1 ==============================
2 LLVM Language Reference Manual
3 ==============================
12 This document is a reference manual for the LLVM assembly language. LLVM
13 is a Static Single Assignment (SSA) based representation that provides
14 type safety, low-level operations, flexibility, and the capability of
15 representing 'all' high-level languages cleanly. It is the common code
16 representation used throughout all phases of the LLVM compilation
22 The LLVM code representation is designed to be used in three different
23 forms: as an in-memory compiler IR, as an on-disk bitcode representation
24 (suitable for fast loading by a Just-In-Time compiler), and as a human
25 readable assembly language representation. This allows LLVM to provide a
26 powerful intermediate representation for efficient compiler
27 transformations and analysis, while providing a natural means to debug
28 and visualize the transformations. The three different forms of LLVM are
29 all equivalent. This document describes the human readable
30 representation and notation.
32 The LLVM representation aims to be light-weight and low-level while
33 being expressive, typed, and extensible at the same time. It aims to be
34 a "universal IR" of sorts, by being at a low enough level that
35 high-level ideas may be cleanly mapped to it (similar to how
36 microprocessors are "universal IR's", allowing many source languages to
37 be mapped to them). By providing type information, LLVM can be used as
38 the target of optimizations: for example, through pointer analysis, it
39 can be proven that a C automatic variable is never accessed outside of
40 the current function, allowing it to be promoted to a simple SSA value
41 instead of a memory location.
48 It is important to note that this document describes 'well formed' LLVM
49 assembly language. There is a difference between what the parser accepts
50 and what is considered 'well formed'. For example, the following
51 instruction is syntactically okay, but not well formed:
57 because the definition of ``%x`` does not dominate all of its uses. The
58 LLVM infrastructure provides a verification pass that may be used to
59 verify that an LLVM module is well formed. This pass is automatically
60 run by the parser after parsing input assembly and by the optimizer
61 before it outputs bitcode. The violations pointed out by the verifier
62 pass indicate bugs in transformation passes or input to the parser.
69 LLVM identifiers come in two basic types: global and local. Global
70 identifiers (functions, global variables) begin with the ``'@'``
71 character. Local identifiers (register names, types) begin with the
72 ``'%'`` character. Additionally, there are three different formats for
73 identifiers, for different purposes:
75 #. Named values are represented as a string of characters with their
76 prefix. For example, ``%foo``, ``@DivisionByZero``,
77 ``%a.really.long.identifier``. The actual regular expression used is
78 '``[%@][-a-zA-Z$._][-a-zA-Z$._0-9]*``'. Identifiers that require other
79 characters in their names can be surrounded with quotes. Special
80 characters may be escaped using ``"\xx"`` where ``xx`` is the ASCII
81 code for the character in hexadecimal. In this way, any character can
82 be used in a name value, even quotes themselves. The ``"\01"`` prefix
83 can be used on global variables to suppress mangling.
84 #. Unnamed values are represented as an unsigned numeric value with
85 their prefix. For example, ``%12``, ``@2``, ``%44``.
86 #. Constants, which are described in the section Constants_ below.
88 LLVM requires that values start with a prefix for two reasons: Compilers
89 don't need to worry about name clashes with reserved words, and the set
90 of reserved words may be expanded in the future without penalty.
91 Additionally, unnamed identifiers allow a compiler to quickly come up
92 with a temporary variable without having to avoid symbol table
95 Reserved words in LLVM are very similar to reserved words in other
96 languages. There are keywords for different opcodes ('``add``',
97 '``bitcast``', '``ret``', etc...), for primitive type names ('``void``',
98 '``i32``', etc...), and others. These reserved words cannot conflict
99 with variable names, because none of them start with a prefix character
100 (``'%'`` or ``'@'``).
102 Here is an example of LLVM code to multiply the integer variable
109 %result = mul i32 %X, 8
111 After strength reduction:
115 %result = shl i32 %X, 3
121 %0 = add i32 %X, %X ; yields i32:%0
122 %1 = add i32 %0, %0 ; yields i32:%1
123 %result = add i32 %1, %1
125 This last way of multiplying ``%X`` by 8 illustrates several important
126 lexical features of LLVM:
128 #. Comments are delimited with a '``;``' and go until the end of line.
129 #. Unnamed temporaries are created when the result of a computation is
130 not assigned to a named value.
131 #. Unnamed temporaries are numbered sequentially (using a per-function
132 incrementing counter, starting with 0). Note that basic blocks and unnamed
133 function parameters are included in this numbering. For example, if the
134 entry basic block is not given a label name and all function parameters are
135 named, then it will get number 0.
137 It also shows a convention that we follow in this document. When
138 demonstrating instructions, we will follow an instruction with a comment
139 that defines the type and name of value produced.
147 LLVM programs are composed of ``Module``'s, each of which is a
148 translation unit of the input programs. Each module consists of
149 functions, global variables, and symbol table entries. Modules may be
150 combined together with the LLVM linker, which merges function (and
151 global variable) definitions, resolves forward declarations, and merges
152 symbol table entries. Here is an example of the "hello world" module:
156 ; Declare the string constant as a global constant.
157 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
159 ; External declaration of the puts function
160 declare i32 @puts(i8* nocapture) nounwind
162 ; Definition of main function
163 define i32 @main() { ; i32()*
164 ; Convert [13 x i8]* to i8 *...
165 %cast210 = getelementptr [13 x i8], [13 x i8]* @.str, i64 0, i64 0
167 ; Call puts function to write out the string to stdout.
168 call i32 @puts(i8* %cast210)
173 !0 = !{i32 42, null, !"string"}
176 This example is made up of a :ref:`global variable <globalvars>` named
177 "``.str``", an external declaration of the "``puts``" function, a
178 :ref:`function definition <functionstructure>` for "``main``" and
179 :ref:`named metadata <namedmetadatastructure>` "``foo``".
181 In general, a module is made up of a list of global values (where both
182 functions and global variables are global values). Global values are
183 represented by a pointer to a memory location (in this case, a pointer
184 to an array of char, and a pointer to a function), and have one of the
185 following :ref:`linkage types <linkage>`.
192 All Global Variables and Functions have one of the following types of
196 Global values with "``private``" linkage are only directly
197 accessible by objects in the current module. In particular, linking
198 code into a module with an private global value may cause the
199 private to be renamed as necessary to avoid collisions. Because the
200 symbol is private to the module, all references can be updated. This
201 doesn't show up in any symbol table in the object file.
203 Similar to private, but the value shows as a local symbol
204 (``STB_LOCAL`` in the case of ELF) in the object file. This
205 corresponds to the notion of the '``static``' keyword in C.
206 ``available_externally``
207 Globals with "``available_externally``" linkage are never emitted
208 into the object file corresponding to the LLVM module. They exist to
209 allow inlining and other optimizations to take place given knowledge
210 of the definition of the global, which is known to be somewhere
211 outside the module. Globals with ``available_externally`` linkage
212 are allowed to be discarded at will, and are otherwise the same as
213 ``linkonce_odr``. This linkage type is only allowed on definitions,
216 Globals with "``linkonce``" linkage are merged with other globals of
217 the same name when linkage occurs. This can be used to implement
218 some forms of inline functions, templates, or other code which must
219 be generated in each translation unit that uses it, but where the
220 body may be overridden with a more definitive definition later.
221 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
222 that ``linkonce`` linkage does not actually allow the optimizer to
223 inline the body of this function into callers because it doesn't
224 know if this definition of the function is the definitive definition
225 within the program or whether it will be overridden by a stronger
226 definition. To enable inlining and other optimizations, use
227 "``linkonce_odr``" linkage.
229 "``weak``" linkage has the same merging semantics as ``linkonce``
230 linkage, except that unreferenced globals with ``weak`` linkage may
231 not be discarded. This is used for globals that are declared "weak"
234 "``common``" linkage is most similar to "``weak``" linkage, but they
235 are used for tentative definitions in C, such as "``int X;``" at
236 global scope. Symbols with "``common``" linkage are merged in the
237 same way as ``weak symbols``, and they may not be deleted if
238 unreferenced. ``common`` symbols may not have an explicit section,
239 must have a zero initializer, and may not be marked
240 ':ref:`constant <globalvars>`'. Functions and aliases may not have
243 .. _linkage_appending:
246 "``appending``" linkage may only be applied to global variables of
247 pointer to array type. When two global variables with appending
248 linkage are linked together, the two global arrays are appended
249 together. This is the LLVM, typesafe, equivalent of having the
250 system linker append together "sections" with identical names when
253 The semantics of this linkage follow the ELF object file model: the
254 symbol is weak until linked, if not linked, the symbol becomes null
255 instead of being an undefined reference.
256 ``linkonce_odr``, ``weak_odr``
257 Some languages allow differing globals to be merged, such as two
258 functions with different semantics. Other languages, such as
259 ``C++``, ensure that only equivalent globals are ever merged (the
260 "one definition rule" --- "ODR"). Such languages can use the
261 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
262 global will only be merged with equivalent globals. These linkage
263 types are otherwise the same as their non-``odr`` versions.
265 If none of the above identifiers are used, the global is externally
266 visible, meaning that it participates in linkage and can be used to
267 resolve external symbol references.
269 It is illegal for a function *declaration* to have any linkage type
270 other than ``external`` or ``extern_weak``.
277 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
278 :ref:`invokes <i_invoke>` can all have an optional calling convention
279 specified for the call. The calling convention of any pair of dynamic
280 caller/callee must match, or the behavior of the program is undefined.
281 The following calling conventions are supported by LLVM, and more may be
284 "``ccc``" - The C calling convention
285 This calling convention (the default if no other calling convention
286 is specified) matches the target C calling conventions. This calling
287 convention supports varargs function calls and tolerates some
288 mismatch in the declared prototype and implemented declaration of
289 the function (as does normal C).
290 "``fastcc``" - The fast calling convention
291 This calling convention attempts to make calls as fast as possible
292 (e.g. by passing things in registers). This calling convention
293 allows the target to use whatever tricks it wants to produce fast
294 code for the target, without having to conform to an externally
295 specified ABI (Application Binary Interface). `Tail calls can only
296 be optimized when this, the GHC or the HiPE convention is
297 used. <CodeGenerator.html#id80>`_ This calling convention does not
298 support varargs and requires the prototype of all callees to exactly
299 match the prototype of the function definition.
300 "``coldcc``" - The cold calling convention
301 This calling convention attempts to make code in the caller as
302 efficient as possible under the assumption that the call is not
303 commonly executed. As such, these calls often preserve all registers
304 so that the call does not break any live ranges in the caller side.
305 This calling convention does not support varargs and requires the
306 prototype of all callees to exactly match the prototype of the
307 function definition. Furthermore the inliner doesn't consider such function
309 "``cc 10``" - GHC convention
310 This calling convention has been implemented specifically for use by
311 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
312 It passes everything in registers, going to extremes to achieve this
313 by disabling callee save registers. This calling convention should
314 not be used lightly but only for specific situations such as an
315 alternative to the *register pinning* performance technique often
316 used when implementing functional programming languages. At the
317 moment only X86 supports this convention and it has the following
320 - On *X86-32* only supports up to 4 bit type parameters. No
321 floating point types are supported.
322 - On *X86-64* only supports up to 10 bit type parameters and 6
323 floating point parameters.
325 This calling convention supports `tail call
326 optimization <CodeGenerator.html#id80>`_ but requires both the
327 caller and callee are using it.
328 "``cc 11``" - The HiPE calling convention
329 This calling convention has been implemented specifically for use by
330 the `High-Performance Erlang
331 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
332 native code compiler of the `Ericsson's Open Source Erlang/OTP
333 system <http://www.erlang.org/download.shtml>`_. It uses more
334 registers for argument passing than the ordinary C calling
335 convention and defines no callee-saved registers. The calling
336 convention properly supports `tail call
337 optimization <CodeGenerator.html#id80>`_ but requires that both the
338 caller and the callee use it. It uses a *register pinning*
339 mechanism, similar to GHC's convention, for keeping frequently
340 accessed runtime components pinned to specific hardware registers.
341 At the moment only X86 supports this convention (both 32 and 64
343 "``webkit_jscc``" - WebKit's JavaScript calling convention
344 This calling convention has been implemented for `WebKit FTL JIT
345 <https://trac.webkit.org/wiki/FTLJIT>`_. It passes arguments on the
346 stack right to left (as cdecl does), and returns a value in the
347 platform's customary return register.
348 "``anyregcc``" - Dynamic calling convention for code patching
349 This is a special convention that supports patching an arbitrary code
350 sequence in place of a call site. This convention forces the call
351 arguments into registers but allows them to be dynamically
352 allocated. This can currently only be used with calls to
353 llvm.experimental.patchpoint because only this intrinsic records
354 the location of its arguments in a side table. See :doc:`StackMaps`.
355 "``preserve_mostcc``" - The `PreserveMost` calling convention
356 This calling convention attempts to make the code in the caller as
357 unintrusive as possible. This convention behaves identically to the `C`
358 calling convention on how arguments and return values are passed, but it
359 uses a different set of caller/callee-saved registers. This alleviates the
360 burden of saving and recovering a large register set before and after the
361 call in the caller. If the arguments are passed in callee-saved registers,
362 then they will be preserved by the callee across the call. This doesn't
363 apply for values returned in callee-saved registers.
365 - On X86-64 the callee preserves all general purpose registers, except for
366 R11. R11 can be used as a scratch register. Floating-point registers
367 (XMMs/YMMs) are not preserved and need to be saved by the caller.
369 The idea behind this convention is to support calls to runtime functions
370 that have a hot path and a cold path. The hot path is usually a small piece
371 of code that doesn't use many registers. The cold path might need to call out to
372 another function and therefore only needs to preserve the caller-saved
373 registers, which haven't already been saved by the caller. The
374 `PreserveMost` calling convention is very similar to the `cold` calling
375 convention in terms of caller/callee-saved registers, but they are used for
376 different types of function calls. `coldcc` is for function calls that are
377 rarely executed, whereas `preserve_mostcc` function calls are intended to be
378 on the hot path and definitely executed a lot. Furthermore `preserve_mostcc`
379 doesn't prevent the inliner from inlining the function call.
381 This calling convention will be used by a future version of the ObjectiveC
382 runtime and should therefore still be considered experimental at this time.
383 Although this convention was created to optimize certain runtime calls to
384 the ObjectiveC runtime, it is not limited to this runtime and might be used
385 by other runtimes in the future too. The current implementation only
386 supports X86-64, but the intention is to support more architectures in the
388 "``preserve_allcc``" - The `PreserveAll` calling convention
389 This calling convention attempts to make the code in the caller even less
390 intrusive than the `PreserveMost` calling convention. This calling
391 convention also behaves identical to the `C` calling convention on how
392 arguments and return values are passed, but it uses a different set of
393 caller/callee-saved registers. This removes the burden of saving and
394 recovering a large register set before and after the call in the caller. If
395 the arguments are passed in callee-saved registers, then they will be
396 preserved by the callee across the call. This doesn't apply for values
397 returned in callee-saved registers.
399 - On X86-64 the callee preserves all general purpose registers, except for
400 R11. R11 can be used as a scratch register. Furthermore it also preserves
401 all floating-point registers (XMMs/YMMs).
403 The idea behind this convention is to support calls to runtime functions
404 that don't need to call out to any other functions.
406 This calling convention, like the `PreserveMost` calling convention, will be
407 used by a future version of the ObjectiveC runtime and should be considered
408 experimental at this time.
409 "``cxx_fast_tlscc``" - The `CXX_FAST_TLS` calling convention for access functions
410 Clang generates an access function to access C++-style TLS. The access
411 function generally has an entry block, an exit block and an initialization
412 block that is run at the first time. The entry and exit blocks can access
413 a few TLS IR variables, each access will be lowered to a platform-specific
416 This calling convention aims to minimize overhead in the caller by
417 preserving as many registers as possible (all the registers that are
418 perserved on the fast path, composed of the entry and exit blocks).
420 This calling convention behaves identical to the `C` calling convention on
421 how arguments and return values are passed, but it uses a different set of
422 caller/callee-saved registers.
424 Given that each platform has its own lowering sequence, hence its own set
425 of preserved registers, we can't use the existing `PreserveMost`.
427 - On X86-64 the callee preserves all general purpose registers, except for
429 "``cc <n>``" - Numbered convention
430 Any calling convention may be specified by number, allowing
431 target-specific calling conventions to be used. Target specific
432 calling conventions start at 64.
434 More calling conventions can be added/defined on an as-needed basis, to
435 support Pascal conventions or any other well-known target-independent
438 .. _visibilitystyles:
443 All Global Variables and Functions have one of the following visibility
446 "``default``" - Default style
447 On targets that use the ELF object file format, default visibility
448 means that the declaration is visible to other modules and, in
449 shared libraries, means that the declared entity may be overridden.
450 On Darwin, default visibility means that the declaration is visible
451 to other modules. Default visibility corresponds to "external
452 linkage" in the language.
453 "``hidden``" - Hidden style
454 Two declarations of an object with hidden visibility refer to the
455 same object if they are in the same shared object. Usually, hidden
456 visibility indicates that the symbol will not be placed into the
457 dynamic symbol table, so no other module (executable or shared
458 library) can reference it directly.
459 "``protected``" - Protected style
460 On ELF, protected visibility indicates that the symbol will be
461 placed in the dynamic symbol table, but that references within the
462 defining module will bind to the local symbol. That is, the symbol
463 cannot be overridden by another module.
465 A symbol with ``internal`` or ``private`` linkage must have ``default``
473 All Global Variables, Functions and Aliases can have one of the following
477 "``dllimport``" causes the compiler to reference a function or variable via
478 a global pointer to a pointer that is set up by the DLL exporting the
479 symbol. On Microsoft Windows targets, the pointer name is formed by
480 combining ``__imp_`` and the function or variable name.
482 "``dllexport``" causes the compiler to provide a global pointer to a pointer
483 in a DLL, so that it can be referenced with the ``dllimport`` attribute. On
484 Microsoft Windows targets, the pointer name is formed by combining
485 ``__imp_`` and the function or variable name. Since this storage class
486 exists for defining a dll interface, the compiler, assembler and linker know
487 it is externally referenced and must refrain from deleting the symbol.
491 Thread Local Storage Models
492 ---------------------------
494 A variable may be defined as ``thread_local``, which means that it will
495 not be shared by threads (each thread will have a separated copy of the
496 variable). Not all targets support thread-local variables. Optionally, a
497 TLS model may be specified:
500 For variables that are only used within the current shared library.
502 For variables in modules that will not be loaded dynamically.
504 For variables defined in the executable and only used within it.
506 If no explicit model is given, the "general dynamic" model is used.
508 The models correspond to the ELF TLS models; see `ELF Handling For
509 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
510 more information on under which circumstances the different models may
511 be used. The target may choose a different TLS model if the specified
512 model is not supported, or if a better choice of model can be made.
514 A model can also be specified in an alias, but then it only governs how
515 the alias is accessed. It will not have any effect in the aliasee.
517 For platforms without linker support of ELF TLS model, the -femulated-tls
518 flag can be used to generate GCC compatible emulated TLS code.
525 LLVM IR allows you to specify both "identified" and "literal" :ref:`structure
526 types <t_struct>`. Literal types are uniqued structurally, but identified types
527 are never uniqued. An :ref:`opaque structural type <t_opaque>` can also be used
528 to forward declare a type that is not yet available.
530 An example of an identified structure specification is:
534 %mytype = type { %mytype*, i32 }
536 Prior to the LLVM 3.0 release, identified types were structurally uniqued. Only
537 literal types are uniqued in recent versions of LLVM.
544 Global variables define regions of memory allocated at compilation time
547 Global variable definitions must be initialized.
549 Global variables in other translation units can also be declared, in which
550 case they don't have an initializer.
552 Either global variable definitions or declarations may have an explicit section
553 to be placed in and may have an optional explicit alignment specified.
555 A variable may be defined as a global ``constant``, which indicates that
556 the contents of the variable will **never** be modified (enabling better
557 optimization, allowing the global data to be placed in the read-only
558 section of an executable, etc). Note that variables that need runtime
559 initialization cannot be marked ``constant`` as there is a store to the
562 LLVM explicitly allows *declarations* of global variables to be marked
563 constant, even if the final definition of the global is not. This
564 capability can be used to enable slightly better optimization of the
565 program, but requires the language definition to guarantee that
566 optimizations based on the 'constantness' are valid for the translation
567 units that do not include the definition.
569 As SSA values, global variables define pointer values that are in scope
570 (i.e. they dominate) all basic blocks in the program. Global variables
571 always define a pointer to their "content" type because they describe a
572 region of memory, and all memory objects in LLVM are accessed through
575 Global variables can be marked with ``unnamed_addr`` which indicates
576 that the address is not significant, only the content. Constants marked
577 like this can be merged with other constants if they have the same
578 initializer. Note that a constant with significant address *can* be
579 merged with a ``unnamed_addr`` constant, the result being a constant
580 whose address is significant.
582 A global variable may be declared to reside in a target-specific
583 numbered address space. For targets that support them, address spaces
584 may affect how optimizations are performed and/or what target
585 instructions are used to access the variable. The default address space
586 is zero. The address space qualifier must precede any other attributes.
588 LLVM allows an explicit section to be specified for globals. If the
589 target supports it, it will emit globals to the section specified.
590 Additionally, the global can placed in a comdat if the target has the necessary
593 By default, global initializers are optimized by assuming that global
594 variables defined within the module are not modified from their
595 initial values before the start of the global initializer. This is
596 true even for variables potentially accessible from outside the
597 module, including those with external linkage or appearing in
598 ``@llvm.used`` or dllexported variables. This assumption may be suppressed
599 by marking the variable with ``externally_initialized``.
601 An explicit alignment may be specified for a global, which must be a
602 power of 2. If not present, or if the alignment is set to zero, the
603 alignment of the global is set by the target to whatever it feels
604 convenient. If an explicit alignment is specified, the global is forced
605 to have exactly that alignment. Targets and optimizers are not allowed
606 to over-align the global if the global has an assigned section. In this
607 case, the extra alignment could be observable: for example, code could
608 assume that the globals are densely packed in their section and try to
609 iterate over them as an array, alignment padding would break this
610 iteration. The maximum alignment is ``1 << 29``.
612 Globals can also have a :ref:`DLL storage class <dllstorageclass>`.
614 Variables and aliases can have a
615 :ref:`Thread Local Storage Model <tls_model>`.
