(float %a)
9580 store i16 %res, i16* @x, align 2
9582 .. _int_convert_from_fp16:
9584 '``llvm.convert.from.fp16``' Intrinsic
9585 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9592 declare float @llvm.convert.from.fp16.f32(i16 %a)
9593 declare double @llvm.convert.from.fp16.f64(i16 %a)
9598 The '``llvm.convert.from.fp16``' intrinsic function performs a
9599 conversion from half precision floating point format to single precision
9600 floating point format.
9605 The intrinsic function contains single argument - the value to be
9611 The '``llvm.convert.from.fp16``' intrinsic function performs a
9612 conversion from half single precision floating point format to single
9613 precision floating point format. The input half-float value is
9614 represented by an ``i16`` value.
9619 .. code-block:: llvm
9621 %a = load i16, i16* @x, align 2
9622 %res = call float @llvm.convert.from.fp16(i16 %a)
9629 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
9630 prefix), are described in the `LLVM Source Level
9631 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
9634 Exception Handling Intrinsics
9635 -----------------------------
9637 The LLVM exception handling intrinsics (which all start with
9638 ``llvm.eh.`` prefix), are described in the `LLVM Exception
9639 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
9643 Trampoline Intrinsics
9644 ---------------------
9646 These intrinsics make it possible to excise one parameter, marked with
9647 the :ref:`nest <nest>` attribute, from a function. The result is a
9648 callable function pointer lacking the nest parameter - the caller does
9649 not need to provide a value for it. Instead, the value to use is stored
9650 in advance in a "trampoline", a block of memory usually allocated on the
9651 stack, which also contains code to splice the nest value into the
9652 argument list. This is used to implement the GCC nested function address
9655 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
9656 then the resulting function pointer has signature ``i32 (i32, i32)*``.
9657 It can be created as follows:
9659 .. code-block:: llvm
9661 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
9662 %tramp1 = getelementptr [10 x i8], [10 x i8]* %tramp, i32 0, i32 0
9663 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
9664 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
9665 %fp = bitcast i8* %p to i32 (i32, i32)*
9667 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
9668 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
9672 '``llvm.init.trampoline``' Intrinsic
9673 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9680 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
9685 This fills the memory pointed to by ``tramp`` with executable code,
9686 turning it into a trampoline.
9691 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
9692 pointers. The ``tramp`` argument must point to a sufficiently large and
9693 sufficiently aligned block of memory; this memory is written to by the
9694 intrinsic. Note that the size and the alignment are target-specific -
9695 LLVM currently provides no portable way of determining them, so a
9696 front-end that generates this intrinsic needs to have some
9697 target-specific knowledge. The ``func`` argument must hold a function
9698 bitcast to an ``i8*``.
9703 The block of memory pointed to by ``tramp`` is filled with target
9704 dependent code, turning it into a function. Then ``tramp`` needs to be
9705 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
9706 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
9707 function's signature is the same as that of ``func`` with any arguments
9708 marked with the ``nest`` attribute removed. At most one such ``nest``
9709 argument is allowed, and it must be of pointer type. Calling the new
9710 function is equivalent to calling ``func`` with the same argument list,
9711 but with ``nval`` used for the missing ``nest`` argument. If, after
9712 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
9713 modified, then the effect of any later call to the returned function
9714 pointer is undefined.
9718 '``llvm.adjust.trampoline``' Intrinsic
9719 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9726 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
9731 This performs any required machine-specific adjustment to the address of
9732 a trampoline (passed as ``tramp``).
9737 ``tramp`` must point to a block of memory which already has trampoline
9738 code filled in by a previous call to
9739 :ref:`llvm.init.trampoline <int_it>`.
9744 On some architectures the address of the code to be executed needs to be
9745 different than the address where the trampoline is actually stored. This
9746 intrinsic returns the executable address corresponding to ``tramp``
9747 after performing the required machine specific adjustments. The pointer
9748 returned can then be :ref:`bitcast and executed <int_trampoline>`.
9750 Masked Vector Load and Store Intrinsics
9751 ---------------------------------------
9753 LLVM provides intrinsics for predicated vector load and store operations. The predicate is specified by a mask operand, which holds one bit per vector element, switching the associated vector lane on or off. The memory addresses corresponding to the "off" lanes are not accessed. When all bits of the mask are on, the intrinsic is identical to a regular vector load or store. When all bits are off, no memory is accessed.