619 [@<GlobalVarName> =] [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal]
620 [unnamed_addr] [AddrSpace] [ExternallyInitialized]
621 <global | constant> <Type> [<InitializerConstant>]
622 [, section "name"] [, comdat [($name)]]
623 [, align <Alignment>]
625 For example, the following defines a global in a numbered address space
626 with an initializer, section, and alignment:
630 @G = addrspace(5) constant float 1.0, section "foo", align 4
632 The following example just declares a global variable
636 @G = external global i32
638 The following example defines a thread-local global with the
639 ``initialexec`` TLS model:
643 @G = thread_local(initialexec) global i32 0, align 4
645 .. _functionstructure:
650 LLVM function definitions consist of the "``define``" keyword, an
651 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
652 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
653 an optional :ref:`calling convention <callingconv>`,
654 an optional ``unnamed_addr`` attribute, a return type, an optional
655 :ref:`parameter attribute <paramattrs>` for the return type, a function
656 name, a (possibly empty) argument list (each with optional :ref:`parameter
657 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
658 an optional section, an optional alignment,
659 an optional :ref:`comdat <langref_comdats>`,
660 an optional :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
661 an optional :ref:`prologue <prologuedata>`,
662 an optional :ref:`personality <personalityfn>`,
663 an optional list of attached :ref:`metadata <metadata>`,
664 an opening curly brace, a list of basic blocks, and a closing curly brace.
666 LLVM function declarations consist of the "``declare``" keyword, an
667 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
668 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
669 an optional :ref:`calling convention <callingconv>`,
670 an optional ``unnamed_addr`` attribute, a return type, an optional
671 :ref:`parameter attribute <paramattrs>` for the return type, a function
672 name, a possibly empty list of arguments, an optional alignment, an optional
673 :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
674 and an optional :ref:`prologue <prologuedata>`.
676 A function definition contains a list of basic blocks, forming the CFG (Control
677 Flow Graph) for the function. Each basic block may optionally start with a label
678 (giving the basic block a symbol table entry), contains a list of instructions,
679 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
680 function return). If an explicit label is not provided, a block is assigned an
681 implicit numbered label, using the next value from the same counter as used for
682 unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
683 entry block does not have an explicit label, it will be assigned label "%0",
684 then the first unnamed temporary in that block will be "%1", etc.
686 The first basic block in a function is special in two ways: it is
687 immediately executed on entrance to the function, and it is not allowed
688 to have predecessor basic blocks (i.e. there can not be any branches to
689 the entry block of a function). Because the block can have no
690 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
692 LLVM allows an explicit section to be specified for functions. If the
693 target supports it, it will emit functions to the section specified.
694 Additionally, the function can be placed in a COMDAT.
696 An explicit alignment may be specified for a function. If not present,
697 or if the alignment is set to zero, the alignment of the function is set
698 by the target to whatever it feels convenient. If an explicit alignment
699 is specified, the function is forced to have at least that much
700 alignment. All alignments must be a power of 2.
702 If the ``unnamed_addr`` attribute is given, the address is known to not
703 be significant and two identical functions can be merged.
707 define [linkage] [visibility] [DLLStorageClass]
709 <ResultType> @<FunctionName> ([argument list])
710 [unnamed_addr] [fn Attrs] [section "name"] [comdat [($name)]]
711 [align N] [gc] [prefix Constant] [prologue Constant]
712 [personality Constant] (!name !N)* { ... }
714 The argument list is a comma separated sequence of arguments where each
715 argument is of the following form:
719 <type> [parameter Attrs] [name]
727 Aliases, unlike function or variables, don't create any new data. They
728 are just a new symbol and metadata for an existing position.
730 Aliases have a name and an aliasee that is either a global value or a
733 Aliases may have an optional :ref:`linkage type <linkage>`, an optional
734 :ref:`visibility style <visibility>`, an optional :ref:`DLL storage class
735 <dllstorageclass>` and an optional :ref:`tls model <tls_model>`.
739 @<Name> = [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal] [unnamed_addr] alias <AliaseeTy>, <AliaseeTy>* @<Aliasee>
741 The linkage must be one of ``private``, ``internal``, ``linkonce``, ``weak``,
742 ``linkonce_odr``, ``weak_odr``, ``external``. Note that some system linkers
743 might not correctly handle dropping a weak symbol that is aliased.
745 Aliases that are not ``unnamed_addr`` are guaranteed to have the same address as
746 the aliasee expression. ``unnamed_addr`` ones are only guaranteed to point
749 Since aliases are only a second name, some restrictions apply, of which
750 some can only be checked when producing an object file:
752 * The expression defining the aliasee must be computable at assembly
753 time. Since it is just a name, no relocations can be used.
755 * No alias in the expression can be weak as the possibility of the
756 intermediate alias being overridden cannot be represented in an
759 * No global value in the expression can be a declaration, since that
760 would require a relocation, which is not possible.
767 Comdat IR provides access to COFF and ELF object file COMDAT functionality.
769 Comdats have a name which represents the COMDAT key. All global objects that
770 specify this key will only end up in the final object file if the linker chooses
771 that key over some other key. Aliases are placed in the same COMDAT that their
772 aliasee computes to, if any.
774 Comdats have a selection kind to provide input on how the linker should
775 choose between keys in two different object files.
779 $<Name> = comdat SelectionKind
781 The selection kind must be one of the following:
784 The linker may choose any COMDAT key, the choice is arbitrary.
786 The linker may choose any COMDAT key but the sections must contain the
789 The linker will choose the section containing the largest COMDAT key.
791 The linker requires that only section with this COMDAT key exist.
793 The linker may choose any COMDAT key but the sections must contain the
796 Note that the Mach-O platform doesn't support COMDATs and ELF only supports
797 ``any`` as a selection kind.
799 Here is an example of a COMDAT group where a function will only be selected if
800 the COMDAT key's section is the largest:
804 $foo = comdat largest
805 @foo = global i32 2, comdat($foo)
807 define void @bar() comdat($foo) {
811 As a syntactic sugar the ``$name`` can be omitted if the name is the same as
817 @foo = global i32 2, comdat
820 In a COFF object file, this will create a COMDAT section with selection kind
821 ``IMAGE_COMDAT_SELECT_LARGEST`` containing the contents of the ``@foo`` symbol
822 and another COMDAT section with selection kind
823 ``IMAGE_COMDAT_SELECT_ASSOCIATIVE`` which is associated with the first COMDAT
824 section and contains the contents of the ``@bar`` symbol.
826 There are some restrictions on the properties of the global object.
827 It, or an alias to it, must have the same name as the COMDAT group when
829 The contents and size of this object may be used during link-time to determine
830 which COMDAT groups get selected depending on the selection kind.
831 Because the name of the object must match the name of the COMDAT group, the
832 linkage of the global object must not be local; local symbols can get renamed
833 if a collision occurs in the symbol table.
835 The combined use of COMDATS and section attributes may yield surprising results.
842 @g1 = global i32 42, section "sec", comdat($foo)
843 @g2 = global i32 42, section "sec", comdat($bar)
845 From the object file perspective, this requires the creation of two sections
846 with the same name. This is necessary because both globals belong to different
847 COMDAT groups and COMDATs, at the object file level, are represented by
850 Note that certain IR constructs like global variables and functions may
851 create COMDATs in the object file in addition to any which are specified using
852 COMDAT IR. This arises when the code generator is configured to emit globals
853 in individual sections (e.g. when `-data-sections` or `-function-sections`
854 is supplied to `llc`).
856 .. _namedmetadatastructure:
861 Named metadata is a collection of metadata. :ref:`Metadata
862 nodes <metadata>` (but not metadata strings) are the only valid
863 operands for a named metadata.
865 #. Named metadata are represented as a string of characters with the
866 metadata prefix. The rules for metadata names are the same as for
867 identifiers, but quoted names are not allowed. ``"\xx"`` type escapes
868 are still valid, which allows any character to be part of a name.
872 ; Some unnamed metadata nodes, which are referenced by the named metadata.
877 !name = !{!0, !1, !2}
884 The return type and each parameter of a function type may have a set of
885 *parameter attributes* associated with them. Parameter attributes are
886 used to communicate additional information about the result or
887 parameters of a function. Parameter attributes are considered to be part
888 of the function, not of the function type, so functions with different
889 parameter attributes can have the same function type.
891 Parameter attributes are simple keywords that follow the type specified.
892 If multiple parameter attributes are needed, they are space separated.
897 declare i32 @printf(i8* noalias nocapture, ...)
898 declare i32 @atoi(i8 zeroext)
899 declare signext i8 @returns_signed_char()
901 Note that any attributes for the function result (``nounwind``,
902 ``readonly``) come immediately after the argument list.
904 Currently, only the following parameter attributes are defined:
907 This indicates to the code generator that the parameter or return
908 value should be zero-extended to the extent required by the target's
909 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
910 the caller (for a parameter) or the callee (for a return value).
912 This indicates to the code generator that the parameter or return
913 value should be sign-extended to the extent required by the target's
914 ABI (which is usually 32-bits) by the caller (for a parameter) or
915 the callee (for a return value).
917 This indicates that this parameter or return value should be treated
918 in a special target-dependent fashion while emitting code for
919 a function call or return (usually, by putting it in a register as
920 opposed to memory, though some targets use it to distinguish between
921 two different kinds of registers). Use of this attribute is
924 This indicates that the pointer parameter should really be passed by
925 value to the function. The attribute implies that a hidden copy of
926 the pointee is made between the caller and the callee, so the callee
927 is unable to modify the value in the caller. This attribute is only
928 valid on LLVM pointer arguments. It is generally used to pass
929 structs and arrays by value, but is also valid on pointers to
930 scalars. The copy is considered to belong to the caller not the
931 callee (for example, ``readonly`` functions should not write to
932 ``byval`` parameters). This is not a valid attribute for return
935 The byval attribute also supports specifying an alignment with the
936 align attribute. It indicates the alignment of the stack slot to
937 form and the known alignment of the pointer specified to the call
938 site. If the alignment is not specified, then the code generator
939 makes a target-specific assumption.
945 The ``inalloca`` argument attribute allows the caller to take the
946 address of outgoing stack arguments. An ``inalloca`` argument must
947 be a pointer to stack memory produced by an ``alloca`` instruction.
948 The alloca, or argument allocation, must also be tagged with the
949 inalloca keyword. Only the last argument may have the ``inalloca``
950 attribute, and that argument is guaranteed to be passed in memory.
952 An argument allocation may be used by a call at most once because
953 the call may deallocate it. The ``inalloca`` attribute cannot be
954 used in conjunction with other attributes that affect argument
955 storage, like ``inreg``, ``nest``, ``sret``, or ``byval``. The
956 ``inalloca`` attribute also disables LLVM's implicit lowering of
957 large aggregate return values, which means that frontend authors
958 must lower them with ``sret`` pointers.
960 When the call site is reached, the argument allocation must have
961 been the most recent stack allocation that is still live, or the
962 results are undefined. It is possible to allocate additional stack
963 space after an argument allocation and before its call site, but it
964 must be cleared off with :ref:`llvm.stackrestore
967 See :doc:`InAlloca` for more information on how to use this
971 This indicates that the pointer parameter specifies the address of a
972 structure that is the return value of the function in the source
973 program. This pointer must be guaranteed by the caller to be valid:
974 loads and stores to the structure may be assumed by the callee
975 not to trap and to be properly aligned. This may only be applied to
976 the first parameter. This is not a valid attribute for return
980 This indicates that the pointer value may be assumed by the optimizer to
981 have the specified alignment.
983 Note that this attribute has additional semantics when combined with the
989 This indicates that objects accessed via pointer values
990 :ref:`based <pointeraliasing>` on the argument or return value are not also
991 accessed, during the execution of the function, via pointer values not
992 *based* on the argument or return value. The attribute on a return value
993 also has additional semantics described below. The caller shares the
994 responsibility with the callee for ensuring that these requirements are met.
995 For further details, please see the discussion of the NoAlias response in
996 :ref:`alias analysis <Must, May, or No>`.
998 Note that this definition of ``noalias`` is intentionally similar
999 to the definition of ``restrict`` in C99 for function arguments.
1001 For function return values, C99's ``restrict`` is not meaningful,
1002 while LLVM's ``noalias`` is. Furthermore, the semantics of the ``noalias``
1003 attribute on return values are stronger than the semantics of the attribute
1004 when used on function arguments. On function return values, the ``noalias``
1005 attribute indicates that the function acts like a system memory allocation
1006 function, returning a pointer to allocated storage disjoint from the
1007 storage for any other object accessible to the caller.
1010 This indicates that the callee does not make any copies of the
1011 pointer that outlive the callee itself. This is not a valid
1012 attribute for return values.
1017 This indicates that the pointer parameter can be excised using the
1018 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
1019 attribute for return values and can only be applied to one parameter.
1022 This indicates that the function always returns the argument as its return
1023 value. This is an optimization hint to the code generator when generating
1024 the caller, allowing tail call optimization and omission of register saves
1025 and restores in some cases; it is not checked or enforced when generating
1026 the callee. The parameter and the function return type must be valid
1027 operands for the :ref:`bitcast instruction <i_bitcast>`. This is not a
1028 valid attribute for return values and can only be applied to one parameter.
1031 This indicates that the parameter or return pointer is not null. This
1032 attribute may only be applied to pointer typed parameters. This is not
1033 checked or enforced by LLVM, the caller must ensure that the pointer
1034 passed in is non-null, or the callee must ensure that the returned pointer
1037 ``dereferenceable(<n>)``
1038 This indicates that the parameter or return pointer is dereferenceable. This
1039 attribute may only be applied to pointer typed parameters. A pointer that
1040 is dereferenceable can be loaded from speculatively without a risk of
1041 trapping. The number of bytes known to be dereferenceable must be provided
1042 in parentheses. It is legal for the number of bytes to be less than the
1043 size of the pointee type. The ``nonnull`` attribute does not imply
1044 dereferenceability (consider a pointer to one element past the end of an
1045 array), however ``dereferenceable(<n>)`` does imply ``nonnull`` in
1046 ``addrspace(0)`` (which is the default address space).
1048 ``dereferenceable_or_null(<n>)``
1049 This indicates that the parameter or return value isn't both
1050 non-null and non-dereferenceable (up to ``<n>`` bytes) at the same
1051 time. All non-null pointers tagged with
1052 ``dereferenceable_or_null(<n>)`` are ``dereferenceable(<n>)``.
1053 For address space 0 ``dereferenceable_or_null(<n>)`` implies that
1054 a pointer is exactly one of ``dereferenceable(<n>)`` or ``null``,
1055 and in other address spaces ``dereferenceable_or_null(<n>)``
1056 implies that a pointer is at least one of ``dereferenceable(<n>)``
1057 or ``null`` (i.e. it may be both ``null`` and
1058 ``dereferenceable(<n>)``). This attribute may only be applied to
1059 pointer typed parameters.
1063 Garbage Collector Strategy Names
1064 --------------------------------
1066 Each function may specify a garbage collector strategy name, which is simply a
1069 .. code-block:: llvm
1071 define void @f() gc "name" { ... }
1073 The supported values of *name* includes those :ref:`built in to LLVM
1074 <builtin-gc-strategies>` and any provided by loaded plugins. Specifying a GC
1075 strategy will cause the compiler to alter its output in order to support the
1076 named garbage collection algorithm. Note that LLVM itself does not contain a
1077 garbage collector, this functionality is restricted to generating machine code
1078 which can interoperate with a collector provided externally.
1085 Prefix data is data associated with a function which the code
1086 generator will emit immediately before the function's entrypoint.
1087 The purpose of this feature is to allow frontends to associate
1088 language-specific runtime metadata with specific functions and make it
1089 available through the function pointer while still allowing the
1090 function pointer to be called.
1092 To access the data for a given function, a program may bitcast the
1093 function pointer to a pointer to the constant's type and dereference
1094 index -1. This implies that the IR symbol points just past the end of
1095 the prefix data. For instance, take the example of a function annotated
1096 with a single ``i32``,
1098 .. code-block:: llvm
1100 define void @f() prefix i32 123 { ... }
1102 The prefix data can be referenced as,
1104 .. code-block:: llvm
1106 %0 = bitcast void* () @f to i32*
1107 %a = getelementptr inbounds i32, i32* %0, i32 -1
1108 %b = load i32, i32* %a
1110 Prefix data is laid out as if it were an initializer for a global variable
1111 of the prefix data's type. The function will be placed such that the
1112 beginning of the prefix data is aligned. This means that if the size
1113 of the prefix data is not a multiple of the alignment size, the
1114 function's entrypoint will not be aligned. If alignment of the
1115 function's entrypoint is desired, padding must be added to the prefix
1118 A function may have prefix data but no body. This has similar semantics
1119 to the ``available_externally`` linkage in that the data may be used by the
1120 optimizers but will not be emitted in the object file.
1127 The ``prologue`` attribute allows arbitrary code (encoded as bytes) to
1128 be inserted prior to the function body. This can be used for enabling
1129 function hot-patching and instrumentation.
1131 To maintain the semantics of ordinary function calls, the prologue data must
1132 have a particular format. Specifically, it must begin with a sequence of
1133 bytes which decode to a sequence of machine instructions, valid for the
1134 module's target, which transfer control to the point immediately succeeding
1135 the prologue data, without performing any other visible action. This allows
1136 the inliner and other passes to reason about the semantics of the function
1137 definition without needing to reason about the prologue data. Obviously this
1138 makes the format of the prologue data highly target dependent.
1140 A trivial example of valid prologue data for the x86 architecture is ``i8 144``,
1141 which encodes the ``nop`` instruction:
1143 .. code-block:: llvm
1145 define void @f() prologue i8 144 { ... }
1147 Generally prologue data can be formed by encoding a relative branch instruction
1148 which skips the metadata, as in this example of valid prologue data for the
1149 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
1151 .. code-block:: llvm
1153 %0 = type <{ i8, i8, i8* }>
1155 define void @f() prologue %0 <{ i8 235, i8 8, i8* @md}> { ... }
1157 A function may have prologue data but no body. This has similar semantics
1158 to the ``available_externally`` linkage in that the data may be used by the
1159 optimizers but will not be emitted in the object file.
1163 Personality Function
1164 --------------------
1166 The ``personality`` attribute permits functions to specify what function
1167 to use for exception handling.
1174 Attribute groups are groups of attributes that are referenced by objects within
1175 the IR. They are important for keeping ``.ll`` files readable, because a lot of
1176 functions will use the same set of attributes. In the degenerative case of a
1177 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
1178 group will capture the important command line flags used to build that file.
1180 An attribute group is a module-level object. To use an attribute group, an
1181 object references the attribute group's ID (e.g. ``#37``). An object may refer
1182 to more than one attribute group. In that situation, the attributes from the
1183 different groups are merged.
1185 Here is an example of attribute groups for a function that should always be
1186 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
1188 .. code-block:: llvm
1190 ; Target-independent attributes:
1191 attributes #0 = { alwaysinline alignstack=4 }
1193 ; Target-dependent attributes:
1194 attributes #1 = { "no-sse" }
1196 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
1197 define void @f() #0 #1 { ... }
1204 Function attributes are set to communicate additional information about
1205 a function. Function attributes are considered to be part of the
1206 function, not of the function type, so functions with different function
1207 attributes can have the same function type.
1209 Function attributes are simple keywords that follow the type specified.
1210 If multiple attributes are needed, they are space separated. For
1213 .. code-block:: llvm
1215 define void @f() noinline { ... }
1216 define void @f() alwaysinline { ... }
1217 define void @f() alwaysinline optsize { ... }
1218 define void @f() optsize { ... }
1221 This attribute indicates that, when emitting the prologue and
1222 epilogue, the backend should forcibly align the stack pointer.
1223 Specify the desired alignment, which must be a power of two, in
1226 This attribute indicates that the inliner should attempt to inline
1227 this function into callers whenever possible, ignoring any active
1228 inlining size threshold for this caller.
1230 This indicates that the callee function at a call site should be
1231 recognized as a built-in function, even though the function's declaration
1232 uses the ``nobuiltin`` attribute. This is only valid at call sites for
1233 direct calls to functions that are declared with the ``nobuiltin``
1236 This attribute indicates that this function is rarely called. When
1237 computing edge weights, basic blocks post-dominated by a cold
1238 function call are also considered to be cold; and, thus, given low
1241 This attribute indicates that the callee is dependent on a convergent
1242 thread execution pattern under certain parallel execution models.
1243 Transformations that are execution model agnostic may not make the execution
1244 of a convergent operation control dependent on any additional values.
1246 This attribute indicates that the source code contained a hint that
1247 inlining this function is desirable (such as the "inline" keyword in
1248 C/C++). It is just a hint; it imposes no requirements on the
1251 This attribute indicates that the function should be added to a
1252 jump-instruction table at code-generation time, and that all address-taken
1253 references to this function should be replaced with a reference to the
1254 appropriate jump-instruction-table function pointer. Note that this creates
1255 a new pointer for the original function, which means that code that depends
1256 on function-pointer identity can break. So, any function annotated with
1257 ``jumptable`` must also be ``unnamed_addr``.
1259 This attribute suggests that optimization passes and code generator
1260 passes make choices that keep the code size of this function as small
1261 as possible and perform optimizations that may sacrifice runtime
1262 performance in order to minimize the size of the generated code.
1264 This attribute disables prologue / epilogue emission for the
1265 function. This can have very system-specific consequences.
1267 This indicates that the callee function at a call site is not recognized as
1268 a built-in function. LLVM will retain the original call and not replace it
1269 with equivalent code based on the semantics of the built-in function, unless
1270 the call site uses the ``builtin`` attribute. This is valid at call sites
1271 and on function declarations and definitions.
1273 This attribute indicates that calls to the function cannot be
1274 duplicated. A call to a ``noduplicate`` function may be moved
1275 within its parent function, but may not be duplicated within
1276 its parent function.
1278 A function containing a ``noduplicate`` call may still
1279 be an inlining candidate, provided that the call is not
1280 duplicated by inlining. That implies that the function has
1281 internal linkage and only has one call site, so the original
1282 call is dead after inlining.
1284 This attributes disables implicit floating point instructions.
1286 This attribute indicates that the inliner should never inline this
1287 function in any situation. This attribute may not be used together
1288 with the ``alwaysinline`` attribute.
1290 This attribute suppresses lazy symbol binding for the function. This
1291 may make calls to the function faster, at the cost of extra program
1292 startup time if the function is not called during program startup.
1294 This attribute indicates that the code generator should not use a
1295 red zone, even if the target-specific ABI normally permits it.
1297 This function attribute indicates that the function never returns
1298 normally. This produces undefined behavior at runtime if the
1299 function ever does dynamically return.
1301 This function attribute indicates that the function does not call itself
1302 either directly or indirectly down any possible call path. This produces
1303 undefined behavior at runtime if the function ever does recurse.
1305 This function attribute indicates that the function never raises an
1306 exception. If the function does raise an exception, its runtime
1307 behavior is undefined. However, functions marked nounwind may still
1308 trap or generate asynchronous exceptions. Exception handling schemes
1309 that are recognized by LLVM to handle asynchronous exceptions, such
1310 as SEH, will still provide their implementation defined semantics.
1312 This function attribute indicates that most optimization passes will skip
1313 this function, with the exception of interprocedural optimization passes.
1314 Code generation defaults to the "fast" instruction selector.
1315 This attribute cannot be used together with the ``alwaysinline``
1316 attribute; this attribute is also incompatible
1317 with the ``minsize`` attribute and the ``optsize`` attribute.
1319 This attribute requires the ``noinline`` attribute to be specified on
1320 the function as well, so the function is never inlined into any caller.
1321 Only functions with the ``alwaysinline`` attribute are valid
1322 candidates for inlining into the body of this function.
1324 This attribute suggests that optimization passes and code generator
1325 passes make choices that keep the code size of this function low,
1326 and otherwise do optimizations specifically to reduce code size as
1327 long as they do not significantly impact runtime performance.
1329 On a function, this attribute indicates that the function computes its
1330 result (or decides to unwind an exception) based strictly on its arguments,
1331 without dereferencing any pointer arguments or otherwise accessing
1332 any mutable state (e.g. memory, control registers, etc) visible to
1333 caller functions. It does not write through any pointer arguments
1334 (including ``byval`` arguments) and never changes any state visible
1335 to callers. This means that it cannot unwind exceptions by calling
1336 the ``C++`` exception throwing methods.
1338 On an argument, this attribute indicates that the function does not
1339 dereference that pointer argument, even though it may read or write the
1340 memory that the pointer points to if accessed through other pointers.
1342 On a function, this attribute indicates that the function does not write
1343 through any pointer arguments (including ``byval`` arguments) or otherwise
1344 modify any state (e.g. memory, control registers, etc) visible to
1345 caller functions. It may dereference pointer arguments and read
1346 state that may be set in the caller. A readonly function always
1347 returns the same value (or unwinds an exception identically) when
1348 called with the same set of arguments and global state. It cannot
1349 unwind an exception by calling the ``C++`` exception throwing
1352 On an argument, this attribute indicates that the function does not write
1353 through this pointer argument, even though it may write to the memory that
1354 the pointer points to.
1356 This attribute indicates that the only memory accesses inside function are
1357 loads and stores from objects pointed to by its pointer-typed arguments,
1358 with arbitrary offsets. Or in other words, all memory operations in the
1359 function can refer to memory only using pointers based on its function
1361 Note that ``argmemonly`` can be used together with ``readonly`` attribute
1362 in order to specify that function reads only from its arguments.
1364 This attribute indicates that this function can return twice. The C
1365 ``setjmp`` is an example of such a function. The compiler disables
1366 some optimizations (like tail calls) in the caller of these
1369 This attribute indicates that
1370 `SafeStack <http://clang.llvm.org/docs/SafeStack.html>`_
1371 protection is enabled for this function.
1373 If a function that has a ``safestack`` attribute is inlined into a
1374 function that doesn't have a ``safestack`` attribute or which has an
1375 ``ssp``, ``sspstrong`` or ``sspreq`` attribute, then the resulting
1376 function will have a ``safestack`` attribute.
1377 ``sanitize_address``
1378 This attribute indicates that AddressSanitizer checks
1379 (dynamic address safety analysis) are enabled for this function.
1381 This attribute indicates that MemorySanitizer checks (dynamic detection
1382 of accesses to uninitialized memory) are enabled for this function.
1384 This attribute indicates that ThreadSanitizer checks
1385 (dynamic thread safety analysis) are enabled for this function.
1387 This attribute indicates that the function should emit a stack
1388 smashing protector. It is in the form of a "canary" --- a random value
1389 placed on the stack before the local variables that's checked upon
1390 return from the function to see if it has been overwritten. A
1391 heuristic is used to determine if a function needs stack protectors
1392 or not. The heuristic used will enable protectors for functions with:
1394 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1395 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1396 - Calls to alloca() with variable sizes or constant sizes greater than
1397 ``ssp-buffer-size``.
1399 Variables that are identified as requiring a protector will be arranged
1400 on the stack such that they are adjacent to the stack protector guard.
1402 If a function that has an ``ssp`` attribute is inlined into a
1403 function that doesn't have an ``ssp`` attribute, then the resulting
1404 function will have an ``ssp`` attribute.
1406 This attribute indicates that the function should *always* emit a
1407 stack smashing protector. This overrides the ``ssp`` function
1410 Variables that are identified as requiring a protector will be arranged
1411 on the stack such that they are adjacent to the stack protector guard.
1412 The specific layout rules are:
1414 #. Large arrays and structures containing large arrays
1415 (``>= ssp-buffer-size``) are closest to the stack protector.
1416 #. Small arrays and structures containing small arrays
1417 (``< ssp-buffer-size``) are 2nd closest to the protector.
1418 #. Variables that have had their address taken are 3rd closest to the
1421 If a function that has an ``sspreq`` attribute is inlined into a
1422 function that doesn't have an ``sspreq`` attribute or which has an
1423 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1424 an ``sspreq`` attribute.
1426 This attribute indicates that the function should emit a stack smashing
1427 protector. This attribute causes a strong heuristic to be used when
1428 determining if a function needs stack protectors. The strong heuristic
1429 will enable protectors for functions with:
1431 - Arrays of any size and type
1432 - Aggregates containing an array of any size and type.
1433 - Calls to alloca().
1434 - Local variables that have had their address taken.
1436 Variables that are identified as requiring a protector will be arranged
1437 on the stack such that they are adjacent to the stack protector guard.
1438 The specific layout rules are:
1440 #. Large arrays and structures containing large arrays
1441 (``>= ssp-buffer-size``) are closest to the stack protector.
1442 #. Small arrays and structures containing small arrays
1443 (``< ssp-buffer-size``) are 2nd closest to the protector.
1444 #. Variables that have had their address taken are 3rd closest to the
1447 This overrides the ``ssp`` function attribute.
1449 If a function that has an ``sspstrong`` attribute is inlined into a
1450 function that doesn't have an ``sspstrong`` attribute, then the
1451 resulting function will have an ``sspstrong`` attribute.
1453 This attribute indicates that the function will delegate to some other
1454 function with a tail call. The prototype of a thunk should not be used for
1455 optimization purposes. The caller is expected to cast the thunk prototype to
1456 match the thunk target prototype.
1458 This attribute indicates that the ABI being targeted requires that
1459 an unwind table entry be produced for this function even if we can
1460 show that no exceptions passes by it. This is normally the case for
1461 the ELF x86-64 abi, but it can be disabled for some compilation
1470 Note: operand bundles are a work in progress, and they should be
1471 considered experimental at this time.
1473 Operand bundles are tagged sets of SSA values that can be associated
1474 with certain LLVM instructions (currently only ``call`` s and
1475 ``invoke`` s). In a way they are like metadata, but dropping them is
1476 incorrect and will change program semantics.
1480 operand bundle set ::= '[' operand bundle (, operand bundle )* ']'
1481 operand bundle ::= tag '(' [ bundle operand ] (, bundle operand )* ')'
1482 bundle operand ::= SSA value
1483 tag ::= string constant
1485 Operand bundles are **not** part of a function's signature, and a
1486 given function may be called from multiple places with different kinds
1487 of operand bundles. This reflects the fact that the operand bundles
1488 are conceptually a part of the ``call`` (or ``invoke``), not the
1489 callee being dispatched to.
1491 Operand bundles are a generic mechanism intended to support
1492 runtime-introspection-like functionality for managed languages. While
1493 the exact semantics of an operand bundle depend on the bundle tag,
1494 there are certain limitations to how much the presence of an operand
1495 bundle can influence the semantics of a program. These restrictions
1496 are described as the semantics of an "unknown" operand bundle. As
1497 long as the behavior of an operand bundle is describable within these
1498 restrictions, LLVM does not need to have special knowledge of the
1499 operand bundle to not miscompile programs containing it.
1501 - The bundle operands for an unknown operand bundle escape in unknown
1502 ways before control is transferred to the callee or invokee.
1503 - Calls and invokes with operand bundles have unknown read / write
1504 effect on the heap on entry and exit (even if the call target is
1505 ``readnone`` or ``readonly``), unless they're overriden with
1506 callsite specific attributes.
1507 - An operand bundle at a call site cannot change the implementation
1508 of the called function. Inter-procedural optimizations work as
1509 usual as long as they take into account the first two properties.
1511 More specific types of operand bundles are described below.
1513 Deoptimization Operand Bundles
1514 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1516 Deoptimization operand bundles are characterized by the ``"deopt"``
1517 operand bundle tag. These operand bundles represent an alternate
1518 "safe" continuation for the call site they're attached to, and can be
1519 used by a suitable runtime to deoptimize the compiled frame at the
1520 specified call site. There can be at most one ``"deopt"`` operand
1521 bundle attached to a call site. Exact details of deoptimization is
1522 out of scope for the language reference, but it usually involves
1523 rewriting a compiled frame into a set of interpreted frames.
1525 From the compiler's perspective, deoptimization operand bundles make
1526 the call sites they're attached to at least ``readonly``. They read
1527 through all of their pointer typed operands (even if they're not
1528 otherwise escaped) and the entire visible heap. Deoptimization
1529 operand bundles do not capture their operands except during
1530 deoptimization, in which case control will not be returned to the
1533 The inliner knows how to inline through calls that have deoptimization
1534 operand bundles. Just like inlining through a normal call site
1535 involves composing the normal and exceptional continuations, inlining
1536 through a call site with a deoptimization operand bundle needs to
1537 appropriately compose the "safe" deoptimization continuation. The
1538 inliner does this by prepending the parent's deoptimization
1539 continuation to every deoptimization continuation in the inlined body.
1540 E.g. inlining ``@f`` into ``@g`` in the following example
1542 .. code-block:: llvm
1545 call void @x() ;; no deopt state
1546 call void @y() [ "deopt"(i32 10) ]
1547 call void @y() [ "deopt"(i32 10), "unknown"(i8* null) ]
1552 call void @f() [ "deopt"(i32 20) ]
1558 .. code-block:: llvm
1561 call void @x() ;; still no deopt state
1562 call void @y() [ "deopt"(i32 20, i32 10) ]
1563 call void @y() [ "deopt"(i32 20, i32 10), "unknown"(i8* null) ]
1567 It is the frontend's responsibility to structure or encode the
1568 deoptimization state in a way that syntactically prepending the
1569 caller's deoptimization state to the callee's deoptimization state is
1570 semantically equivalent to composing the caller's deoptimization
1571 continuation after the callee's deoptimization continuation.
1575 Module-Level Inline Assembly
1576 ----------------------------
1578 Modules may contain "module-level inline asm" blocks, which corresponds
1579 to the GCC "file scope inline asm" blocks. These blocks are internally
1580 concatenated by LLVM and treated as a single unit, but may be separated
1581 in the ``.ll`` file if desired. The syntax is very simple:
1583 .. code-block:: llvm
1585 module asm "inline asm code goes here"
1586 module asm "more can go here"
1588 The strings can contain any character by escaping non-printable
1589 characters. The escape sequence used is simply "\\xx" where "xx" is the
1590 two digit hex code for the number.
1592 Note that the assembly string *must* be parseable by LLVM's integrated assembler
1593 (unless it is disabled), even when emitting a ``.s`` file.
1595 .. _langref_datalayout:
1600 A module may specify a target specific data layout string that specifies
1601 how data is to be laid out in memory. The syntax for the data layout is
1604 .. code-block:: llvm
1606 target datalayout = "layout specification"
1608 The *layout specification* consists of a list of specifications
1609 separated by the minus sign character ('-'). Each specification starts
1610 with a letter and may include other information after the letter to
1611 define some aspect of the data layout. The specifications accepted are
1615 Specifies that the target lays out data in big-endian form. That is,
1616 the bits with the most significance have the lowest address
1619 Specifies that the target lays out data in little-endian form. That
1620 is, the bits with the least significance have the lowest address
1623 Specifies the natural alignment of the stack in bits. Alignment
1624 promotion of stack variables is limited to the natural stack
1625 alignment to avoid dynamic stack realignment. The stack alignment
1626 must be a multiple of 8-bits. If omitted, the natural stack
1627 alignment defaults to "unspecified", which does not prevent any
1628 alignment promotions.
1629 ``p[n]:<size>:<abi>:<pref>``
1630 This specifies the *size* of a pointer and its ``<abi>`` and
1631 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1632 bits. The address space, ``n``, is optional, and if not specified,
1633 denotes the default address space 0. The value of ``n`` must be
1634 in the range [1,2^23).
1635 ``i<size>:<abi>:<pref>``
1636 This specifies the alignment for an integer type of a given bit
1637 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1638 ``v<size>:<abi>:<pref>``
1639 This specifies the alignment for a vector type of a given bit
1641 ``f<size>:<abi>:<pref>``
1642 This specifies the alignment for a floating point type of a given bit
1643 ``<size>``. Only values of ``<size>`` that are supported by the target
1644 will work. 32 (float) and 64 (double) are supported on all targets; 80
1645 or 128 (different flavors of long double) are also supported on some
1648 This specifies the alignment for an object of aggregate type.
1650 If present, specifies that llvm names are mangled in the output. The
1653 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1654 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1655 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1656 symbols get a ``_`` prefix.
1657 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1658 functions also get a suffix based on the frame size.
1659 * ``x``: Windows x86 COFF prefix: Similar to Windows COFF, but use a ``_``
1660 prefix for ``__cdecl`` functions.
1661 ``n<size1>:<size2>:<size3>...``
1662 This specifies a set of native integer widths for the target CPU in
1663 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1664 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1665 this set are considered to support most general arithmetic operations
1668 On every specification that takes a ``<abi>:<pref>``, specifying the
1669 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1670 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1672 When constructing the data layout for a given target, LLVM starts with a
1673 default set of specifications which are then (possibly) overridden by
1674 the specifications in the ``datalayout`` keyword. The default
1675 specifications are given in this list:
1677 - ``E`` - big endian
1678 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1679 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1680 same as the default address space.
1681 - ``S0`` - natural stack alignment is unspecified
1682 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1683 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1684 - ``i16:16:16`` - i16 is 16-bit aligned
1685 - ``i32:32:32`` - i32 is 32-bit aligned
1686 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1687 alignment of 64-bits
1688 - ``f16:16:16`` - half is 16-bit aligned
1689 - ``f32:32:32`` - float is 32-bit aligned
1690 - ``f64:64:64`` - double is 64-bit aligned
1691 - ``f128:128:128`` - quad is 128-bit aligned
1692 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1693 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1694 - ``a:0:64`` - aggregates are 64-bit aligned
1696 When LLVM is determining the alignment for a given type, it uses the
1699 #. If the type sought is an exact match for one of the specifications,
1700 that specification is used.
1701 #. If no match is found, and the type sought is an integer type, then
1702 the smallest integer type that is larger than the bitwidth of the
1703 sought type is used. If none of the specifications are larger than
1704 the bitwidth then the largest integer type is used. For example,
1705 given the default specifications above, the i7 type will use the
1706 alignment of i8 (next largest) while both i65 and i256 will use the
1707 alignment of i64 (largest specified).
1708 #. If no match is found, and the type sought is a vector type, then the
1709 largest vector type that is smaller than the sought vector type will
1710 be used as a fall back. This happens because <128 x double> can be
1711 implemented in terms of 64 <2 x double>, for example.
1713 The function of the data layout string may not be what you expect.
1714 Notably, this is not a specification from the frontend of what alignment
1715 the code generator should use.
1717 Instead, if specified, the target data layout is required to match what
1718 the ultimate *code generator* expects. This string is used by the
1719 mid-level optimizers to improve code, and this only works if it matches
1720 what the ultimate code generator uses. There is no way to generate IR
1721 that does not embed this target-specific detail into the IR. If you
1722 don't specify the string, the default specifications will be used to
1723 generate a Data Layout and the optimization phases will operate
1724 accordingly and introduce target specificity into the IR with respect to
1725 these default specifications.
1732 A module may specify a target triple string that describes the target
1733 host. The syntax for the target triple is simply:
1735 .. code-block:: llvm
1737 target triple = "x86_64-apple-macosx10.7.0"
1739 The *target triple* string consists of a series of identifiers delimited
1740 by the minus sign character ('-'). The canonical forms are:
1744 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1745 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1747 This information is passed along to the backend so that it generates
1748 code for the proper architecture. It's possible to override this on the
1749 command line with the ``-mtriple`` command line option.
1751 .. _pointeraliasing:
1753 Pointer Aliasing Rules
1754 ----------------------
1756 Any memory access must be done through a pointer value associated with
1757 an address range of the memory access, otherwise the behavior is
1758 undefined. Pointer values are associated with address ranges according
1759 to the following rules:
1761 - A pointer value is associated with the addresses associated with any
1762 value it is *based* on.
1763 - An address of a global variable is associated with the address range
1764 of the variable's storage.
1765 - The result value of an allocation instruction is associated with the
1766 address range of the allocated storage.
1767 - A null pointer in the default address-space is associated with no
1769 - An integer constant other than zero or a pointer value returned from
1770 a function not defined within LLVM may be associated with address
1771 ranges allocated through mechanisms other than those provided by
1772 LLVM. Such ranges shall not overlap with any ranges of addresses
1773 allocated by mechanisms provided by LLVM.
1775 A pointer value is *based* on another pointer value according to the
1778 - A pointer value formed from a ``getelementptr`` operation is *based*
1779 on the first value operand of the ``getelementptr``.
1780 - The result value of a ``bitcast`` is *based* on the operand of the
1782 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1783 values that contribute (directly or indirectly) to the computation of
1784 the pointer's value.
1785 - The "*based* on" relationship is transitive.
1787 Note that this definition of *"based"* is intentionally similar to the
1788 definition of *"based"* in C99, though it is slightly weaker.
1790 LLVM IR does not associate types with memory. The result type of a
1791 ``load`` merely indicates the size and alignment of the memory from
1792 which to load, as well as the interpretation of the value. The first
1793 operand type of a ``store`` similarly only indicates the size and
1794 alignment of the store.
1796 Consequently, type-based alias analysis, aka TBAA, aka
1797 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1798 :ref:`Metadata <metadata>` may be used to encode additional information
1799 which specialized optimization passes may use to implement type-based
1804 Volatile Memory Accesses
1805 ------------------------
1807 Certain memory accesses, such as :ref:`load <i_load>`'s,
1808 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1809 marked ``volatile``. The optimizers must not change the number of
1810 volatile operations or change their order of execution relative to other
1811 volatile operations. The optimizers *may* change the order of volatile
1812 operations relative to non-volatile operations. This is not Java's
1813 "volatile" and has no cross-thread synchronization behavior.
1815 IR-level volatile loads and stores cannot safely be optimized into
1816 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1817 flagged volatile. Likewise, the backend should never split or merge
1818 target-legal volatile load/store instructions.
1820 .. admonition:: Rationale
1822 Platforms may rely on volatile loads and stores of natively supported
1823 data width to be executed as single instruction. For example, in C
1824 this holds for an l-value of volatile primitive type with native
1825 hardware support, but not necessarily for aggregate types. The
1826 frontend upholds these expectations, which are intentionally
1827 unspecified in the IR. The rules above ensure that IR transformations
1828 do not violate the frontend's contract with the language.
1832 Memory Model for Concurrent Operations
1833 --------------------------------------
1835 The LLVM IR does not define any way to start parallel threads of
1836 execution or to register signal handlers. Nonetheless, there are
1837 platform-specific ways to create them, and we define LLVM IR's behavior
1838 in their presence. This model is inspired by the C++0x memory model.
1840 For a more informal introduction to this model, see the :doc:`Atomics`.
1842 We define a *happens-before* partial order as the least partial order
1845 - Is a superset of single-thread program order, and
1846 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1847 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1848 techniques, like pthread locks, thread creation, thread joining,
1849 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1850 Constraints <ordering>`).
1852 Note that program order does not introduce *happens-before* edges
1853 between a thread and signals executing inside that thread.
1855 Every (defined) read operation (load instructions, memcpy, atomic
1856 loads/read-modify-writes, etc.) R reads a series of bytes written by
1857 (defined) write operations (store instructions, atomic
1858 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1859 section, initialized globals are considered to have a write of the
1860 initializer which is atomic and happens before any other read or write
1861 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1862 may see any write to the same byte, except:
1864 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1865 write\ :sub:`2` happens before R\ :sub:`byte`, then
1866 R\ :sub:`byte` does not see write\ :sub:`1`.
1867 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1868 R\ :sub:`byte` does not see write\ :sub:`3`.
1870 Given that definition, R\ :sub:`byte` is defined as follows:
1872 - If R is volatile, the result is target-dependent. (Volatile is
1873 supposed to give guarantees which can support ``sig_atomic_t`` in
1874 C/C++, and may be used for accesses to addresses that do not behave
1875 like normal memory. It does not generally provide cross-thread
1877 - Otherwise, if there is no write to the same byte that happens before
1878 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1879 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1880 R\ :sub:`byte` returns the value written by that write.
1881 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1882 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1883 Memory Ordering Constraints <ordering>` section for additional
1884 constraints on how the choice is made.
1885 - Otherwise R\ :sub:`byte` returns ``undef``.
1887 R returns the value composed of the series of bytes it read. This
1888 implies that some bytes within the value may be ``undef`` **without**
1889 the entire value being ``undef``. Note that this only defines the
1890 semantics of the operation; it doesn't mean that targets will emit more
1891 than one instruction to read the series of bytes.
1893 Note that in cases where none of the atomic intrinsics are used, this
1894 model places only one restriction on IR transformations on top of what
1895 is required for single-threaded execution: introducing a store to a byte
1896 which might not otherwise be stored is not allowed in general.
1897 (Specifically, in the case where another thread might write to and read
1898 from an address, introducing a store can change a load that may see
1899 exactly one write into a load that may see multiple writes.)
1903 Atomic Memory Ordering Constraints
1904 ----------------------------------
1906 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1907 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1908 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1909 ordering parameters that determine which other atomic instructions on
1910 the same address they *synchronize with*. These semantics are borrowed
1911 from Java and C++0x, but are somewhat more colloquial. If these
1912 descriptions aren't precise enough, check those specs (see spec
1913 references in the :doc:`atomics guide <Atomics>`).
1914 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1915 differently since they don't take an address. See that instruction's
1916 documentation for details.
1918 For a simpler introduction to the ordering constraints, see the
1922 The set of values that can be read is governed by the happens-before
1923 partial order. A value cannot be read unless some operation wrote
1924 it. This is intended to provide a guarantee strong enough to model
1925 Java's non-volatile shared variables. This ordering cannot be
1926 specified for read-modify-write operations; it is not strong enough
1927 to make them atomic in any interesting way.