9757 '``llvm.masked.load.*``' Intrinsics
9758 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9762 This is an overloaded intrinsic. The loaded data is a vector of any integer or floating point data type.
9766 declare <16 x float> @llvm.masked.load.v16f32 (<16 x float>* <ptr>, i32 <alignment>, <16 x i1> <mask>, <16 x float> <passthru>)
9767 declare <2 x double> @llvm.masked.load.v2f64 (<2 x double>* <ptr>, i32 <alignment>, <2 x i1> <mask>, <2 x double> <passthru>)
9772 Reads a vector from memory according to the provided mask. The mask holds a bit for each vector lane, and is used to prevent memory accesses to the masked-off lanes. The masked-off lanes in the result vector are taken from the corresponding lanes in the passthru operand.
9778 The first operand is the base pointer for the load. The second operand is the alignment of the source location. It must be a constant integer value. The third operand, mask, is a vector of boolean 'i1' values with the same number of elements as the return type. The fourth is a pass-through value that is used to fill the masked-off lanes of the result. The return type, underlying type of the base pointer and the type of passthru operand are the same vector types.
9784 The '``llvm.masked.load``' intrinsic is designed for conditional reading of selected vector elements in a single IR operation. It is useful for targets that support vector masked loads and allows vectorizing predicated basic blocks on these targets. Other targets may support this intrinsic differently, for example by lowering it into a sequence of branches that guard scalar load operations.
9785 The result of this operation is equivalent to a regular vector load instruction followed by a 'select' between the loaded and the passthru values, predicated on the same mask. However, using this intrinsic prevents exceptions on memory access to masked-off lanes.
9790 %res = call <16 x float> @llvm.masked.load.v16f32 (<16 x float>* %ptr, i32 4, <16 x i1>%mask, <16 x float> %passthru)
9792 ;; The result of the two following instructions is identical aside from potential memory access exception
9793 %loadlal = load <16 x float>, <16 x float>* %ptr, align 4
9794 %res = select <16 x i1> %mask, <16 x float> %loadlal, <16 x float> %passthru
9798 '``llvm.masked.store.*``' Intrinsics
9799 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9803 This is an overloaded intrinsic. The data stored in memory is a vector of any integer or floating point data type.
9807 declare void @llvm.masked.store.v8i32 (<8 x i32> <value>, <8 x i32> * <ptr>, i32 <alignment>, <8 x i1> <mask>)
9808 declare void @llvm.masked.store.v16f32(<16 x i32> <value>, <16 x i32>* <ptr>, i32 <alignment>, <16 x i1> <mask>)
9813 Writes a vector to memory according to the provided mask. The mask holds a bit for each vector lane, and is used to prevent memory accesses to the masked-off lanes.
9818 The first operand is the vector value to be written to memory. The second operand is the base pointer for the store, it has the same underlying type as the value operand. The third operand is the alignment of the destination location. The fourth operand, mask, is a vector of boolean values. The types of the mask and the value operand must have the same number of vector elements.
9824 The '``llvm.masked.store``' intrinsics is designed for conditional writing of selected vector elements in a single IR operation. It is useful for targets that support vector masked store and allows vectorizing predicated basic blocks on these targets. Other targets may support this intrinsic differently, for example by lowering it into a sequence of branches that guard scalar store operations.
9825 The result of this operation is equivalent to a load-modify-store sequence. However, using this intrinsic prevents exceptions and data races on memory access to masked-off lanes.
9829 call void @llvm.masked.store.v16f32(<16 x float> %value, <16 x float>* %ptr, i32 4, <16 x i1> %mask)
9831 ;; The result of the following instructions is identical aside from potential data races and memory access exceptions
9832 %oldval = load <16 x float>, <16 x float>* %ptr, align 4
9833 %res = select <16 x i1> %mask, <16 x float> %value, <16 x float> %oldval
9834 store <16 x float> %res, <16 x float>* %ptr, align 4
9840 This class of intrinsics provides information about the lifetime of
9841 memory objects and ranges where variables are immutable.
9845 '``llvm.lifetime.start``' Intrinsic
9846 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9853 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
9858 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
9864 The first argument is a constant integer representing the size of the
9865 object, or -1 if it is variable sized. The second argument is a pointer
9871 This intrinsic indicates that before this point in the code, the value
9872 of the memory pointed to by ``ptr`` is dead. This means that it is known
9873 to never be used and has an undefined value. A load from the pointer
9874 that precedes this intrinsic can be replaced with ``'undef'``.