1929 In addition to the guarantees of ``unordered``, there is a single
1930 total order for modifications by ``monotonic`` operations on each
1931 address. All modification orders must be compatible with the
1932 happens-before order. There is no guarantee that the modification
1933 orders can be combined to a global total order for the whole program
1934 (and this often will not be possible). The read in an atomic
1935 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1936 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1937 order immediately before the value it writes. If one atomic read
1938 happens before another atomic read of the same address, the later
1939 read must see the same value or a later value in the address's
1940 modification order. This disallows reordering of ``monotonic`` (or
1941 stronger) operations on the same address. If an address is written
1942 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1943 read that address repeatedly, the other threads must eventually see
1944 the write. This corresponds to the C++0x/C1x
1945 ``memory_order_relaxed``.
1947 In addition to the guarantees of ``monotonic``, a
1948 *synchronizes-with* edge may be formed with a ``release`` operation.
1949 This is intended to model C++'s ``memory_order_acquire``.
1951 In addition to the guarantees of ``monotonic``, if this operation
1952 writes a value which is subsequently read by an ``acquire``
1953 operation, it *synchronizes-with* that operation. (This isn't a
1954 complete description; see the C++0x definition of a release
1955 sequence.) This corresponds to the C++0x/C1x
1956 ``memory_order_release``.
1957 ``acq_rel`` (acquire+release)
1958 Acts as both an ``acquire`` and ``release`` operation on its
1959 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1960 ``seq_cst`` (sequentially consistent)
1961 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1962 operation that only reads, ``release`` for an operation that only
1963 writes), there is a global total order on all
1964 sequentially-consistent operations on all addresses, which is
1965 consistent with the *happens-before* partial order and with the
1966 modification orders of all the affected addresses. Each
1967 sequentially-consistent read sees the last preceding write to the
1968 same address in this global order. This corresponds to the C++0x/C1x
1969 ``memory_order_seq_cst`` and Java volatile.
1973 If an atomic operation is marked ``singlethread``, it only *synchronizes
1974 with* or participates in modification and seq\_cst total orderings with
1975 other operations running in the same thread (for example, in signal
1983 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1984 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1985 :ref:`frem <i_frem>`, :ref:`fcmp <i_fcmp>`) have the following flags that can
1986 be set to enable otherwise unsafe floating point operations
1989 No NaNs - Allow optimizations to assume the arguments and result are not
1990 NaN. Such optimizations are required to retain defined behavior over
1991 NaNs, but the value of the result is undefined.
1994 No Infs - Allow optimizations to assume the arguments and result are not
1995 +/-Inf. Such optimizations are required to retain defined behavior over
1996 +/-Inf, but the value of the result is undefined.
1999 No Signed Zeros - Allow optimizations to treat the sign of a zero
2000 argument or result as insignificant.
2003 Allow Reciprocal - Allow optimizations to use the reciprocal of an
2004 argument rather than perform division.
2007 Fast - Allow algebraically equivalent transformations that may
2008 dramatically change results in floating point (e.g. reassociate). This
2009 flag implies all the others.
2013 Use-list Order Directives
2014 -------------------------
2016 Use-list directives encode the in-memory order of each use-list, allowing the
2017 order to be recreated. ``<order-indexes>`` is a comma-separated list of
2018 indexes that are assigned to the referenced value's uses. The referenced
2019 value's use-list is immediately sorted by these indexes.
2021 Use-list directives may appear at function scope or global scope. They are not
2022 instructions, and have no effect on the semantics of the IR. When they're at
2023 function scope, they must appear after the terminator of the final basic block.
2025 If basic blocks have their address taken via ``blockaddress()`` expressions,
2026 ``uselistorder_bb`` can be used to reorder their use-lists from outside their
2033 uselistorder <ty> <value>, { <order-indexes> }
2034 uselistorder_bb @function, %block { <order-indexes> }
2040 define void @foo(i32 %arg1, i32 %arg2) {
2042 ; ... instructions ...
2044 ; ... instructions ...
2046 ; At function scope.
2047 uselistorder i32 %arg1, { 1, 0, 2 }
2048 uselistorder label %bb, { 1, 0 }
2052 uselistorder i32* @global, { 1, 2, 0 }
2053 uselistorder i32 7, { 1, 0 }
2054 uselistorder i32 (i32) @bar, { 1, 0 }
2055 uselistorder_bb @foo, %bb, { 5, 1, 3, 2, 0, 4 }
2062 The LLVM type system is one of the most important features of the
2063 intermediate representation. Being typed enables a number of
2064 optimizations to be performed on the intermediate representation
2065 directly, without having to do extra analyses on the side before the
2066 transformation. A strong type system makes it easier to read the
2067 generated code and enables novel analyses and transformations that are
2068 not feasible to perform on normal three address code representations.
2078 The void type does not represent any value and has no size.
2096 The function type can be thought of as a function signature. It consists of a
2097 return type and a list of formal parameter types. The return type of a function
2098 type is a void type or first class type --- except for :ref:`label <t_label>`
2099 and :ref:`metadata <t_metadata>` types.
2105 <returntype> (<parameter list>)
2107 ...where '``<parameter list>``' is a comma-separated list of type
2108 specifiers. Optionally, the parameter list may include a type ``...``, which
2109 indicates that the function takes a variable number of arguments. Variable
2110 argument functions can access their arguments with the :ref:`variable argument
2111 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
2112 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
2116 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2117 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
2118 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2119 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
2120 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2121 | ``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. |
2122 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2123 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
2124 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2131 The :ref:`first class <t_firstclass>` types are perhaps the most important.
2132 Values of these types are the only ones which can be produced by
2140 These are the types that are valid in registers from CodeGen's perspective.
2149 The integer type is a very simple type that simply specifies an
2150 arbitrary bit width for the integer type desired. Any bit width from 1
2151 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
2159 The number of bits the integer will occupy is specified by the ``N``
2165 +----------------+------------------------------------------------+
2166 | ``i1`` | a single-bit integer. |
2167 +----------------+------------------------------------------------+
2168 | ``i32`` | a 32-bit integer. |
2169 +----------------+------------------------------------------------+
2170 | ``i1942652`` | a really big integer of over 1 million bits. |
2171 +----------------+------------------------------------------------+
2175 Floating Point Types
2176 """"""""""""""""""""
2185 - 16-bit floating point value
2188 - 32-bit floating point value
2191 - 64-bit floating point value
2194 - 128-bit floating point value (112-bit mantissa)
2197 - 80-bit floating point value (X87)
2200 - 128-bit floating point value (two 64-bits)
2207 The x86_mmx type represents a value held in an MMX register on an x86
2208 machine. The operations allowed on it are quite limited: parameters and
2209 return values, load and store, and bitcast. User-specified MMX
2210 instructions are represented as intrinsic or asm calls with arguments
2211 and/or results of this type. There are no arrays, vectors or constants
2228 The pointer type is used to specify memory locations. Pointers are
2229 commonly used to reference objects in memory.
2231 Pointer types may have an optional address space attribute defining the
2232 numbered address space where the pointed-to object resides. The default
2233 address space is number zero. The semantics of non-zero address spaces
2234 are target-specific.
2236 Note that LLVM does not permit pointers to void (``void*``) nor does it
2237 permit pointers to labels (``label*``). Use ``i8*`` instead.
2247 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2248 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
2249 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2250 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
2251 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2252 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
2253 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2262 A vector type is a simple derived type that represents a vector of
2263 elements. Vector types are used when multiple primitive data are
2264 operated in parallel using a single instruction (SIMD). A vector type
2265 requires a size (number of elements) and an underlying primitive data
2266 type. Vector types are considered :ref:`first class <t_firstclass>`.
2272 < <# elements> x <elementtype> >
2274 The number of elements is a constant integer value larger than 0;
2275 elementtype may be any integer, floating point or pointer type. Vectors
2276 of size zero are not allowed.
2280 +-------------------+--------------------------------------------------+
2281 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
2282 +-------------------+--------------------------------------------------+
2283 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
2284 +-------------------+--------------------------------------------------+
2285 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
2286 +-------------------+--------------------------------------------------+
2287 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
2288 +-------------------+--------------------------------------------------+
2297 The label type represents code labels.
2312 The token type is used when a value is associated with an instruction
2313 but all uses of the value must not attempt to introspect or obscure it.
2314 As such, it is not appropriate to have a :ref:`phi <i_phi>` or
2315 :ref:`select <i_select>` of type token.
2332 The metadata type represents embedded metadata. No derived types may be
2333 created from metadata except for :ref:`function <t_function>` arguments.
2346 Aggregate Types are a subset of derived types that can contain multiple
2347 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
2348 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
2358 The array type is a very simple derived type that arranges elements
2359 sequentially in memory. The array type requires a size (number of
2360 elements) and an underlying data type.
2366 [<# elements> x <elementtype>]
2368 The number of elements is a constant integer value; ``elementtype`` may
2369 be any type with a size.
2373 +------------------+--------------------------------------+
2374 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
2375 +------------------+--------------------------------------+
2376 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
2377 +------------------+--------------------------------------+
2378 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
2379 +------------------+--------------------------------------+
2381 Here are some examples of multidimensional arrays:
2383 +-----------------------------+----------------------------------------------------------+
2384 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
2385 +-----------------------------+----------------------------------------------------------+
2386 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
2387 +-----------------------------+----------------------------------------------------------+
2388 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
2389 +-----------------------------+----------------------------------------------------------+
2391 There is no restriction on indexing beyond the end of the array implied
2392 by a static type (though there are restrictions on indexing beyond the
2393 bounds of an allocated object in some cases). This means that
2394 single-dimension 'variable sized array' addressing can be implemented in
2395 LLVM with a zero length array type. An implementation of 'pascal style
2396 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
2406 The structure type is used to represent a collection of data members
2407 together in memory. The elements of a structure may be any type that has
2410 Structures in memory are accessed using '``load``' and '``store``' by
2411 getting a pointer to a field with the '``getelementptr``' instruction.
2412 Structures in registers are accessed using the '``extractvalue``' and
2413 '``insertvalue``' instructions.
2415 Structures may optionally be "packed" structures, which indicate that
2416 the alignment of the struct is one byte, and that there is no padding
2417 between the elements. In non-packed structs, padding between field types
2418 is inserted as defined by the DataLayout string in the module, which is
2419 required to match what the underlying code generator expects.
2421 Structures can either be "literal" or "identified". A literal structure
2422 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
2423 identified types are always defined at the top level with a name.
2424 Literal types are uniqued by their contents and can never be recursive
2425 or opaque since there is no way to write one. Identified types can be
2426 recursive, can be opaqued, and are never uniqued.
2432 %T1 = type { <type list> } ; Identified normal struct type
2433 %T2 = type <{ <type list> }> ; Identified packed struct type
2437 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2438 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
2439 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2440 | ``{ 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``. |
2441 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2442 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
2443 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2447 Opaque Structure Types
2448 """"""""""""""""""""""
2452 Opaque structure types are used to represent named structure types that
2453 do not have a body specified. This corresponds (for example) to the C
2454 notion of a forward declared structure.
2465 +--------------+-------------------+
2466 | ``opaque`` | An opaque type. |
2467 +--------------+-------------------+
2474 LLVM has several different basic types of constants. This section
2475 describes them all and their syntax.
2480 **Boolean constants**
2481 The two strings '``true``' and '``false``' are both valid constants
2483 **Integer constants**
2484 Standard integers (such as '4') are constants of the
2485 :ref:`integer <t_integer>` type. Negative numbers may be used with
2487 **Floating point constants**
2488 Floating point constants use standard decimal notation (e.g.
2489 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2490 hexadecimal notation (see below). The assembler requires the exact
2491 decimal value of a floating-point constant. For example, the
2492 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2493 decimal in binary. Floating point constants must have a :ref:`floating
2494 point <t_floating>` type.
2495 **Null pointer constants**
2496 The identifier '``null``' is recognized as a null pointer constant
2497 and must be of :ref:`pointer type <t_pointer>`.
2499 The identifier '``none``' is recognized as an empty token constant
2500 and must be of :ref:`token type <t_token>`.
2502 The one non-intuitive notation for constants is the hexadecimal form of
2503 floating point constants. For example, the form
2504 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2505 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2506 constants are required (and the only time that they are generated by the
2507 disassembler) is when a floating point constant must be emitted but it
2508 cannot be represented as a decimal floating point number in a reasonable
2509 number of digits. For example, NaN's, infinities, and other special
2510 values are represented in their IEEE hexadecimal format so that assembly
2511 and disassembly do not cause any bits to change in the constants.
2513 When using the hexadecimal form, constants of types half, float, and
2514 double are represented using the 16-digit form shown above (which
2515 matches the IEEE754 representation for double); half and float values
2516 must, however, be exactly representable as IEEE 754 half and single
2517 precision, respectively. Hexadecimal format is always used for long
2518 double, and there are three forms of long double. The 80-bit format used
2519 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2520 128-bit format used by PowerPC (two adjacent doubles) is represented by
2521 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2522 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2523 will only work if they match the long double format on your target.
2524 The IEEE 16-bit format (half precision) is represented by ``0xH``
2525 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2526 (sign bit at the left).
2528 There are no constants of type x86_mmx.
2530 .. _complexconstants:
2535 Complex constants are a (potentially recursive) combination of simple
2536 constants and smaller complex constants.
2538 **Structure constants**
2539 Structure constants are represented with notation similar to
2540 structure type definitions (a comma separated list of elements,
2541 surrounded by braces (``{}``)). For example:
2542 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2543 "``@G = external global i32``". Structure constants must have
2544 :ref:`structure type <t_struct>`, and the number and types of elements
2545 must match those specified by the type.
2547 Array constants are represented with notation similar to array type
2548 definitions (a comma separated list of elements, surrounded by
2549 square brackets (``[]``)). For example:
2550 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2551 :ref:`array type <t_array>`, and the number and types of elements must
2552 match those specified by the type. As a special case, character array
2553 constants may also be represented as a double-quoted string using the ``c``
2554 prefix. For example: "``c"Hello World\0A\00"``".
2555 **Vector constants**
2556 Vector constants are represented with notation similar to vector
2557 type definitions (a comma separated list of elements, surrounded by
2558 less-than/greater-than's (``<>``)). For example:
2559 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2560 must have :ref:`vector type <t_vector>`, and the number and types of
2561 elements must match those specified by the type.
2562 **Zero initialization**
2563 The string '``zeroinitializer``' can be used to zero initialize a
2564 value to zero of *any* type, including scalar and
2565 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2566 having to print large zero initializers (e.g. for large arrays) and
2567 is always exactly equivalent to using explicit zero initializers.
2569 A metadata node is a constant tuple without types. For example:
2570 "``!{!0, !{!2, !0}, !"test"}``". Metadata can reference constant values,
2571 for example: "``!{!0, i32 0, i8* @global, i64 (i64)* @function, !"str"}``".
2572 Unlike other typed constants that are meant to be interpreted as part of
2573 the instruction stream, metadata is a place to attach additional
2574 information such as debug info.
2576 Global Variable and Function Addresses
2577 --------------------------------------
2579 The addresses of :ref:`global variables <globalvars>` and
2580 :ref:`functions <functionstructure>` are always implicitly valid
2581 (link-time) constants. These constants are explicitly referenced when
2582 the :ref:`identifier for the global <identifiers>` is used and always have
2583 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2586 .. code-block:: llvm
2590 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2597 The string '``undef``' can be used anywhere a constant is expected, and
2598 indicates that the user of the value may receive an unspecified
2599 bit-pattern. Undefined values may be of any type (other than '``label``'
2600 or '``void``') and be used anywhere a constant is permitted.
2602 Undefined values are useful because they indicate to the compiler that
2603 the program is well defined no matter what value is used. This gives the
2604 compiler more freedom to optimize. Here are some examples of
2605 (potentially surprising) transformations that are valid (in pseudo IR):
2607 .. code-block:: llvm
2617 This is safe because all of the output bits are affected by the undef
2618 bits. Any output bit can have a zero or one depending on the input bits.
2620 .. code-block:: llvm
2631 These logical operations have bits that are not always affected by the
2632 input. For example, if ``%X`` has a zero bit, then the output of the
2633 '``and``' operation will always be a zero for that bit, no matter what
2634 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2635 optimize or assume that the result of the '``and``' is '``undef``'.
2636 However, it is safe to assume that all bits of the '``undef``' could be
2637 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2638 all the bits of the '``undef``' operand to the '``or``' could be set,
2639 allowing the '``or``' to be folded to -1.
2641 .. code-block:: llvm
2643 %A = select undef, %X, %Y
2644 %B = select undef, 42, %Y
2645 %C = select %X, %Y, undef
2655 This set of examples shows that undefined '``select``' (and conditional
2656 branch) conditions can go *either way*, but they have to come from one
2657 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2658 both known to have a clear low bit, then ``%A`` would have to have a
2659 cleared low bit. However, in the ``%C`` example, the optimizer is
2660 allowed to assume that the '``undef``' operand could be the same as
2661 ``%Y``, allowing the whole '``select``' to be eliminated.
2663 .. code-block:: llvm
2665 %A = xor undef, undef
2682 This example points out that two '``undef``' operands are not
2683 necessarily the same. This can be surprising to people (and also matches
2684 C semantics) where they assume that "``X^X``" is always zero, even if
2685 ``X`` is undefined. This isn't true for a number of reasons, but the
2686 short answer is that an '``undef``' "variable" can arbitrarily change
2687 its value over its "live range". This is true because the variable
2688 doesn't actually *have a live range*. Instead, the value is logically
2689 read from arbitrary registers that happen to be around when needed, so
2690 the value is not necessarily consistent over time. In fact, ``%A`` and
2691 ``%C`` need to have the same semantics or the core LLVM "replace all
2692 uses with" concept would not hold.
2694 .. code-block:: llvm
2702 These examples show the crucial difference between an *undefined value*
2703 and *undefined behavior*. An undefined value (like '``undef``') is
2704 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2705 operation can be constant folded to '``undef``', because the '``undef``'
2706 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2707 However, in the second example, we can make a more aggressive
2708 assumption: because the ``undef`` is allowed to be an arbitrary value,
2709 we are allowed to assume that it could be zero. Since a divide by zero
2710 has *undefined behavior*, we are allowed to assume that the operation
2711 does not execute at all. This allows us to delete the divide and all
2712 code after it. Because the undefined operation "can't happen", the
2713 optimizer can assume that it occurs in dead code.
2715 .. code-block:: llvm
2717 a: store undef -> %X
2718 b: store %X -> undef
2723 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2724 value can be assumed to not have any effect; we can assume that the
2725 value is overwritten with bits that happen to match what was already
2726 there. However, a store *to* an undefined location could clobber
2727 arbitrary memory, therefore, it has undefined behavior.
2734 Poison values are similar to :ref:`undef values <undefvalues>`, however
2735 they also represent the fact that an instruction or constant expression
2736 that cannot evoke side effects has nevertheless detected a condition
2737 that results in undefined behavior.
2739 There is currently no way of representing a poison value in the IR; they
2740 only exist when produced by operations such as :ref:`add <i_add>` with
2743 Poison value behavior is defined in terms of value *dependence*:
2745 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2746 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2747 their dynamic predecessor basic block.
2748 - Function arguments depend on the corresponding actual argument values
2749 in the dynamic callers of their functions.
2750 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2751 instructions that dynamically transfer control back to them.
2752 - :ref:`Invoke <i_invoke>` instructions depend on the
2753 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2754 call instructions that dynamically transfer control back to them.
2755 - Non-volatile loads and stores depend on the most recent stores to all
2756 of the referenced memory addresses, following the order in the IR
2757 (including loads and stores implied by intrinsics such as
2758 :ref:`@llvm.memcpy <int_memcpy>`.)
2759 - An instruction with externally visible side effects depends on the
2760 most recent preceding instruction with externally visible side
2761 effects, following the order in the IR. (This includes :ref:`volatile
2762 operations <volatile>`.)
2763 - An instruction *control-depends* on a :ref:`terminator
2764 instruction <terminators>` if the terminator instruction has
2765 multiple successors and the instruction is always executed when
2766 control transfers to one of the successors, and may not be executed
2767 when control is transferred to another.
2768 - Additionally, an instruction also *control-depends* on a terminator
2769 instruction if the set of instructions it otherwise depends on would
2770 be different if the terminator had transferred control to a different
2772 - Dependence is transitive.
2774 Poison values have the same behavior as :ref:`undef values <undefvalues>`,
2775 with the additional effect that any instruction that has a *dependence*
2776 on a poison value has undefined behavior.
2778 Here are some examples:
2780 .. code-block:: llvm
2783 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2784 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2785 %poison_yet_again = getelementptr i32, i32* @h, i32 %still_poison
2786 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2788 store i32 %poison, i32* @g ; Poison value stored to memory.
2789 %poison2 = load i32, i32* @g ; Poison value loaded back from memory.
2791 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2793 %narrowaddr = bitcast i32* @g to i16*
2794 %wideaddr = bitcast i32* @g to i64*
2795 %poison3 = load i16, i16* %narrowaddr ; Returns a poison value.
2796 %poison4 = load i64, i64* %wideaddr ; Returns a poison value.
2798 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2799 br i1 %cmp, label %true, label %end ; Branch to either destination.
2802 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2803 ; it has undefined behavior.
2807 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2808 ; Both edges into this PHI are
2809 ; control-dependent on %cmp, so this
2810 ; always results in a poison value.
2812 store volatile i32 0, i32* @g ; This would depend on the store in %true
2813 ; if %cmp is true, or the store in %entry
2814 ; otherwise, so this is undefined behavior.
2816 br i1 %cmp, label %second_true, label %second_end
2817 ; The same branch again, but this time the
2818 ; true block doesn't have side effects.
2825 store volatile i32 0, i32* @g ; This time, the instruction always depends
2826 ; on the store in %end. Also, it is
2827 ; control-equivalent to %end, so this is
2828 ; well-defined (ignoring earlier undefined
2829 ; behavior in this example).
2833 Addresses of Basic Blocks
2834 -------------------------
2836 ``blockaddress(@function, %block)``
2838 The '``blockaddress``' constant computes the address of the specified
2839 basic block in the specified function, and always has an ``i8*`` type.
2840 Taking the address of the entry block is illegal.
2842 This value only has defined behavior when used as an operand to the
2843 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2844 against null. Pointer equality tests between labels addresses results in
2845 undefined behavior --- though, again, comparison against null is ok, and
2846 no label is equal to the null pointer. This may be passed around as an
2847 opaque pointer sized value as long as the bits are not inspected. This
2848 allows ``ptrtoint`` and arithmetic to be performed on these values so
2849 long as the original value is reconstituted before the ``indirectbr``
2852 Finally, some targets may provide defined semantics when using the value
2853 as the operand to an inline assembly, but that is target specific.
2857 Constant Expressions
2858 --------------------
2860 Constant expressions are used to allow expressions involving other
2861 constants to be used as constants. Constant expressions may be of any
2862 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2863 that does not have side effects (e.g. load and call are not supported).
2864 The following is the syntax for constant expressions:
2866 ``trunc (CST to TYPE)``
2867 Truncate a constant to another type. The bit size of CST must be
2868 larger than the bit size of TYPE. Both types must be integers.
2869 ``zext (CST to TYPE)``
2870 Zero extend a constant to another type. The bit size of CST must be
2871 smaller than the bit size of TYPE. Both types must be integers.
2872 ``sext (CST to TYPE)``
2873 Sign extend a constant to another type. The bit size of CST must be
2874 smaller than the bit size of TYPE. Both types must be integers.
2875 ``fptrunc (CST to TYPE)``
2876 Truncate a floating point constant to another floating point type.
2877 The size of CST must be larger than the size of TYPE. Both types
2878 must be floating point.
2879 ``fpext (CST to TYPE)``
2880 Floating point extend a constant to another type. The size of CST
2881 must be smaller or equal to the size of TYPE. Both types must be
2883 ``fptoui (CST to TYPE)``
2884 Convert a floating point constant to the corresponding unsigned
2885 integer constant. TYPE must be a scalar or vector integer type. CST
2886 must be of scalar or vector floating point type. Both CST and TYPE
2887 must be scalars, or vectors of the same number of elements. If the
2888 value won't fit in the integer type, the results are undefined.
2889 ``fptosi (CST to TYPE)``
2890 Convert a floating point constant to the corresponding signed
2891 integer constant. TYPE must be a scalar or vector integer type. CST
2892 must be of scalar or vector floating point type. Both CST and TYPE
2893 must be scalars, or vectors of the same number of elements. If the
2894 value won't fit in the integer type, the results are undefined.
2895 ``uitofp (CST to TYPE)``
2896 Convert an unsigned integer constant to the corresponding floating
2897 point constant. TYPE must be a scalar or vector floating point type.
2898 CST must be of scalar or vector integer type. Both CST and TYPE must
2899 be scalars, or vectors of the same number of elements. If the value
2900 won't fit in the floating point type, the results are undefined.
2901 ``sitofp (CST to TYPE)``
2902 Convert a signed integer constant to the corresponding floating
2903 point constant. TYPE must be a scalar or vector floating point type.
2904 CST must be of scalar or vector integer type. Both CST and TYPE must
2905 be scalars, or vectors of the same number of elements. If the value
2906 won't fit in the floating point type, the results are undefined.
2907 ``ptrtoint (CST to TYPE)``
2908 Convert a pointer typed constant to the corresponding integer
2909 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2910 pointer type. The ``CST`` value is zero extended, truncated, or
2911 unchanged to make it fit in ``TYPE``.
2912 ``inttoptr (CST to TYPE)``
2913 Convert an integer constant to a pointer constant. TYPE must be a
2914 pointer type. CST must be of integer type. The CST value is zero
2915 extended, truncated, or unchanged to make it fit in a pointer size.
2916 This one is *really* dangerous!
2917 ``bitcast (CST to TYPE)``
2918 Convert a constant, CST, to another TYPE. The constraints of the
2919 operands are the same as those for the :ref:`bitcast
2920 instruction <i_bitcast>`.
2921 ``addrspacecast (CST to TYPE)``
2922 Convert a constant pointer or constant vector of pointer, CST, to another
2923 TYPE in a different address space. The constraints of the operands are the
2924 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2925 ``getelementptr (TY, CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (TY, CSTPTR, IDX0, IDX1, ...)``
2926 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2927 constants. As with the :ref:`getelementptr <i_getelementptr>`
2928 instruction, the index list may have zero or more indexes, which are
2929 required to make sense for the type of "pointer to TY".
2930 ``select (COND, VAL1, VAL2)``
2931 Perform the :ref:`select operation <i_select>` on constants.
2932 ``icmp COND (VAL1, VAL2)``
2933 Performs the :ref:`icmp operation <i_icmp>` on constants.
2934 ``fcmp COND (VAL1, VAL2)``
2935 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2936 ``extractelement (VAL, IDX)``
2937 Perform the :ref:`extractelement operation <i_extractelement>` on
2939 ``insertelement (VAL, ELT, IDX)``
2940 Perform the :ref:`insertelement operation <i_insertelement>` on
2942 ``shufflevector (VEC1, VEC2, IDXMASK)``
2943 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2945 ``extractvalue (VAL, IDX0, IDX1, ...)``
2946 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2947 constants. The index list is interpreted in a similar manner as
2948 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2949 least one index value must be specified.
2950 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2951 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2952 The index list is interpreted in a similar manner as indices in a
2953 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2954 value must be specified.
2955 ``OPCODE (LHS, RHS)``
2956 Perform the specified operation of the LHS and RHS constants. OPCODE
2957 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2958 binary <bitwiseops>` operations. The constraints on operands are
2959 the same as those for the corresponding instruction (e.g. no bitwise
2960 operations on floating point values are allowed).
2967 Inline Assembler Expressions
2968 ----------------------------
2970 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2971 Inline Assembly <moduleasm>`) through the use of a special value. This value
2972 represents the inline assembler as a template string (containing the
2973 instructions to emit), a list of operand constraints (stored as a string), a
2974 flag that indicates whether or not the inline asm expression has side effects,
2975 and a flag indicating whether the function containing the asm needs to align its
2976 stack conservatively.
2978 The template string supports argument substitution of the operands using "``$``"
2979 followed by a number, to indicate substitution of the given register/memory
2980 location, as specified by the constraint string. "``${NUM:MODIFIER}``" may also
2981 be used, where ``MODIFIER`` is a target-specific annotation for how to print the
2982 operand (See :ref:`inline-asm-modifiers`).
2984 A literal "``$``" may be included by using "``$$``" in the template. To include
2985 other special characters into the output, the usual "``\XX``" escapes may be
2986 used, just as in other strings. Note that after template substitution, the
2987 resulting assembly string is parsed by LLVM's integrated assembler unless it is
2988 disabled -- even when emitting a ``.s`` file -- and thus must contain assembly
2989 syntax known to LLVM.
2991 LLVM's support for inline asm is modeled closely on the requirements of Clang's
2992 GCC-compatible inline-asm support. Thus, the feature-set and the constraint and
2993 modifier codes listed here are similar or identical to those in GCC's inline asm
2994 support. However, to be clear, the syntax of the template and constraint strings
2995 described here is *not* the same as the syntax accepted by GCC and Clang, and,
2996 while most constraint letters are passed through as-is by Clang, some get
2997 translated to other codes when converting from the C source to the LLVM
3000 An example inline assembler expression is:
3002 .. code-block:: llvm
3004 i32 (i32) asm "bswap $0", "=r,r"
3006 Inline assembler expressions may **only** be used as the callee operand
3007 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
3008 Thus, typically we have:
3010 .. code-block:: llvm
3012 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
3014 Inline asms with side effects not visible in the constraint list must be
3015 marked as having side effects. This is done through the use of the
3016 '``sideeffect``' keyword, like so:
3018 .. code-block:: llvm
3020 call void asm sideeffect "eieio", ""()
3022 In some cases inline asms will contain code that will not work unless
3023 the stack is aligned in some way, such as calls or SSE instructions on
3024 x86, yet will not contain code that does that alignment within the asm.
3025 The compiler should make conservative assumptions about what the asm
3026 might contain and should generate its usual stack alignment code in the
3027 prologue if the '``alignstack``' keyword is present:
3029 .. code-block:: llvm
3031 call void asm alignstack "eieio", ""()
3033 Inline asms also support using non-standard assembly dialects. The
3034 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
3035 the inline asm is using the Intel dialect. Currently, ATT and Intel are
3036 the only supported dialects. An example is:
3038 .. code-block:: llvm
3040 call void asm inteldialect "eieio", ""()
3042 If multiple keywords appear the '``sideeffect``' keyword must come
3043 first, the '``alignstack``' keyword second and the '``inteldialect``'
3046 Inline Asm Constraint String
3047 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3049 The constraint list is a comma-separated string, each element containing one or
3050 more constraint codes.
3052 For each element in the constraint list an appropriate register or memory
3053 operand will be chosen, and it will be made available to assembly template
3054 string expansion as ``$0`` for the first constraint in the list, ``$1`` for the
3057 There are three different types of constraints, which are distinguished by a
3058 prefix symbol in front of the constraint code: Output, Input, and Clobber. The
3059 constraints must always be given in that order: outputs first, then inputs, then
3060 clobbers. They cannot be intermingled.
3062 There are also three different categories of constraint codes:
3064 - Register constraint. This is either a register class, or a fixed physical
3065 register. This kind of constraint will allocate a register, and if necessary,
3066 bitcast the argument or result to the appropriate type.
3067 - Memory constraint. This kind of constraint is for use with an instruction
3068 taking a memory operand. Different constraints allow for different addressing
3069 modes used by the target.
3070 - Immediate value constraint. This kind of constraint is for an integer or other
3071 immediate value which can be rendered directly into an instruction. The
3072 various target-specific constraints allow the selection of a value in the
3073 proper range for the instruction you wish to use it with.
3078 Output constraints are specified by an "``=``" prefix (e.g. "``=r``"). This
3079 indicates that the assembly will write to this operand, and the operand will
3080 then be made available as a return value of the ``asm`` expression. Output
3081 constraints do not consume an argument from the call instruction. (Except, see
3082 below about indirect outputs).
3084 Normally, it is expected that no output locations are written to by the assembly
3085 expression until *all* of the inputs have been read. As such, LLVM may assign
3086 the same register to an output and an input. If this is not safe (e.g. if the
3087 assembly contains two instructions, where the first writes to one output, and
3088 the second reads an input and writes to a second output), then the "``&``"
3089 modifier must be used (e.g. "``=&r``") to specify that the output is an
3090 "early-clobber" output. Marking an ouput as "early-clobber" ensures that LLVM
3091 will not use the same register for any inputs (other than an input tied to this
3097 Input constraints do not have a prefix -- just the constraint codes. Each input
3098 constraint will consume one argument from the call instruction. It is not
3099 permitted for the asm to write to any input register or memory location (unless
3100 that input is tied to an output). Note also that multiple inputs may all be
3101 assigned to the same register, if LLVM can determine that they necessarily all
3102 contain the same value.
3104 Instead of providing a Constraint Code, input constraints may also "tie"
3105 themselves to an output constraint, by providing an integer as the constraint
3106 string. Tied inputs still consume an argument from the call instruction, and
3107 take up a position in the asm template numbering as is usual -- they will simply
3108 be constrained to always use the same register as the output they've been tied
3109 to. For example, a constraint string of "``=r,0``" says to assign a register for
3110 output, and use that register as an input as well (it being the 0'th
3113 It is permitted to tie an input to an "early-clobber" output. In that case, no
3114 *other* input may share the same register as the input tied to the early-clobber
3115 (even when the other input has the same value).
3117 You may only tie an input to an output which has a register constraint, not a
3118 memory constraint. Only a single input may be tied to an output.
3120 There is also an "interesting" feature which deserves a bit of explanation: if a
3121 register class constraint allocates a register which is too small for the value
3122 type operand provided as input, the input value will be split into multiple
3123 registers, and all of them passed to the inline asm.
3125 However, this feature is often not as useful as you might think.
3127 Firstly, the registers are *not* guaranteed to be consecutive. So, on those
3128 architectures that have instructions which operate on multiple consecutive
3129 instructions, this is not an appropriate way to support them. (e.g. the 32-bit
3130 SparcV8 has a 64-bit load, which instruction takes a single 32-bit register. The
3131 hardware then loads into both the named register, and the next register. This
3132 feature of inline asm would not be useful to support that.)
3134 A few of the targets provide a template string modifier allowing explicit access
3135 to the second register of a two-register operand (e.g. MIPS ``L``, ``M``, and
3136 ``D``). On such an architecture, you can actually access the second allocated
3137 register (yet, still, not any subsequent ones). But, in that case, you're still
3138 probably better off simply splitting the value into two separate operands, for
3139 clarity. (e.g. see the description of the ``A`` constraint on X86, which,
3140 despite existing only for use with this feature, is not really a good idea to
3143 Indirect inputs and outputs
3144 """""""""""""""""""""""""""
3146 Indirect output or input constraints can be specified by the "``*``" modifier
3147 (which goes after the "``=``" in case of an output). This indicates that the asm
3148 will write to or read from the contents of an *address* provided as an input
3149 argument. (Note that in this way, indirect outputs act more like an *input* than
3150 an output: just like an input, they consume an argument of the call expression,
3151 rather than producing a return value. An indirect output constraint is an
3152 "output" only in that the asm is expected to write to the contents of the input
3153 memory location, instead of just read from it).
3155 This is most typically used for memory constraint, e.g. "``=*m``", to pass the
3156 address of a variable as a value.
3158 It is also possible to use an indirect *register* constraint, but only on output
3159 (e.g. "``=*r``"). This will cause LLVM to allocate a register for an output
3160 value normally, and then, separately emit a store to the address provided as
3161 input, after the provided inline asm. (It's not clear what value this
3162 functionality provides, compared to writing the store explicitly after the asm
3163 statement, and it can only produce worse code, since it bypasses many
3164 optimization passes. I would recommend not using it.)
3170 A clobber constraint is indicated by a "``~``" prefix. A clobber does not
3171 consume an input operand, nor generate an output. Clobbers cannot use any of the
3172 general constraint code letters -- they may use only explicit register
3173 constraints, e.g. "``~{eax}``". The one exception is that a clobber string of
3174 "``~{memory}``" indicates that the assembly writes to arbitrary undeclared
3175 memory locations -- not only the memory pointed to by a declared indirect
3181 After a potential prefix comes constraint code, or codes.
3183 A Constraint Code is either a single letter (e.g. "``r``"), a "``^``" character
3184 followed by two letters (e.g. "``^wc``"), or "``{``" register-name "``}``"
3187 The one and two letter constraint codes are typically chosen to be the same as
3188 GCC's constraint codes.
3190 A single constraint may include one or more than constraint code in it, leaving
3191 it up to LLVM to choose which one to use. This is included mainly for
3192 compatibility with the translation of GCC inline asm coming from clang.
3194 There are two ways to specify alternatives, and either or both may be used in an
3195 inline asm constraint list:
3197 1) Append the codes to each other, making a constraint code set. E.g. "``im``"
3198 or "``{eax}m``". This means "choose any of the options in the set". The
3199 choice of constraint is made independently for each constraint in the
3202 2) Use "``|``" between constraint code sets, creating alternatives. Every
3203 constraint in the constraint list must have the same number of alternative
3204 sets. With this syntax, the same alternative in *all* of the items in the
3205 constraint list will be chosen together.
3207 Putting those together, you might have a two operand constraint string like
3208 ``"rm|r,ri|rm"``. This indicates that if operand 0 is ``r`` or ``m``, then
3209 operand 1 may be one of ``r`` or ``i``. If operand 0 is ``r``, then operand 1
3210 may be one of ``r`` or ``m``. But, operand 0 and 1 cannot both be of type m.
3212 However, the use of either of the alternatives features is *NOT* recommended, as
3213 LLVM is not able to make an intelligent choice about which one to use. (At the
3214 point it currently needs to choose, not enough information is available to do so
3215 in a smart way.) Thus, it simply tries to make a choice that's most likely to
3216 compile, not one that will be optimal performance. (e.g., given "``rm``", it'll
3217 always choose to use memory, not registers). And, if given multiple registers,
3218 or multiple register classes, it will simply choose the first one. (In fact, it
3219 doesn't currently even ensure explicitly specified physical registers are
3220 unique, so specifying multiple physical registers as alternatives, like
3221 ``{r11}{r12},{r11}{r12}``, will assign r11 to both operands, not at all what was
3224 Supported Constraint Code List
3225 """"""""""""""""""""""""""""""
3227 The constraint codes are, in general, expected to behave the same way they do in
3228 GCC. LLVM's support is often implemented on an 'as-needed' basis, to support C
3229 inline asm code which was supported by GCC. A mismatch in behavior between LLVM
3230 and GCC likely indicates a bug in LLVM.
3232 Some constraint codes are typically supported by all targets:
3234 - ``r``: A register in the target's general purpose register class.
3235 - ``m``: A memory address operand. It is target-specific what addressing modes
3236 are supported, typical examples are register, or register + register offset,
3237 or register + immediate offset (of some target-specific size).
3238 - ``i``: An integer constant (of target-specific width). Allows either a simple
3239 immediate, or a relocatable value.
3240 - ``n``: An integer constant -- *not* including relocatable values.
3241 - ``s``: An integer constant, but allowing *only* relocatable values.
3242 - ``X``: Allows an operand of any kind, no constraint whatsoever. Typically
3243 useful to pass a label for an asm branch or call.
3245 .. FIXME: but that surely isn't actually okay to jump out of an asm
3246 block without telling llvm about the control transfer???)
3248 - ``{register-name}``: Requires exactly the named physical register.
3250 Other constraints are target-specific:
3254 - ``z``: An immediate integer 0. Outputs ``WZR`` or ``XZR``, as appropriate.
3255 - ``I``: An immediate integer valid for an ``ADD`` or ``SUB`` instruction,
3256 i.e. 0 to 4095 with optional shift by 12.
3257 - ``J``: An immediate integer that, when negated, is valid for an ``ADD`` or
3258 ``SUB`` instruction, i.e. -1 to -4095 with optional left shift by 12.
3259 - ``K``: An immediate integer that is valid for the 'bitmask immediate 32' of a
3260 logical instruction like ``AND``, ``EOR``, or ``ORR`` with a 32-bit register.
3261 - ``L``: An immediate integer that is valid for the 'bitmask immediate 64' of a
3262 logical instruction like ``AND``, ``EOR``, or ``ORR`` with a 64-bit register.
3263 - ``M``: An immediate integer for use with the ``MOV`` assembly alias on a
3264 32-bit register. This is a superset of ``K``: in addition to the bitmask
3265 immediate, also allows immediate integers which can be loaded with a single
3266 ``MOVZ`` or ``MOVL`` instruction.
3267 - ``N``: An immediate integer for use with the ``MOV`` assembly alias on a
3268 64-bit register. This is a superset of ``L``.
3269 - ``Q``: Memory address operand must be in a single register (no
3270 offsets). (However, LLVM currently does this for the ``m`` constraint as
3272 - ``r``: A 32 or 64-bit integer register (W* or X*).
3273 - ``w``: A 32, 64, or 128-bit floating-point/SIMD register.
3274 - ``x``: A lower 128-bit floating-point/SIMD register (``V0`` to ``V15``).
3278 - ``r``: A 32 or 64-bit integer register.
3279 - ``[0-9]v``: The 32-bit VGPR register, number 0-9.
3280 - ``[0-9]s``: The 32-bit SGPR register, number 0-9.
3285 - ``Q``, ``Um``, ``Un``, ``Uq``, ``Us``, ``Ut``, ``Uv``, ``Uy``: Memory address
3286 operand. Treated the same as operand ``m``, at the moment.
3288 ARM and ARM's Thumb2 mode:
3290 - ``j``: An immediate integer between 0 and 65535 (valid for ``MOVW``)
3291 - ``I``: An immediate integer valid for a data-processing instruction.
3292 - ``J``: An immediate integer between -4095 and 4095.
3293 - ``K``: An immediate integer whose bitwise inverse is valid for a
3294 data-processing instruction. (Can be used with template modifier "``B``" to
3295 print the inverted value).
3296 - ``L``: An immediate integer whose negation is valid for a data-processing
3297 instruction. (Can be used with template modifier "``n``" to print the negated
3299 - ``M``: A power of two or a integer between 0 and 32.
3300 - ``N``: Invalid immediate constraint.
3301 - ``O``: Invalid immediate constraint.
3302 - ``r``: A general-purpose 32-bit integer register (``r0-r15``).
3303 - ``l``: In Thumb2 mode, low 32-bit GPR registers (``r0-r7``). In ARM mode, same
3305 - ``h``: In Thumb2 mode, a high 32-bit GPR register (``r8-r15``). In ARM mode,
3307 - ``w``: A 32, 64, or 128-bit floating-point/SIMD register: ``s0-s31``,
3308 ``d0-d31``, or ``q0-q15``.
3309 - ``x``: A 32, 64, or 128-bit floating-point/SIMD register: ``s0-s15``,
3310 ``d0-d7``, or ``q0-q3``.
3311 - ``t``: A floating-point/SIMD register, only supports 32-bit values:
3316 - ``I``: An immediate integer between 0 and 255.
3317 - ``J``: An immediate integer between -255 and -1.
3318 - ``K``: An immediate integer between 0 and 255, with optional left-shift by
3320 - ``L``: An immediate integer between -7 and 7.
3321 - ``M``: An immediate integer which is a multiple of 4 between 0 and 1020.
3322 - ``N``: An immediate integer between 0 and 31.
3323 - ``O``: An immediate integer which is a multiple of 4 between -508 and 508.
3324 - ``r``: A low 32-bit GPR register (``r0-r7``).
3325 - ``l``: A low 32-bit GPR register (``r0-r7``).
3326 - ``h``: A high GPR register (``r0-r7``).
3327 - ``w``: A 32, 64, or 128-bit floating-point/SIMD register: ``s0-s31``,
3328 ``d0-d31``, or ``q0-q15``.
3329 - ``x``: A 32, 64, or 128-bit floating-point/SIMD register: ``s0-s15``,
3330 ``d0-d7``, or ``q0-q3``.
3331 - ``t``: A floating-point/SIMD register, only supports 32-bit values:
3337 - ``o``, ``v``: A memory address operand, treated the same as constraint ``m``,
3339 - ``r``: A 32 or 64-bit register.
3343 - ``r``: An 8 or 16-bit register.
3347 - ``I``: An immediate signed 16-bit integer.
3348 - ``J``: An immediate integer zero.
3349 - ``K``: An immediate unsigned 16-bit integer.
3350 - ``L``: An immediate 32-bit integer, where the lower 16 bits are 0.
3351 - ``N``: An immediate integer between -65535 and -1.
3352 - ``O``: An immediate signed 15-bit integer.
3353 - ``P``: An immediate integer between 1 and 65535.
3354 - ``m``: A memory address operand. In MIPS-SE mode, allows a base address
3355 register plus 16-bit immediate offset. In MIPS mode, just a base register.
3356 - ``R``: A memory address operand. In MIPS-SE mode, allows a base address
3357 register plus a 9-bit signed offset. In MIPS mode, the same as constraint
3359 - ``ZC``: A memory address operand, suitable for use in a ``pref``, ``ll``, or
3360 ``sc`` instruction on the given subtarget (details vary).
3361 - ``r``, ``d``, ``y``: A 32 or 64-bit GPR register.
3362 - ``f``: A 32 or 64-bit FPU register (``F0-F31``), or a 128-bit MSA register
3363 (``W0-W31``). In the case of MSA registers, it is recommended to use the ``w``
3364 argument modifier for compatibility with GCC.
3365 - ``c``: A 32-bit or 64-bit GPR register suitable for indirect jump (always
3367 - ``l``: The ``lo`` register, 32 or 64-bit.
3372 - ``b``: A 1-bit integer register.
3373 - ``c`` or ``h``: A 16-bit integer register.