9878 '``llvm.lifetime.end``' Intrinsic
9879 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9886 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
9891 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
9897 The first argument is a constant integer representing the size of the
9898 object, or -1 if it is variable sized. The second argument is a pointer
9904 This intrinsic indicates that after this point in the code, the value of
9905 the memory pointed to by ``ptr`` is dead. This means that it is known to
9906 never be used and has an undefined value. Any stores into the memory
9907 object following this intrinsic may be removed as dead.
9909 '``llvm.invariant.start``' Intrinsic
9910 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9917 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
9922 The '``llvm.invariant.start``' intrinsic specifies that the contents of
9923 a memory object will not change.
9928 The first argument is a constant integer representing the size of the
9929 object, or -1 if it is variable sized. The second argument is a pointer
9935 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
9936 the return value, the referenced memory location is constant and
9939 '``llvm.invariant.end``' Intrinsic
9940 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9947 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
9952 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
9953 memory object are mutable.
9958 The first argument is the matching ``llvm.invariant.start`` intrinsic.
9959 The second argument is a constant integer representing the size of the
9960 object, or -1 if it is variable sized and the third argument is a
9961 pointer to the object.
9966 This intrinsic indicates that the memory is mutable again.
9971 This class of intrinsics is designed to be generic and has no specific
9974 '``llvm.var.annotation``' Intrinsic
9975 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9982 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9987 The '``llvm.var.annotation``' intrinsic.
9992 The first argument is a pointer to a value, the second is a pointer to a
9993 global string, the third is a pointer to a global string which is the
9994 source file name, and the last argument is the line number.
9999 This intrinsic allows annotation of local variables with arbitrary
10000 strings. This can be useful for special purpose optimizations that want
10001 to look for these annotations. These have no other defined use; they are
10002 ignored by code generation and optimization.
10004 '``llvm.ptr.annotation.*``' Intrinsic
10005 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10010 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
10011 pointer to an integer of any width. *NOTE* you must specify an address space for
10012 the pointer. The identifier for the default address space is the integer
10017 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
10018 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
10019 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
10020 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
10021 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
10026 The '``llvm.ptr.annotation``' intrinsic.
10031 The first argument is a pointer to an integer value of arbitrary bitwidth
10032 (result of some expression), the second is a pointer to a global string, the
10033 third is a pointer to a global string which is the source file name, and the
10034 last argument is the line number. It returns the value of the first argument.
10039 This intrinsic allows annotation of a pointer to an integer with arbitrary
10040 strings. This can be useful for special purpose optimizations that want to look
10041 for these annotations. These have no other defined use; they are ignored by code
10042 generation and optimization.
10044 '``llvm.annotation.*``' Intrinsic
10045 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10050 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
10051 any integer bit width.
10055 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
10056 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
10057 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
10058 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
10059 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
10064 The '``llvm.annotation``' intrinsic.
10069 The first argument is an integer value (result of some expression), the
10070 second is a pointer to a global string, the third is a pointer to a
10071 global string which is the source file name, and the last argument is
10072 the line number. It returns the value of the first argument.
10077 This intrinsic allows annotations to be put on arbitrary expressions
10078 with arbitrary strings. This can be useful for special purpose
10079 optimizations that want to look for these annotations. These have no
10080 other defined use; they are ignored by code generation and optimization.
10082 '``llvm.trap``' Intrinsic
10083 ^^^^^^^^^^^^^^^^^^^^^^^^^
10090 declare void @llvm.trap() noreturn nounwind
10095 The '``llvm.trap``' intrinsic.
10105 This intrinsic is lowered to the target dependent trap instruction. If
10106 the target does not have a trap instruction, this intrinsic will be
10107 lowered to a call of the ``abort()`` function.
10109 '``llvm.debugtrap``' Intrinsic
10110 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10117 declare void @llvm.debugtrap() nounwind
10122 The '``llvm.debugtrap``' intrinsic.
10132 This intrinsic is lowered to code which is intended to cause an
10133 execution trap with the intention of requesting the attention of a
10136 '``llvm.stackprotector``' Intrinsic
10137 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10144 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
10149 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
10150 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
10151 is placed on the stack before local variables.
10156 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
10157 The first argument is the value loaded from the stack guard
10158 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
10159 enough space to hold the value of the guard.