3374 - ``r``: A 32-bit integer register.
3375 - ``l`` or ``N``: A 64-bit integer register.
3376 - ``f``: A 32-bit float register.
3377 - ``d``: A 64-bit float register.
3382 - ``I``: An immediate signed 16-bit integer.
3383 - ``J``: An immediate unsigned 16-bit integer, shifted left 16 bits.
3384 - ``K``: An immediate unsigned 16-bit integer.
3385 - ``L``: An immediate signed 16-bit integer, shifted left 16 bits.
3386 - ``M``: An immediate integer greater than 31.
3387 - ``N``: An immediate integer that is an exact power of 2.
3388 - ``O``: The immediate integer constant 0.
3389 - ``P``: An immediate integer constant whose negation is a signed 16-bit
3391 - ``es``, ``o``, ``Q``, ``Z``, ``Zy``: A memory address operand, currently
3392 treated the same as ``m``.
3393 - ``r``: A 32 or 64-bit integer register.
3394 - ``b``: A 32 or 64-bit integer register, excluding ``R0`` (that is:
3396 - ``f``: A 32 or 64-bit float register (``F0-F31``), or when QPX is enabled, a
3397 128 or 256-bit QPX register (``Q0-Q31``; aliases the ``F`` registers).
3398 - ``v``: For ``4 x f32`` or ``4 x f64`` types, when QPX is enabled, a
3399 128 or 256-bit QPX register (``Q0-Q31``), otherwise a 128-bit
3400 altivec vector register (``V0-V31``).
3402 .. FIXME: is this a bug that v accepts QPX registers? I think this
3403 is supposed to only use the altivec vector registers?
3405 - ``y``: Condition register (``CR0-CR7``).
3406 - ``wc``: An individual CR bit in a CR register.
3407 - ``wa``, ``wd``, ``wf``: Any 128-bit VSX vector register, from the full VSX
3408 register set (overlapping both the floating-point and vector register files).
3409 - ``ws``: A 32 or 64-bit floating point register, from the full VSX register
3414 - ``I``: An immediate 13-bit signed integer.
3415 - ``r``: A 32-bit integer register.
3419 - ``I``: An immediate unsigned 8-bit integer.
3420 - ``J``: An immediate unsigned 12-bit integer.
3421 - ``K``: An immediate signed 16-bit integer.
3422 - ``L``: An immediate signed 20-bit integer.
3423 - ``M``: An immediate integer 0x7fffffff.
3424 - ``Q``, ``R``, ``S``, ``T``: A memory address operand, treated the same as
3425 ``m``, at the moment.
3426 - ``r`` or ``d``: A 32, 64, or 128-bit integer register.
3427 - ``a``: A 32, 64, or 128-bit integer address register (excludes R0, which in an
3428 address context evaluates as zero).
3429 - ``h``: A 32-bit value in the high part of a 64bit data register
3431 - ``f``: A 32, 64, or 128-bit floating point register.
3435 - ``I``: An immediate integer between 0 and 31.
3436 - ``J``: An immediate integer between 0 and 64.
3437 - ``K``: An immediate signed 8-bit integer.
3438 - ``L``: An immediate integer, 0xff or 0xffff or (in 64-bit mode only)
3440 - ``M``: An immediate integer between 0 and 3.
3441 - ``N``: An immediate unsigned 8-bit integer.
3442 - ``O``: An immediate integer between 0 and 127.
3443 - ``e``: An immediate 32-bit signed integer.
3444 - ``Z``: An immediate 32-bit unsigned integer.
3445 - ``o``, ``v``: Treated the same as ``m``, at the moment.
3446 - ``q``: An 8, 16, 32, or 64-bit register which can be accessed as an 8-bit
3447 ``l`` integer register. On X86-32, this is the ``a``, ``b``, ``c``, and ``d``
3448 registers, and on X86-64, it is all of the integer registers.
3449 - ``Q``: An 8, 16, 32, or 64-bit register which can be accessed as an 8-bit
3450 ``h`` integer register. This is the ``a``, ``b``, ``c``, and ``d`` registers.
3451 - ``r`` or ``l``: An 8, 16, 32, or 64-bit integer register.
3452 - ``R``: An 8, 16, 32, or 64-bit "legacy" integer register -- one which has
3453 existed since i386, and can be accessed without the REX prefix.
3454 - ``f``: A 32, 64, or 80-bit '387 FPU stack pseudo-register.
3455 - ``y``: A 64-bit MMX register, if MMX is enabled.
3456 - ``x``: If SSE is enabled: a 32 or 64-bit scalar operand, or 128-bit vector
3457 operand in a SSE register. If AVX is also enabled, can also be a 256-bit
3458 vector operand in an AVX register. If AVX-512 is also enabled, can also be a
3459 512-bit vector operand in an AVX512 register, Otherwise, an error.
3460 - ``Y``: The same as ``x``, if *SSE2* is enabled, otherwise an error.
3461 - ``A``: Special case: allocates EAX first, then EDX, for a single operand (in
3462 32-bit mode, a 64-bit integer operand will get split into two registers). It
3463 is not recommended to use this constraint, as in 64-bit mode, the 64-bit
3464 operand will get allocated only to RAX -- if two 32-bit operands are needed,
3465 you're better off splitting it yourself, before passing it to the asm
3470 - ``r``: A 32-bit integer register.
3473 .. _inline-asm-modifiers:
3475 Asm template argument modifiers
3476 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3478 In the asm template string, modifiers can be used on the operand reference, like
3481 The modifiers are, in general, expected to behave the same way they do in
3482 GCC. LLVM's support is often implemented on an 'as-needed' basis, to support C
3483 inline asm code which was supported by GCC. A mismatch in behavior between LLVM
3484 and GCC likely indicates a bug in LLVM.
3488 - ``c``: Print an immediate integer constant unadorned, without
3489 the target-specific immediate punctuation (e.g. no ``$`` prefix).
3490 - ``n``: Negate and print immediate integer constant unadorned, without the
3491 target-specific immediate punctuation (e.g. no ``$`` prefix).
3492 - ``l``: Print as an unadorned label, without the target-specific label
3493 punctuation (e.g. no ``$`` prefix).
3497 - ``w``: Print a GPR register with a ``w*`` name instead of ``x*`` name. E.g.,
3498 instead of ``x30``, print ``w30``.
3499 - ``x``: Print a GPR register with a ``x*`` name. (this is the default, anyhow).
3500 - ``b``, ``h``, ``s``, ``d``, ``q``: Print a floating-point/SIMD register with a
3501 ``b*``, ``h*``, ``s*``, ``d*``, or ``q*`` name, rather than the default of
3510 - ``a``: Print an operand as an address (with ``[`` and ``]`` surrounding a
3514 - ``y``: Print a VFP single-precision register as an indexed double (e.g. print
3515 as ``d4[1]`` instead of ``s9``)
3516 - ``B``: Bitwise invert and print an immediate integer constant without ``#``
3518 - ``L``: Print the low 16-bits of an immediate integer constant.
3519 - ``M``: Print as a register set suitable for ldm/stm. Also prints *all*
3520 register operands subsequent to the specified one (!), so use carefully.
3521 - ``Q``: Print the low-order register of a register-pair, or the low-order
3522 register of a two-register operand.
3523 - ``R``: Print the high-order register of a register-pair, or the high-order
3524 register of a two-register operand.
3525 - ``H``: Print the second register of a register-pair. (On a big-endian system,
3526 ``H`` is equivalent to ``Q``, and on little-endian system, ``H`` is equivalent
3529 .. FIXME: H doesn't currently support printing the second register
3530 of a two-register operand.
3532 - ``e``: Print the low doubleword register of a NEON quad register.
3533 - ``f``: Print the high doubleword register of a NEON quad register.
3534 - ``m``: Print the base register of a memory operand without the ``[`` and ``]``
3539 - ``L``: Print the second register of a two-register operand. Requires that it
3540 has been allocated consecutively to the first.
3542 .. FIXME: why is it restricted to consecutive ones? And there's
3543 nothing that ensures that happens, is there?
3545 - ``I``: Print the letter 'i' if the operand is an integer constant, otherwise
3546 nothing. Used to print 'addi' vs 'add' instructions.
3550 No additional modifiers.
3554 - ``X``: Print an immediate integer as hexadecimal
3555 - ``x``: Print the low 16 bits of an immediate integer as hexadecimal.
3556 - ``d``: Print an immediate integer as decimal.
3557 - ``m``: Subtract one and print an immediate integer as decimal.
3558 - ``z``: Print $0 if an immediate zero, otherwise print normally.
3559 - ``L``: Print the low-order register of a two-register operand, or prints the
3560 address of the low-order word of a double-word memory operand.
3562 .. FIXME: L seems to be missing memory operand support.
3564 - ``M``: Print the high-order register of a two-register operand, or prints the
3565 address of the high-order word of a double-word memory operand.
3567 .. FIXME: M seems to be missing memory operand support.
3569 - ``D``: Print the second register of a two-register operand, or prints the
3570 second word of a double-word memory operand. (On a big-endian system, ``D`` is
3571 equivalent to ``L``, and on little-endian system, ``D`` is equivalent to
3573 - ``w``: No effect. Provided for compatibility with GCC which requires this
3574 modifier in order to print MSA registers (``W0-W31``) with the ``f``
3583 - ``L``: Print the second register of a two-register operand. Requires that it
3584 has been allocated consecutively to the first.
3586 .. FIXME: why is it restricted to consecutive ones? And there's
3587 nothing that ensures that happens, is there?
3589 - ``I``: Print the letter 'i' if the operand is an integer constant, otherwise
3590 nothing. Used to print 'addi' vs 'add' instructions.
3591 - ``y``: For a memory operand, prints formatter for a two-register X-form
3592 instruction. (Currently always prints ``r0,OPERAND``).
3593 - ``U``: Prints 'u' if the memory operand is an update form, and nothing
3594 otherwise. (NOTE: LLVM does not support update form, so this will currently
3595 always print nothing)
3596 - ``X``: Prints 'x' if the memory operand is an indexed form. (NOTE: LLVM does
3597 not support indexed form, so this will currently always print nothing)
3605 SystemZ implements only ``n``, and does *not* support any of the other
3606 target-independent modifiers.
3610 - ``c``: Print an unadorned integer or symbol name. (The latter is
3611 target-specific behavior for this typically target-independent modifier).
3612 - ``A``: Print a register name with a '``*``' before it.
3613 - ``b``: Print an 8-bit register name (e.g. ``al``); do nothing on a memory
3615 - ``h``: Print the upper 8-bit register name (e.g. ``ah``); do nothing on a
3617 - ``w``: Print the 16-bit register name (e.g. ``ax``); do nothing on a memory
3619 - ``k``: Print the 32-bit register name (e.g. ``eax``); do nothing on a memory
3621 - ``q``: Print the 64-bit register name (e.g. ``rax``), if 64-bit registers are
3622 available, otherwise the 32-bit register name; do nothing on a memory operand.
3623 - ``n``: Negate and print an unadorned integer, or, for operands other than an
3624 immediate integer (e.g. a relocatable symbol expression), print a '-' before
3625 the operand. (The behavior for relocatable symbol expressions is a
3626 target-specific behavior for this typically target-independent modifier)
3627 - ``H``: Print a memory reference with additional offset +8.
3628 - ``P``: Print a memory reference or operand for use as the argument of a call
3629 instruction. (E.g. omit ``(rip)``, even though it's PC-relative.)
3633 No additional modifiers.
3639 The call instructions that wrap inline asm nodes may have a
3640 "``!srcloc``" MDNode attached to it that contains a list of constant
3641 integers. If present, the code generator will use the integer as the
3642 location cookie value when report errors through the ``LLVMContext``
3643 error reporting mechanisms. This allows a front-end to correlate backend
3644 errors that occur with inline asm back to the source code that produced
3647 .. code-block:: llvm
3649 call void asm sideeffect "something bad", ""(), !srcloc !42
3651 !42 = !{ i32 1234567 }
3653 It is up to the front-end to make sense of the magic numbers it places
3654 in the IR. If the MDNode contains multiple constants, the code generator
3655 will use the one that corresponds to the line of the asm that the error
3663 LLVM IR allows metadata to be attached to instructions in the program
3664 that can convey extra information about the code to the optimizers and
3665 code generator. One example application of metadata is source-level
3666 debug information. There are two metadata primitives: strings and nodes.
3668 Metadata does not have a type, and is not a value. If referenced from a
3669 ``call`` instruction, it uses the ``metadata`` type.
3671 All metadata are identified in syntax by a exclamation point ('``!``').
3673 .. _metadata-string:
3675 Metadata Nodes and Metadata Strings
3676 -----------------------------------
3678 A metadata string is a string surrounded by double quotes. It can
3679 contain any character by escaping non-printable characters with
3680 "``\xx``" where "``xx``" is the two digit hex code. For example:
3683 Metadata nodes are represented with notation similar to structure
3684 constants (a comma separated list of elements, surrounded by braces and
3685 preceded by an exclamation point). Metadata nodes can have any values as
3686 their operand. For example:
3688 .. code-block:: llvm
3690 !{ !"test\00", i32 10}
3692 Metadata nodes that aren't uniqued use the ``distinct`` keyword. For example:
3694 .. code-block:: llvm
3696 !0 = distinct !{!"test\00", i32 10}
3698 ``distinct`` nodes are useful when nodes shouldn't be merged based on their
3699 content. They can also occur when transformations cause uniquing collisions
3700 when metadata operands change.
3702 A :ref:`named metadata <namedmetadatastructure>` is a collection of
3703 metadata nodes, which can be looked up in the module symbol table. For
3706 .. code-block:: llvm
3710 Metadata can be used as function arguments. Here ``llvm.dbg.value``
3711 function is using two metadata arguments:
3713 .. code-block:: llvm
3715 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
3717 Metadata can be attached to an instruction. Here metadata ``!21`` is attached
3718 to the ``add`` instruction using the ``!dbg`` identifier:
3720 .. code-block:: llvm
3722 %indvar.next = add i64 %indvar, 1, !dbg !21
3724 Metadata can also be attached to a function definition. Here metadata ``!22``
3725 is attached to the ``foo`` function using the ``!dbg`` identifier:
3727 .. code-block:: llvm
3729 define void @foo() !dbg !22 {
3733 More information about specific metadata nodes recognized by the
3734 optimizers and code generator is found below.
3736 .. _specialized-metadata:
3738 Specialized Metadata Nodes
3739 ^^^^^^^^^^^^^^^^^^^^^^^^^^
3741 Specialized metadata nodes are custom data structures in metadata (as opposed
3742 to generic tuples). Their fields are labelled, and can be specified in any
3745 These aren't inherently debug info centric, but currently all the specialized
3746 metadata nodes are related to debug info.
3753 ``DICompileUnit`` nodes represent a compile unit. The ``enums:``,
3754 ``retainedTypes:``, ``subprograms:``, ``globals:`` and ``imports:`` fields are
3755 tuples containing the debug info to be emitted along with the compile unit,
3756 regardless of code optimizations (some nodes are only emitted if there are
3757 references to them from instructions).
3759 .. code-block:: llvm
3761 !0 = !DICompileUnit(language: DW_LANG_C99, file: !1, producer: "clang",
3762 isOptimized: true, flags: "-O2", runtimeVersion: 2,
3763 splitDebugFilename: "abc.debug", emissionKind: 1,
3764 enums: !2, retainedTypes: !3, subprograms: !4,
3765 globals: !5, imports: !6)
3767 Compile unit descriptors provide the root scope for objects declared in a
3768 specific compilation unit. File descriptors are defined using this scope.
3769 These descriptors are collected by a named metadata ``!llvm.dbg.cu``. They
3770 keep track of subprograms, global variables, type information, and imported
3771 entities (declarations and namespaces).
3778 ``DIFile`` nodes represent files. The ``filename:`` can include slashes.
3780 .. code-block:: llvm
3782 !0 = !DIFile(filename: "path/to/file", directory: "/path/to/dir")
3784 Files are sometimes used in ``scope:`` fields, and are the only valid target
3785 for ``file:`` fields.
3792 ``DIBasicType`` nodes represent primitive types, such as ``int``, ``bool`` and
3793 ``float``. ``tag:`` defaults to ``DW_TAG_base_type``.
3795 .. code-block:: llvm
3797 !0 = !DIBasicType(name: "unsigned char", size: 8, align: 8,
3798 encoding: DW_ATE_unsigned_char)
3799 !1 = !DIBasicType(tag: DW_TAG_unspecified_type, name: "decltype(nullptr)")
3801 The ``encoding:`` describes the details of the type. Usually it's one of the
3804 .. code-block:: llvm
3810 DW_ATE_signed_char = 6
3812 DW_ATE_unsigned_char = 8
3814 .. _DISubroutineType:
3819 ``DISubroutineType`` nodes represent subroutine types. Their ``types:`` field
3820 refers to a tuple; the first operand is the return type, while the rest are the
3821 types of the formal arguments in order. If the first operand is ``null``, that
3822 represents a function with no return value (such as ``void foo() {}`` in C++).
3824 .. code-block:: llvm
3826 !0 = !BasicType(name: "int", size: 32, align: 32, DW_ATE_signed)
3827 !1 = !BasicType(name: "char", size: 8, align: 8, DW_ATE_signed_char)
3828 !2 = !DISubroutineType(types: !{null, !0, !1}) ; void (int, char)
3835 ``DIDerivedType`` nodes represent types derived from other types, such as
3838 .. code-block:: llvm
3840 !0 = !DIBasicType(name: "unsigned char", size: 8, align: 8,
3841 encoding: DW_ATE_unsigned_char)
3842 !1 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !0, size: 32,
3845 The following ``tag:`` values are valid:
3847 .. code-block:: llvm
3849 DW_TAG_formal_parameter = 5
3851 DW_TAG_pointer_type = 15
3852 DW_TAG_reference_type = 16
3854 DW_TAG_ptr_to_member_type = 31
3855 DW_TAG_const_type = 38
3856 DW_TAG_volatile_type = 53
3857 DW_TAG_restrict_type = 55
3859 ``DW_TAG_member`` is used to define a member of a :ref:`composite type
3860 <DICompositeType>` or :ref:`subprogram <DISubprogram>`. The type of the member
3861 is the ``baseType:``. The ``offset:`` is the member's bit offset.
3862 ``DW_TAG_formal_parameter`` is used to define a member which is a formal
3863 argument of a subprogram.
3865 ``DW_TAG_typedef`` is used to provide a name for the ``baseType:``.
3867 ``DW_TAG_pointer_type``, ``DW_TAG_reference_type``, ``DW_TAG_const_type``,
3868 ``DW_TAG_volatile_type`` and ``DW_TAG_restrict_type`` are used to qualify the
3871 Note that the ``void *`` type is expressed as a type derived from NULL.
3873 .. _DICompositeType:
3878 ``DICompositeType`` nodes represent types composed of other types, like
3879 structures and unions. ``elements:`` points to a tuple of the composed types.
3881 If the source language supports ODR, the ``identifier:`` field gives the unique
3882 identifier used for type merging between modules. When specified, other types
3883 can refer to composite types indirectly via a :ref:`metadata string
3884 <metadata-string>` that matches their identifier.
3886 .. code-block:: llvm
3888 !0 = !DIEnumerator(name: "SixKind", value: 7)
3889 !1 = !DIEnumerator(name: "SevenKind", value: 7)
3890 !2 = !DIEnumerator(name: "NegEightKind", value: -8)
3891 !3 = !DICompositeType(tag: DW_TAG_enumeration_type, name: "Enum", file: !12,
3892 line: 2, size: 32, align: 32, identifier: "_M4Enum",
3893 elements: !{!0, !1, !2})
3895 The following ``tag:`` values are valid:
3897 .. code-block:: llvm
3899 DW_TAG_array_type = 1
3900 DW_TAG_class_type = 2
3901 DW_TAG_enumeration_type = 4
3902 DW_TAG_structure_type = 19
3903 DW_TAG_union_type = 23
3904 DW_TAG_subroutine_type = 21
3905 DW_TAG_inheritance = 28
3908 For ``DW_TAG_array_type``, the ``elements:`` should be :ref:`subrange
3909 descriptors <DISubrange>`, each representing the range of subscripts at that
3910 level of indexing. The ``DIFlagVector`` flag to ``flags:`` indicates that an
3911 array type is a native packed vector.
3913 For ``DW_TAG_enumeration_type``, the ``elements:`` should be :ref:`enumerator
3914 descriptors <DIEnumerator>`, each representing the definition of an enumeration
3915 value for the set. All enumeration type descriptors are collected in the
3916 ``enums:`` field of the :ref:`compile unit <DICompileUnit>`.
3918 For ``DW_TAG_structure_type``, ``DW_TAG_class_type``, and
3919 ``DW_TAG_union_type``, the ``elements:`` should be :ref:`derived types
3920 <DIDerivedType>` with ``tag: DW_TAG_member`` or ``tag: DW_TAG_inheritance``.
3927 ``DISubrange`` nodes are the elements for ``DW_TAG_array_type`` variants of
3928 :ref:`DICompositeType`. ``count: -1`` indicates an empty array.
3930 .. code-block:: llvm
3932 !0 = !DISubrange(count: 5, lowerBound: 0) ; array counting from 0
3933 !1 = !DISubrange(count: 5, lowerBound: 1) ; array counting from 1
3934 !2 = !DISubrange(count: -1) ; empty array.
3941 ``DIEnumerator`` nodes are the elements for ``DW_TAG_enumeration_type``
3942 variants of :ref:`DICompositeType`.
3944 .. code-block:: llvm
3946 !0 = !DIEnumerator(name: "SixKind", value: 7)
3947 !1 = !DIEnumerator(name: "SevenKind", value: 7)
3948 !2 = !DIEnumerator(name: "NegEightKind", value: -8)
3950 DITemplateTypeParameter
3951 """""""""""""""""""""""
3953 ``DITemplateTypeParameter`` nodes represent type parameters to generic source
3954 language constructs. They are used (optionally) in :ref:`DICompositeType` and
3955 :ref:`DISubprogram` ``templateParams:`` fields.
3957 .. code-block:: llvm
3959 !0 = !DITemplateTypeParameter(name: "Ty", type: !1)
3961 DITemplateValueParameter
3962 """"""""""""""""""""""""
3964 ``DITemplateValueParameter`` nodes represent value parameters to generic source
3965 language constructs. ``tag:`` defaults to ``DW_TAG_template_value_parameter``,
3966 but if specified can also be set to ``DW_TAG_GNU_template_template_param`` or
3967 ``DW_TAG_GNU_template_param_pack``. They are used (optionally) in
3968 :ref:`DICompositeType` and :ref:`DISubprogram` ``templateParams:`` fields.
3970 .. code-block:: llvm
3972 !0 = !DITemplateValueParameter(name: "Ty", type: !1, value: i32 7)
3977 ``DINamespace`` nodes represent namespaces in the source language.
3979 .. code-block:: llvm
3981 !0 = !DINamespace(name: "myawesomeproject", scope: !1, file: !2, line: 7)
3986 ``DIGlobalVariable`` nodes represent global variables in the source language.
3988 .. code-block:: llvm
3990 !0 = !DIGlobalVariable(name: "foo", linkageName: "foo", scope: !1,
3991 file: !2, line: 7, type: !3, isLocal: true,
3992 isDefinition: false, variable: i32* @foo,
3995 All global variables should be referenced by the `globals:` field of a
3996 :ref:`compile unit <DICompileUnit>`.
4003 ``DISubprogram`` nodes represent functions from the source language. A
4004 ``DISubprogram`` may be attached to a function definition using ``!dbg``
4005 metadata. The ``variables:`` field points at :ref:`variables <DILocalVariable>`
4006 that must be retained, even if their IR counterparts are optimized out of
4007 the IR. The ``type:`` field must point at an :ref:`DISubroutineType`.
4009 .. code-block:: llvm
4011 define void @_Z3foov() !dbg !0 {
4015 !0 = distinct !DISubprogram(name: "foo", linkageName: "_Zfoov", scope: !1,
4016 file: !2, line: 7, type: !3, isLocal: true,
4017 isDefinition: false, scopeLine: 8,
4019 virtuality: DW_VIRTUALITY_pure_virtual,
4020 virtualIndex: 10, flags: DIFlagPrototyped,
4021 isOptimized: true, templateParams: !5,
4022 declaration: !6, variables: !7)
4029 ``DILexicalBlock`` nodes describe nested blocks within a :ref:`subprogram
4030 <DISubprogram>`. The line number and column numbers are used to distinguish
4031 two lexical blocks at same depth. They are valid targets for ``scope:``
4034 .. code-block:: llvm
4036 !0 = distinct !DILexicalBlock(scope: !1, file: !2, line: 7, column: 35)
4038 Usually lexical blocks are ``distinct`` to prevent node merging based on
4041 .. _DILexicalBlockFile:
4046 ``DILexicalBlockFile`` nodes are used to discriminate between sections of a
4047 :ref:`lexical block <DILexicalBlock>`. The ``file:`` field can be changed to
4048 indicate textual inclusion, or the ``discriminator:`` field can be used to
4049 discriminate between control flow within a single block in the source language.
4051 .. code-block:: llvm
4053 !0 = !DILexicalBlock(scope: !3, file: !4, line: 7, column: 35)
4054 !1 = !DILexicalBlockFile(scope: !0, file: !4, discriminator: 0)
4055 !2 = !DILexicalBlockFile(scope: !0, file: !4, discriminator: 1)
4062 ``DILocation`` nodes represent source debug locations. The ``scope:`` field is
4063 mandatory, and points at an :ref:`DILexicalBlockFile`, an
4064 :ref:`DILexicalBlock`, or an :ref:`DISubprogram`.
4066 .. code-block:: llvm
4068 !0 = !DILocation(line: 2900, column: 42, scope: !1, inlinedAt: !2)
4070 .. _DILocalVariable:
4075 ``DILocalVariable`` nodes represent local variables in the source language. If
4076 the ``arg:`` field is set to non-zero, then this variable is a subprogram
4077 parameter, and it will be included in the ``variables:`` field of its
4078 :ref:`DISubprogram`.
4080 .. code-block:: llvm
4082 !0 = !DILocalVariable(name: "this", arg: 1, scope: !3, file: !2, line: 7,
4083 type: !3, flags: DIFlagArtificial)
4084 !1 = !DILocalVariable(name: "x", arg: 2, scope: !4, file: !2, line: 7,
4086 !2 = !DILocalVariable(name: "y", scope: !5, file: !2, line: 7, type: !3)
4091 ``DIExpression`` nodes represent DWARF expression sequences. They are used in
4092 :ref:`debug intrinsics<dbg_intrinsics>` (such as ``llvm.dbg.declare``) to
4093 describe how the referenced LLVM variable relates to the source language
4096 The current supported vocabulary is limited:
4098 - ``DW_OP_deref`` dereferences the working expression.
4099 - ``DW_OP_plus, 93`` adds ``93`` to the working expression.
4100 - ``DW_OP_bit_piece, 16, 8`` specifies the offset and size (``16`` and ``8``
4101 here, respectively) of the variable piece from the working expression.
4103 .. code-block:: llvm
4105 !0 = !DIExpression(DW_OP_deref)
4106 !1 = !DIExpression(DW_OP_plus, 3)
4107 !2 = !DIExpression(DW_OP_bit_piece, 3, 7)
4108 !3 = !DIExpression(DW_OP_deref, DW_OP_plus, 3, DW_OP_bit_piece, 3, 7)
4113 ``DIObjCProperty`` nodes represent Objective-C property nodes.
4115 .. code-block:: llvm
4117 !3 = !DIObjCProperty(name: "foo", file: !1, line: 7, setter: "setFoo",
4118 getter: "getFoo", attributes: 7, type: !2)
4123 ``DIImportedEntity`` nodes represent entities (such as modules) imported into a
4126 .. code-block:: llvm
4128 !2 = !DIImportedEntity(tag: DW_TAG_imported_module, name: "foo", scope: !0,
4129 entity: !1, line: 7)
4134 In LLVM IR, memory does not have types, so LLVM's own type system is not
4135 suitable for doing TBAA. Instead, metadata is added to the IR to
4136 describe a type system of a higher level language. This can be used to
4137 implement typical C/C++ TBAA, but it can also be used to implement
4138 custom alias analysis behavior for other languages.
4140 The current metadata format is very simple. TBAA metadata nodes have up
4141 to three fields, e.g.:
4143 .. code-block:: llvm
4145 !0 = !{ !"an example type tree" }
4146 !1 = !{ !"int", !0 }
4147 !2 = !{ !"float", !0 }
4148 !3 = !{ !"const float", !2, i64 1 }
4150 The first field is an identity field. It can be any value, usually a
4151 metadata string, which uniquely identifies the type. The most important
4152 name in the tree is the name of the root node. Two trees with different
4153 root node names are entirely disjoint, even if they have leaves with
4156 The second field identifies the type's parent node in the tree, or is
4157 null or omitted for a root node. A type is considered to alias all of
4158 its descendants and all of its ancestors in the tree. Also, a type is
4159 considered to alias all types in other trees, so that bitcode produced
4160 from multiple front-ends is handled conservatively.
4162 If the third field is present, it's an integer which if equal to 1
4163 indicates that the type is "constant" (meaning
4164 ``pointsToConstantMemory`` should return true; see `other useful
4165 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
4167 '``tbaa.struct``' Metadata
4168 ^^^^^^^^^^^^^^^^^^^^^^^^^^
4170 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
4171 aggregate assignment operations in C and similar languages, however it
4172 is defined to copy a contiguous region of memory, which is more than
4173 strictly necessary for aggregate types which contain holes due to
4174 padding. Also, it doesn't contain any TBAA information about the fields
4177 ``!tbaa.struct`` metadata can describe which memory subregions in a
4178 memcpy are padding and what the TBAA tags of the struct are.
4180 The current metadata format is very simple. ``!tbaa.struct`` metadata
4181 nodes are a list of operands which are in conceptual groups of three.
4182 For each group of three, the first operand gives the byte offset of a
4183 field in bytes, the second gives its size in bytes, and the third gives
4186 .. code-block:: llvm
4188 !4 = !{ i64 0, i64 4, !1, i64 8, i64 4, !2 }
4190 This describes a struct with two fields. The first is at offset 0 bytes
4191 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
4192 and has size 4 bytes and has tbaa tag !2.
4194 Note that the fields need not be contiguous. In this example, there is a
4195 4 byte gap between the two fields. This gap represents padding which
4196 does not carry useful data and need not be preserved.
4198 '``noalias``' and '``alias.scope``' Metadata
4199 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4201 ``noalias`` and ``alias.scope`` metadata provide the ability to specify generic
4202 noalias memory-access sets. This means that some collection of memory access
4203 instructions (loads, stores, memory-accessing calls, etc.) that carry
4204 ``noalias`` metadata can specifically be specified not to alias with some other
4205 collection of memory access instructions that carry ``alias.scope`` metadata.
4206 Each type of metadata specifies a list of scopes where each scope has an id and
4207 a domain. When evaluating an aliasing query, if for some domain, the set
4208 of scopes with that domain in one instruction's ``alias.scope`` list is a
4209 subset of (or equal to) the set of scopes for that domain in another
4210 instruction's ``noalias`` list, then the two memory accesses are assumed not to
4213 The metadata identifying each domain is itself a list containing one or two
4214 entries. The first entry is the name of the domain. Note that if the name is a
4215 string then it can be combined across functions and translation units. A
4216 self-reference can be used to create globally unique domain names. A
4217 descriptive string may optionally be provided as a second list entry.
4219 The metadata identifying each scope is also itself a list containing two or
4220 three entries. The first entry is the name of the scope. Note that if the name
4221 is a string then it can be combined across functions and translation units. A
4222 self-reference can be used to create globally unique scope names. A metadata
4223 reference to the scope's domain is the second entry. A descriptive string may
4224 optionally be provided as a third list entry.
4228 .. code-block:: llvm
4230 ; Two scope domains:
4234 ; Some scopes in these domains:
4240 !5 = !{!4} ; A list containing only scope !4
4244 ; These two instructions don't alias:
4245 %0 = load float, float* %c, align 4, !alias.scope !5
4246 store float %0, float* %arrayidx.i, align 4, !noalias !5
4248 ; These two instructions also don't alias (for domain !1, the set of scopes
4249 ; in the !alias.scope equals that in the !noalias list):
4250 %2 = load float, float* %c, align 4, !alias.scope !5
4251 store float %2, float* %arrayidx.i2, align 4, !noalias !6
4253 ; These two instructions may alias (for domain !0, the set of scopes in
4254 ; the !noalias list is not a superset of, or equal to, the scopes in the
4255 ; !alias.scope list):
4256 %2 = load float, float* %c, align 4, !alias.scope !6
4257 store float %0, float* %arrayidx.i, align 4, !noalias !7
4259 '``fpmath``' Metadata
4260 ^^^^^^^^^^^^^^^^^^^^^
4262 ``fpmath`` metadata may be attached to any instruction of floating point
4263 type. It can be used to express the maximum acceptable error in the
4264 result of that instruction, in ULPs, thus potentially allowing the
4265 compiler to use a more efficient but less accurate method of computing
4266 it. ULP is defined as follows:
4268 If ``x`` is a real number that lies between two finite consecutive
4269 floating-point numbers ``a`` and ``b``, without being equal to one
4270 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
4271 distance between the two non-equal finite floating-point numbers
4272 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
4274 The metadata node shall consist of a single positive floating point
4275 number representing the maximum relative error, for example:
4277 .. code-block:: llvm
4279 !0 = !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
4283 '``range``' Metadata
4284 ^^^^^^^^^^^^^^^^^^^^
4286 ``range`` metadata may be attached only to ``load``, ``call`` and ``invoke`` of
4287 integer types. It expresses the possible ranges the loaded value or the value
4288 returned by the called function at this call site is in. The ranges are
4289 represented with a flattened list of integers. The loaded value or the value
4290 returned is known to be in the union of the ranges defined by each consecutive
4291 pair. Each pair has the following properties:
4293 - The type must match the type loaded by the instruction.
4294 - The pair ``a,b`` represents the range ``[a,b)``.
4295 - Both ``a`` and ``b`` are constants.
4296 - The range is allowed to wrap.
4297 - The range should not represent the full or empty set. That is,
4300 In addition, the pairs must be in signed order of the lower bound and
4301 they must be non-contiguous.
4305 .. code-block:: llvm
4307 %a = load i8, i8* %x, align 1, !range !0 ; Can only be 0 or 1
4308 %b = load i8, i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
4309 %c = call i8 @foo(), !range !2 ; Can only be 0, 1, 3, 4 or 5
4310 %d = invoke i8 @bar() to label %cont
4311 unwind label %lpad, !range !3 ; Can only be -2, -1, 3, 4 or 5
4313 !0 = !{ i8 0, i8 2 }
4314 !1 = !{ i8 255, i8 2 }
4315 !2 = !{ i8 0, i8 2, i8 3, i8 6 }
4316 !3 = !{ i8 -2, i8 0, i8 3, i8 6 }
4318 '``unpredictable``' Metadata
4319 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4321 ``unpredictable`` metadata may be attached to any branch or switch
4322 instruction. It can be used to express the unpredictability of control
4323 flow. Similar to the llvm.expect intrinsic, it may be used to alter
4324 optimizations related to compare and branch instructions. The metadata
4325 is treated as a boolean value; if it exists, it signals that the branch
4326 or switch that it is attached to is completely unpredictable.
4331 It is sometimes useful to attach information to loop constructs. Currently,
4332 loop metadata is implemented as metadata attached to the branch instruction
4333 in the loop latch block. This type of metadata refer to a metadata node that is
4334 guaranteed to be separate for each loop. The loop identifier metadata is
4335 specified with the name ``llvm.loop``.
4337 The loop identifier metadata is implemented using a metadata that refers to
4338 itself to avoid merging it with any other identifier metadata, e.g.,
4339 during module linkage or function inlining. That is, each loop should refer
4340 to their own identification metadata even if they reside in separate functions.
4341 The following example contains loop identifier metadata for two separate loop
4344 .. code-block:: llvm
4349 The loop identifier metadata can be used to specify additional
4350 per-loop metadata. Any operands after the first operand can be treated
4351 as user-defined metadata. For example the ``llvm.loop.unroll.count``
4352 suggests an unroll factor to the loop unroller:
4354 .. code-block:: llvm
4356 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
4359 !1 = !{!"llvm.loop.unroll.count", i32 4}
4361 '``llvm.loop.vectorize``' and '``llvm.loop.interleave``'
4362 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4364 Metadata prefixed with ``llvm.loop.vectorize`` or ``llvm.loop.interleave`` are
4365 used to control per-loop vectorization and interleaving parameters such as
4366 vectorization width and interleave count. These metadata should be used in
4367 conjunction with ``llvm.loop`` loop identification metadata. The
4368 ``llvm.loop.vectorize`` and ``llvm.loop.interleave`` metadata are only
4369 optimization hints and the optimizer will only interleave and vectorize loops if
4370 it believes it is safe to do so. The ``llvm.mem.parallel_loop_access`` metadata
4371 which contains information about loop-carried memory dependencies can be helpful
4372 in determining the safety of these transformations.
4374 '``llvm.loop.interleave.count``' Metadata
4375 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4377 This metadata suggests an interleave count to the loop interleaver.
4378 The first operand is the string ``llvm.loop.interleave.count`` and the
4379 second operand is an integer specifying the interleave count. For
4382 .. code-block:: llvm
4384 !0 = !{!"llvm.loop.interleave.count", i32 4}
4386 Note that setting ``llvm.loop.interleave.count`` to 1 disables interleaving
4387 multiple iterations of the loop. If ``llvm.loop.interleave.count`` is set to 0
4388 then the interleave count will be determined automatically.
4390 '``llvm.loop.vectorize.enable``' Metadata
4391 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4393 This metadata selectively enables or disables vectorization for the loop. The
4394 first operand is the string ``llvm.loop.vectorize.enable`` and the second operand
4395 is a bit. If the bit operand value is 1 vectorization is enabled. A value of
4396 0 disables vectorization:
4398 .. code-block:: llvm
4400 !0 = !{!"llvm.loop.vectorize.enable", i1 0}
4401 !1 = !{!"llvm.loop.vectorize.enable", i1 1}
4403 '``llvm.loop.vectorize.width``' Metadata
4404 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4406 This metadata sets the target width of the vectorizer. The first
4407 operand is the string ``llvm.loop.vectorize.width`` and the second
4408 operand is an integer specifying the width. For example:
4410 .. code-block:: llvm
4412 !0 = !{!"llvm.loop.vectorize.width", i32 4}
4414 Note that setting ``llvm.loop.vectorize.width`` to 1 disables
4415 vectorization of the loop. If ``llvm.loop.vectorize.width`` is set to
4416 0 or if the loop does not have this metadata the width will be
4417 determined automatically.
4419 '``llvm.loop.unroll``'
4420 ^^^^^^^^^^^^^^^^^^^^^^
4422 Metadata prefixed with ``llvm.loop.unroll`` are loop unrolling
4423 optimization hints such as the unroll factor. ``llvm.loop.unroll``
4424 metadata should be used in conjunction with ``llvm.loop`` loop
4425 identification metadata. The ``llvm.loop.unroll`` metadata are only
4426 optimization hints and the unrolling will only be performed if the
4427 optimizer believes it is safe to do so.
4429 '``llvm.loop.unroll.count``' Metadata
4430 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4432 This metadata suggests an unroll factor to the loop unroller. The
4433 first operand is the string ``llvm.loop.unroll.count`` and the second
4434 operand is a positive integer specifying the unroll factor. For
4437 .. code-block:: llvm
4439 !0 = !{!"llvm.loop.unroll.count", i32 4}
4441 If the trip count of the loop is less than the unroll count the loop
4442 will be partially unrolled.
4444 '``llvm.loop.unroll.disable``' Metadata
4445 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4447 This metadata disables loop unrolling. The metadata has a single operand
4448 which is the string ``llvm.loop.unroll.disable``. For example:
4450 .. code-block:: llvm
4452 !0 = !{!"llvm.loop.unroll.disable"}
4454 '``llvm.loop.unroll.runtime.disable``' Metadata
4455 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4457 This metadata disables runtime loop unrolling. The metadata has a single
4458 operand which is the string ``llvm.loop.unroll.runtime.disable``. For example:
4460 .. code-block:: llvm
4462 !0 = !{!"llvm.loop.unroll.runtime.disable"}
4464 '``llvm.loop.unroll.enable``' Metadata
4465 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4467 This metadata suggests that the loop should be fully unrolled if the trip count
4468 is known at compile time and partially unrolled if the trip count is not known
4469 at compile time. The metadata has a single operand which is the string
4470 ``llvm.loop.unroll.enable``. For example:
4472 .. code-block:: llvm
4474 !0 = !{!"llvm.loop.unroll.enable"}
4476 '``llvm.loop.unroll.full``' Metadata
4477 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4479 This metadata suggests that the loop should be unrolled fully. The
4480 metadata has a single operand which is the string ``llvm.loop.unroll.full``.
4483 .. code-block:: llvm
4485 !0 = !{!"llvm.loop.unroll.full"}
4490 Metadata types used to annotate memory accesses with information helpful
4491 for optimizations are prefixed with ``llvm.mem``.
4493 '``llvm.mem.parallel_loop_access``' Metadata
4494 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4496 The ``llvm.mem.parallel_loop_access`` metadata refers to a loop identifier,
4497 or metadata containing a list of loop identifiers for nested loops.
4498 The metadata is attached to memory accessing instructions and denotes that
4499 no loop carried memory dependence exist between it and other instructions denoted
4500 with the same loop identifier.
4502 Precisely, given two instructions ``m1`` and ``m2`` that both have the
4503 ``llvm.mem.parallel_loop_access`` metadata, with ``L1`` and ``L2`` being the
4504 set of loops associated with that metadata, respectively, then there is no loop
4505 carried dependence between ``m1`` and ``m2`` for loops in both ``L1`` and
4508 As a special case, if all memory accessing instructions in a loop have
4509 ``llvm.mem.parallel_loop_access`` metadata that refers to that loop, then the
4510 loop has no loop carried memory dependences and is considered to be a parallel
4513 Note that if not all memory access instructions have such metadata referring to
4514 the loop, then the loop is considered not being trivially parallel. Additional
4515 memory dependence analysis is required to make that determination. As a fail
4516 safe mechanism, this causes loops that were originally parallel to be considered
4517 sequential (if optimization passes that are unaware of the parallel semantics
4518 insert new memory instructions into the loop body).
4520 Example of a loop that is considered parallel due to its correct use of
4521 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
4522 metadata types that refer to the same loop identifier metadata.
4524 .. code-block:: llvm
4528 %val0 = load i32, i32* %arrayidx, !llvm.mem.parallel_loop_access !0
4530 store i32 %val0, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
4532 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
4538 It is also possible to have nested parallel loops. In that case the
4539 memory accesses refer to a list of loop identifier metadata nodes instead of
4540 the loop identifier metadata node directly:
4542 .. code-block:: llvm
4546 %val1 = load i32, i32* %arrayidx3, !llvm.mem.parallel_loop_access !2
4548 br label %inner.for.body
4552 %val0 = load i32, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
4554 store i32 %val0, i32* %arrayidx2, !llvm.mem.parallel_loop_access !0
4556 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
4560 store i32 %val1, i32* %arrayidx4, !llvm.mem.parallel_loop_access !2
4562 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
4564 outer.for.end: ; preds = %for.body
4566 !0 = !{!1, !2} ; a list of loop identifiers
4567 !1 = !{!1} ; an identifier for the inner loop
4568 !2 = !{!2} ; an identifier for the outer loop
4573 The ``llvm.bitsets`` global metadata is used to implement
4574 :doc:`bitsets <BitSets>`.
4576 '``invariant.group``' Metadata
4577 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4579 The ``invariant.group`` metadata may be attached to ``load``/``store`` instructions.
4580 The existence of the ``invariant.group`` metadata on the instruction tells
4581 the optimizer that every ``load`` and ``store`` to the same pointer operand
4582 within the same invariant group can be assumed to load or store the same
4583 value (but see the ``llvm.invariant.group.barrier`` intrinsic which affects
4584 when two pointers are considered the same).
4588 .. code-block:: llvm
4590 @unknownPtr = external global i8
4593 store i8 42, i8* %ptr, !invariant.group !0
4594 call void @foo(i8* %ptr)
4596 %a = load i8, i8* %ptr, !invariant.group !0 ; Can assume that value under %ptr didn't change
4597 call void @foo(i8* %ptr)
4598 %b = load i8, i8* %ptr, !invariant.group !1 ; Can't assume anything, because group changed
4600 %newPtr = call i8* @getPointer(i8* %ptr)
4601 %c = load i8, i8* %newPtr, !invariant.group !0 ; Can't assume anything, because we only have information about %ptr
4603 %unknownValue = load i8, i8* @unknownPtr
4604 store i8 %unknownValue, i8* %ptr, !invariant.group !0 ; Can assume that %unknownValue == 42
4606 call void @foo(i8* %ptr)
4607 %newPtr2 = call i8* @llvm.invariant.group.barrier(i8* %ptr)
4608 %d = load i8, i8* %newPtr2, !invariant.group !0 ; Can't step through invariant.group.barrier to get value of %ptr
4611 declare void @foo(i8*)
4612 declare i8* @getPointer(i8*)
4613 declare i8* @llvm.invariant.group.barrier(i8*)
4615 !0 = !{!"magic ptr"}
4616 !1 = !{!"other ptr"}
4620 Module Flags Metadata
4621 =====================
4623 Information about the module as a whole is difficult to convey to LLVM's
4624 subsystems. The LLVM IR isn't sufficient to transmit this information.