10164 This intrinsic causes the prologue/epilogue inserter to force the position of
10165 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
10166 to ensure that if a local variable on the stack is overwritten, it will destroy
10167 the value of the guard. When the function exits, the guard on the stack is
10168 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
10169 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
10170 calling the ``__stack_chk_fail()`` function.
10172 '``llvm.stackprotectorcheck``' Intrinsic
10173 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10180 declare void @llvm.stackprotectorcheck(i8** <guard>)
10185 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
10186 created stack protector and if they are not equal calls the
10187 ``__stack_chk_fail()`` function.
10192 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
10193 the variable ``@__stack_chk_guard``.
10198 This intrinsic is provided to perform the stack protector check by comparing
10199 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
10200 values do not match call the ``__stack_chk_fail()`` function.
10202 The reason to provide this as an IR level intrinsic instead of implementing it
10203 via other IR operations is that in order to perform this operation at the IR
10204 level without an intrinsic, one would need to create additional basic blocks to
10205 handle the success/failure cases. This makes it difficult to stop the stack
10206 protector check from disrupting sibling tail calls in Codegen. With this
10207 intrinsic, we are able to generate the stack protector basic blocks late in
10208 codegen after the tail call decision has occurred.
10210 '``llvm.objectsize``' Intrinsic
10211 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10218 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
10219 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
10224 The ``llvm.objectsize`` intrinsic is designed to provide information to
10225 the optimizers to determine at compile time whether a) an operation
10226 (like memcpy) will overflow a buffer that corresponds to an object, or
10227 b) that a runtime check for overflow isn't necessary. An object in this
10228 context means an allocation of a specific class, structure, array, or
10234 The ``llvm.objectsize`` intrinsic takes two arguments. The first
10235 argument is a pointer to or into the ``object``. The second argument is
10236 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
10237 or -1 (if false) when the object size is unknown. The second argument
10238 only accepts constants.
10243 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
10244 the size of the object concerned. If the size cannot be determined at
10245 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
10246 on the ``min`` argument).
10248 '``llvm.expect``' Intrinsic
10249 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10254 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
10259 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
10260 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
10261 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
10266 The ``llvm.expect`` intrinsic provides information about expected (the
10267 most probable) value of ``val``, which can be used by optimizers.
10272 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
10273 a value. The second argument is an expected value, this needs to be a
10274 constant value, variables are not allowed.
10279 This intrinsic is lowered to the ``val``.
10281 '``llvm.assume``' Intrinsic
10282 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10289 declare void @llvm.assume(i1 %cond)
10294 The ``llvm.assume`` allows the optimizer to assume that the provided
10295 condition is true. This information can then be used in simplifying other parts
10301 The condition which the optimizer may assume is always true.
10306 The intrinsic allows the optimizer to assume that the provided condition is
10307 always true whenever the control flow reaches the intrinsic call. No code is
10308 generated for this intrinsic, and instructions that contribute only to the
10309 provided condition are not used for code generation. If the condition is
10310 violated during execution, the behavior is undefined.
10312 Note that the optimizer might limit the transformations performed on values
10313 used by the ``llvm.assume`` intrinsic in order to preserve the instructions
10314 only used to form the intrinsic's input argument. This might prove undesirable
10315 if the extra information provided by the ``llvm.assume`` intrinsic does not cause
10316 sufficient overall improvement in code quality. For this reason,
10317 ``llvm.assume`` should not be used to document basic mathematical invariants
10318 that the optimizer can otherwise deduce or facts that are of little use to the
10323 '``llvm.bitset.test``' Intrinsic
10324 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10331 declare i1 @llvm.bitset.test(i8* %ptr, metadata %bitset) nounwind readnone
10337 The first argument is a pointer to be tested. The second argument is a
10338 metadata string containing the name of a :doc:`bitset <BitSets>`.
10343 The ``llvm.bitset.test`` intrinsic tests whether the given pointer is a
10344 member of the given bitset.
10346 '``llvm.donothing``' Intrinsic
10347 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10354 declare void @llvm.donothing() nounwind readnone
10359 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's one of only
10360 two intrinsics (besides ``llvm.experimental.patchpoint``) that can be called
10361 with an invoke instruction.
10371 This intrinsic does nothing, and it's removed by optimizers and ignored
10374 Stack Map Intrinsics
10375 --------------------
10377 LLVM provides experimental intrinsics to support runtime patching
10378 mechanisms commonly desired in dynamic language JITs. These intrinsics
10379 are described in :doc:`StackMaps`.