4625 The ``llvm.module.flags`` named metadata exists in order to facilitate
4626 this. These flags are in the form of key / value pairs --- much like a
4627 dictionary --- making it easy for any subsystem who cares about a flag to
4630 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
4631 Each triplet has the following form:
4633 - The first element is a *behavior* flag, which specifies the behavior
4634 when two (or more) modules are merged together, and it encounters two
4635 (or more) metadata with the same ID. The supported behaviors are
4637 - The second element is a metadata string that is a unique ID for the
4638 metadata. Each module may only have one flag entry for each unique ID (not
4639 including entries with the **Require** behavior).
4640 - The third element is the value of the flag.
4642 When two (or more) modules are merged together, the resulting
4643 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
4644 each unique metadata ID string, there will be exactly one entry in the merged
4645 modules ``llvm.module.flags`` metadata table, and the value for that entry will
4646 be determined by the merge behavior flag, as described below. The only exception
4647 is that entries with the *Require* behavior are always preserved.
4649 The following behaviors are supported:
4660 Emits an error if two values disagree, otherwise the resulting value
4661 is that of the operands.
4665 Emits a warning if two values disagree. The result value will be the
4666 operand for the flag from the first module being linked.
4670 Adds a requirement that another module flag be present and have a
4671 specified value after linking is performed. The value must be a
4672 metadata pair, where the first element of the pair is the ID of the
4673 module flag to be restricted, and the second element of the pair is
4674 the value the module flag should be restricted to. This behavior can
4675 be used to restrict the allowable results (via triggering of an
4676 error) of linking IDs with the **Override** behavior.
4680 Uses the specified value, regardless of the behavior or value of the
4681 other module. If both modules specify **Override**, but the values
4682 differ, an error will be emitted.
4686 Appends the two values, which are required to be metadata nodes.
4690 Appends the two values, which are required to be metadata
4691 nodes. However, duplicate entries in the second list are dropped
4692 during the append operation.
4694 It is an error for a particular unique flag ID to have multiple behaviors,
4695 except in the case of **Require** (which adds restrictions on another metadata
4696 value) or **Override**.
4698 An example of module flags:
4700 .. code-block:: llvm
4702 !0 = !{ i32 1, !"foo", i32 1 }
4703 !1 = !{ i32 4, !"bar", i32 37 }
4704 !2 = !{ i32 2, !"qux", i32 42 }
4705 !3 = !{ i32 3, !"qux",
4710 !llvm.module.flags = !{ !0, !1, !2, !3 }
4712 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
4713 if two or more ``!"foo"`` flags are seen is to emit an error if their
4714 values are not equal.
4716 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
4717 behavior if two or more ``!"bar"`` flags are seen is to use the value
4720 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
4721 behavior if two or more ``!"qux"`` flags are seen is to emit a
4722 warning if their values are not equal.
4724 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
4730 The behavior is to emit an error if the ``llvm.module.flags`` does not
4731 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
4734 Objective-C Garbage Collection Module Flags Metadata
4735 ----------------------------------------------------
4737 On the Mach-O platform, Objective-C stores metadata about garbage
4738 collection in a special section called "image info". The metadata
4739 consists of a version number and a bitmask specifying what types of
4740 garbage collection are supported (if any) by the file. If two or more
4741 modules are linked together their garbage collection metadata needs to
4742 be merged rather than appended together.
4744 The Objective-C garbage collection module flags metadata consists of the
4745 following key-value pairs:
4754 * - ``Objective-C Version``
4755 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
4757 * - ``Objective-C Image Info Version``
4758 - **[Required]** --- The version of the image info section. Currently
4761 * - ``Objective-C Image Info Section``
4762 - **[Required]** --- The section to place the metadata. Valid values are
4763 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
4764 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
4765 Objective-C ABI version 2.
4767 * - ``Objective-C Garbage Collection``
4768 - **[Required]** --- Specifies whether garbage collection is supported or
4769 not. Valid values are 0, for no garbage collection, and 2, for garbage
4770 collection supported.
4772 * - ``Objective-C GC Only``
4773 - **[Optional]** --- Specifies that only garbage collection is supported.
4774 If present, its value must be 6. This flag requires that the
4775 ``Objective-C Garbage Collection`` flag have the value 2.
4777 Some important flag interactions:
4779 - If a module with ``Objective-C Garbage Collection`` set to 0 is
4780 merged with a module with ``Objective-C Garbage Collection`` set to
4781 2, then the resulting module has the
4782 ``Objective-C Garbage Collection`` flag set to 0.
4783 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
4784 merged with a module with ``Objective-C GC Only`` set to 6.
4786 Automatic Linker Flags Module Flags Metadata
4787 --------------------------------------------
4789 Some targets support embedding flags to the linker inside individual object
4790 files. Typically this is used in conjunction with language extensions which
4791 allow source files to explicitly declare the libraries they depend on, and have
4792 these automatically be transmitted to the linker via object files.
4794 These flags are encoded in the IR using metadata in the module flags section,
4795 using the ``Linker Options`` key. The merge behavior for this flag is required
4796 to be ``AppendUnique``, and the value for the key is expected to be a metadata
4797 node which should be a list of other metadata nodes, each of which should be a
4798 list of metadata strings defining linker options.
4800 For example, the following metadata section specifies two separate sets of
4801 linker options, presumably to link against ``libz`` and the ``Cocoa``
4804 !0 = !{ i32 6, !"Linker Options",
4807 !{ !"-framework", !"Cocoa" } } }
4808 !llvm.module.flags = !{ !0 }
4810 The metadata encoding as lists of lists of options, as opposed to a collapsed
4811 list of options, is chosen so that the IR encoding can use multiple option
4812 strings to specify e.g., a single library, while still having that specifier be
4813 preserved as an atomic element that can be recognized by a target specific
4814 assembly writer or object file emitter.
4816 Each individual option is required to be either a valid option for the target's
4817 linker, or an option that is reserved by the target specific assembly writer or
4818 object file emitter. No other aspect of these options is defined by the IR.
4820 C type width Module Flags Metadata
4821 ----------------------------------
4823 The ARM backend emits a section into each generated object file describing the
4824 options that it was compiled with (in a compiler-independent way) to prevent
4825 linking incompatible objects, and to allow automatic library selection. Some
4826 of these options are not visible at the IR level, namely wchar_t width and enum
4829 To pass this information to the backend, these options are encoded in module
4830 flags metadata, using the following key-value pairs:
4840 - * 0 --- sizeof(wchar_t) == 4
4841 * 1 --- sizeof(wchar_t) == 2
4844 - * 0 --- Enums are at least as large as an ``int``.
4845 * 1 --- Enums are stored in the smallest integer type which can
4846 represent all of its values.
4848 For example, the following metadata section specifies that the module was
4849 compiled with a ``wchar_t`` width of 4 bytes, and the underlying type of an
4850 enum is the smallest type which can represent all of its values::
4852 !llvm.module.flags = !{!0, !1}
4853 !0 = !{i32 1, !"short_wchar", i32 1}
4854 !1 = !{i32 1, !"short_enum", i32 0}
4856 .. _intrinsicglobalvariables:
4858 Intrinsic Global Variables
4859 ==========================
4861 LLVM has a number of "magic" global variables that contain data that
4862 affect code generation or other IR semantics. These are documented here.
4863 All globals of this sort should have a section specified as
4864 "``llvm.metadata``". This section and all globals that start with
4865 "``llvm.``" are reserved for use by LLVM.
4869 The '``llvm.used``' Global Variable
4870 -----------------------------------
4872 The ``@llvm.used`` global is an array which has
4873 :ref:`appending linkage <linkage_appending>`. This array contains a list of
4874 pointers to named global variables, functions and aliases which may optionally
4875 have a pointer cast formed of bitcast or getelementptr. For example, a legal
4878 .. code-block:: llvm
4883 @llvm.used = appending global [2 x i8*] [
4885 i8* bitcast (i32* @Y to i8*)
4886 ], section "llvm.metadata"
4888 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
4889 and linker are required to treat the symbol as if there is a reference to the
4890 symbol that it cannot see (which is why they have to be named). For example, if
4891 a variable has internal linkage and no references other than that from the
4892 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
4893 references from inline asms and other things the compiler cannot "see", and
4894 corresponds to "``attribute((used))``" in GNU C.
4896 On some targets, the code generator must emit a directive to the
4897 assembler or object file to prevent the assembler and linker from
4898 molesting the symbol.
4900 .. _gv_llvmcompilerused:
4902 The '``llvm.compiler.used``' Global Variable
4903 --------------------------------------------
4905 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
4906 directive, except that it only prevents the compiler from touching the
4907 symbol. On targets that support it, this allows an intelligent linker to
4908 optimize references to the symbol without being impeded as it would be
4911 This is a rare construct that should only be used in rare circumstances,
4912 and should not be exposed to source languages.
4914 .. _gv_llvmglobalctors:
4916 The '``llvm.global_ctors``' Global Variable
4917 -------------------------------------------
4919 .. code-block:: llvm
4921 %0 = type { i32, void ()*, i8* }
4922 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor, i8* @data }]
4924 The ``@llvm.global_ctors`` array contains a list of constructor
4925 functions, priorities, and an optional associated global or function.
4926 The functions referenced by this array will be called in ascending order
4927 of priority (i.e. lowest first) when the module is loaded. The order of
4928 functions with the same priority is not defined.
4930 If the third field is present, non-null, and points to a global variable
4931 or function, the initializer function will only run if the associated
4932 data from the current module is not discarded.
4934 .. _llvmglobaldtors:
4936 The '``llvm.global_dtors``' Global Variable
4937 -------------------------------------------
4939 .. code-block:: llvm
4941 %0 = type { i32, void ()*, i8* }
4942 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor, i8* @data }]
4944 The ``@llvm.global_dtors`` array contains a list of destructor
4945 functions, priorities, and an optional associated global or function.
4946 The functions referenced by this array will be called in descending
4947 order of priority (i.e. highest first) when the module is unloaded. The
4948 order of functions with the same priority is not defined.
4950 If the third field is present, non-null, and points to a global variable
4951 or function, the destructor function will only run if the associated
4952 data from the current module is not discarded.
4954 Instruction Reference
4955 =====================
4957 The LLVM instruction set consists of several different classifications
4958 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
4959 instructions <binaryops>`, :ref:`bitwise binary
4960 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
4961 :ref:`other instructions <otherops>`.
4965 Terminator Instructions
4966 -----------------------
4968 As mentioned :ref:`previously <functionstructure>`, every basic block in a
4969 program ends with a "Terminator" instruction, which indicates which
4970 block should be executed after the current block is finished. These
4971 terminator instructions typically yield a '``void``' value: they produce
4972 control flow, not values (the one exception being the
4973 ':ref:`invoke <i_invoke>`' instruction).
4975 The terminator instructions are: ':ref:`ret <i_ret>`',
4976 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
4977 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
4978 ':ref:`resume <i_resume>`', ':ref:`catchpad <i_catchpad>`',
4979 ':ref:`catchendpad <i_catchendpad>`',
4980 ':ref:`catchret <i_catchret>`',
4981 ':ref:`cleanupendpad <i_cleanupendpad>`',
4982 ':ref:`cleanupret <i_cleanupret>`',
4983 ':ref:`terminatepad <i_terminatepad>`',
4984 and ':ref:`unreachable <i_unreachable>`'.
4988 '``ret``' Instruction
4989 ^^^^^^^^^^^^^^^^^^^^^
4996 ret <type> <value> ; Return a value from a non-void function
4997 ret void ; Return from void function
5002 The '``ret``' instruction is used to return control flow (and optionally
5003 a value) from a function back to the caller.
5005 There are two forms of the '``ret``' instruction: one that returns a
5006 value and then causes control flow, and one that just causes control
5012 The '``ret``' instruction optionally accepts a single argument, the
5013 return value. The type of the return value must be a ':ref:`first
5014 class <t_firstclass>`' type.
5016 A function is not :ref:`well formed <wellformed>` if it it has a non-void
5017 return type and contains a '``ret``' instruction with no return value or
5018 a return value with a type that does not match its type, or if it has a
5019 void return type and contains a '``ret``' instruction with a return
5025 When the '``ret``' instruction is executed, control flow returns back to
5026 the calling function's context. If the caller is a
5027 ":ref:`call <i_call>`" instruction, execution continues at the
5028 instruction after the call. If the caller was an
5029 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
5030 beginning of the "normal" destination block. If the instruction returns
5031 a value, that value shall set the call or invoke instruction's return
5037 .. code-block:: llvm
5039 ret i32 5 ; Return an integer value of 5
5040 ret void ; Return from a void function
5041 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
5045 '``br``' Instruction
5046 ^^^^^^^^^^^^^^^^^^^^
5053 br i1 <cond>, label <iftrue>, label <iffalse>
5054 br label <dest> ; Unconditional branch
5059 The '``br``' instruction is used to cause control flow to transfer to a
5060 different basic block in the current function. There are two forms of
5061 this instruction, corresponding to a conditional branch and an
5062 unconditional branch.
5067 The conditional branch form of the '``br``' instruction takes a single
5068 '``i1``' value and two '``label``' values. The unconditional form of the
5069 '``br``' instruction takes a single '``label``' value as a target.
5074 Upon execution of a conditional '``br``' instruction, the '``i1``'
5075 argument is evaluated. If the value is ``true``, control flows to the
5076 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
5077 to the '``iffalse``' ``label`` argument.
5082 .. code-block:: llvm
5085 %cond = icmp eq i32 %a, %b
5086 br i1 %cond, label %IfEqual, label %IfUnequal
5094 '``switch``' Instruction
5095 ^^^^^^^^^^^^^^^^^^^^^^^^
5102 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
5107 The '``switch``' instruction is used to transfer control flow to one of
5108 several different places. It is a generalization of the '``br``'
5109 instruction, allowing a branch to occur to one of many possible
5115 The '``switch``' instruction uses three parameters: an integer
5116 comparison value '``value``', a default '``label``' destination, and an
5117 array of pairs of comparison value constants and '``label``'s. The table
5118 is not allowed to contain duplicate constant entries.
5123 The ``switch`` instruction specifies a table of values and destinations.
5124 When the '``switch``' instruction is executed, this table is searched
5125 for the given value. If the value is found, control flow is transferred
5126 to the corresponding destination; otherwise, control flow is transferred
5127 to the default destination.
5132 Depending on properties of the target machine and the particular
5133 ``switch`` instruction, this instruction may be code generated in
5134 different ways. For example, it could be generated as a series of
5135 chained conditional branches or with a lookup table.
5140 .. code-block:: llvm
5142 ; Emulate a conditional br instruction
5143 %Val = zext i1 %value to i32
5144 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
5146 ; Emulate an unconditional br instruction
5147 switch i32 0, label %dest [ ]
5149 ; Implement a jump table:
5150 switch i32 %val, label %otherwise [ i32 0, label %onzero
5152 i32 2, label %ontwo ]
5156 '``indirectbr``' Instruction
5157 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5164 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
5169 The '``indirectbr``' instruction implements an indirect branch to a
5170 label within the current function, whose address is specified by
5171 "``address``". Address must be derived from a
5172 :ref:`blockaddress <blockaddress>` constant.
5177 The '``address``' argument is the address of the label to jump to. The
5178 rest of the arguments indicate the full set of possible destinations
5179 that the address may point to. Blocks are allowed to occur multiple
5180 times in the destination list, though this isn't particularly useful.
5182 This destination list is required so that dataflow analysis has an
5183 accurate understanding of the CFG.
5188 Control transfers to the block specified in the address argument. All
5189 possible destination blocks must be listed in the label list, otherwise
5190 this instruction has undefined behavior. This implies that jumps to
5191 labels defined in other functions have undefined behavior as well.
5196 This is typically implemented with a jump through a register.
5201 .. code-block:: llvm
5203 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
5207 '``invoke``' Instruction
5208 ^^^^^^^^^^^^^^^^^^^^^^^^
5215 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
5216 [operand bundles] to label <normal label> unwind label <exception label>
5221 The '``invoke``' instruction causes control to transfer to a specified
5222 function, with the possibility of control flow transfer to either the
5223 '``normal``' label or the '``exception``' label. If the callee function
5224 returns with the "``ret``" instruction, control flow will return to the
5225 "normal" label. If the callee (or any indirect callees) returns via the
5226 ":ref:`resume <i_resume>`" instruction or other exception handling
5227 mechanism, control is interrupted and continued at the dynamically
5228 nearest "exception" label.
5230 The '``exception``' label is a `landing
5231 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
5232 '``exception``' label is required to have the
5233 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
5234 information about the behavior of the program after unwinding happens,
5235 as its first non-PHI instruction. The restrictions on the
5236 "``landingpad``" instruction's tightly couples it to the "``invoke``"
5237 instruction, so that the important information contained within the
5238 "``landingpad``" instruction can't be lost through normal code motion.
5243 This instruction requires several arguments:
5245 #. The optional "cconv" marker indicates which :ref:`calling
5246 convention <callingconv>` the call should use. If none is
5247 specified, the call defaults to using C calling conventions.
5248 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
5249 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
5251 #. '``ptr to function ty``': shall be the signature of the pointer to
5252 function value being invoked. In most cases, this is a direct
5253 function invocation, but indirect ``invoke``'s are just as possible,
5254 branching off an arbitrary pointer to function value.
5255 #. '``function ptr val``': An LLVM value containing a pointer to a
5256 function to be invoked.
5257 #. '``function args``': argument list whose types match the function
5258 signature argument types and parameter attributes. All arguments must
5259 be of :ref:`first class <t_firstclass>` type. If the function signature
5260 indicates the function accepts a variable number of arguments, the
5261 extra arguments can be specified.
5262 #. '``normal label``': the label reached when the called function
5263 executes a '``ret``' instruction.
5264 #. '``exception label``': the label reached when a callee returns via
5265 the :ref:`resume <i_resume>` instruction or other exception handling
5267 #. The optional :ref:`function attributes <fnattrs>` list. Only
5268 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
5269 attributes are valid here.
5270 #. The optional :ref:`operand bundles <opbundles>` list.
5275 This instruction is designed to operate as a standard '``call``'
5276 instruction in most regards. The primary difference is that it
5277 establishes an association with a label, which is used by the runtime
5278 library to unwind the stack.
5280 This instruction is used in languages with destructors to ensure that
5281 proper cleanup is performed in the case of either a ``longjmp`` or a
5282 thrown exception. Additionally, this is important for implementation of
5283 '``catch``' clauses in high-level languages that support them.
5285 For the purposes of the SSA form, the definition of the value returned
5286 by the '``invoke``' instruction is deemed to occur on the edge from the
5287 current block to the "normal" label. If the callee unwinds then no
5288 return value is available.
5293 .. code-block:: llvm
5295 %retval = invoke i32 @Test(i32 15) to label %Continue
5296 unwind label %TestCleanup ; i32:retval set
5297 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
5298 unwind label %TestCleanup ; i32:retval set
5302 '``resume``' Instruction
5303 ^^^^^^^^^^^^^^^^^^^^^^^^
5310 resume <type> <value>
5315 The '``resume``' instruction is a terminator instruction that has no
5321 The '``resume``' instruction requires one argument, which must have the
5322 same type as the result of any '``landingpad``' instruction in the same
5328 The '``resume``' instruction resumes propagation of an existing
5329 (in-flight) exception whose unwinding was interrupted with a
5330 :ref:`landingpad <i_landingpad>` instruction.
5335 .. code-block:: llvm
5337 resume { i8*, i32 } %exn
5341 '``catchpad``' Instruction
5342 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5349 <resultval> = catchpad [<args>*]
5350 to label <normal label> unwind label <exception label>
5355 The '``catchpad``' instruction is used by `LLVM's exception handling
5356 system <ExceptionHandling.html#overview>`_ to specify that a basic block
5357 is a catch block --- one where a personality routine attempts to transfer
5358 control to catch an exception.
5359 The ``args`` correspond to whatever information the personality
5360 routine requires to know if this is an appropriate place to catch the
5361 exception. Control is transfered to the ``exception`` label if the
5362 ``catchpad`` is not an appropriate handler for the in-flight exception.
5363 The ``normal`` label should contain the code found in the ``catch``
5364 portion of a ``try``/``catch`` sequence. The ``resultval`` has the type
5365 :ref:`token <t_token>` and is used to match the ``catchpad`` to
5366 corresponding :ref:`catchrets <i_catchret>`.
5371 The instruction takes a list of arbitrary values which are interpreted
5372 by the :ref:`personality function <personalityfn>`.
5374 The ``catchpad`` must be provided a ``normal`` label to transfer control
5375 to if the ``catchpad`` matches the exception and an ``exception``
5376 label to transfer control to if it doesn't.
5381 When the call stack is being unwound due to an exception being thrown,
5382 the exception is compared against the ``args``. If it doesn't match,
5383 then control is transfered to the ``exception`` basic block.
5384 As with calling conventions, how the personality function results are
5385 represented in LLVM IR is target specific.
5387 The ``catchpad`` instruction has several restrictions:
5389 - A catch block is a basic block which is the unwind destination of
5390 an exceptional instruction.
5391 - A catch block must have a '``catchpad``' instruction as its
5392 first non-PHI instruction.
5393 - A catch block's ``exception`` edge must refer to a catch block or a
5395 - There can be only one '``catchpad``' instruction within the
5397 - A basic block that is not a catch block may not include a
5398 '``catchpad``' instruction.
5399 - A catch block which has another catch block as a predecessor may not have
5400 any other predecessors.
5401 - It is undefined behavior for control to transfer from a ``catchpad`` to a
5402 ``ret`` without first executing a ``catchret`` that consumes the
5403 ``catchpad`` or unwinding through its ``catchendpad``.
5404 - It is undefined behavior for control to transfer from a ``catchpad`` to
5405 itself without first executing a ``catchret`` that consumes the
5406 ``catchpad`` or unwinding through its ``catchendpad``.
5411 .. code-block:: llvm
5413 ;; A catch block which can catch an integer.
5414 %tok = catchpad [i8** @_ZTIi]
5415 to label %int.handler unwind label %terminate
5419 '``catchendpad``' Instruction
5420 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5427 catchendpad unwind label <nextaction>
5428 catchendpad unwind to caller
5433 The '``catchendpad``' instruction is used by `LLVM's exception handling
5434 system <ExceptionHandling.html#overview>`_ to communicate to the
5435 :ref:`personality function <personalityfn>` which invokes are associated
5436 with a chain of :ref:`catchpad <i_catchpad>` instructions; propagating an
5437 exception out of a catch handler is represented by unwinding through its
5438 ``catchendpad``. Unwinding to the outer scope when a chain of catch handlers
5439 do not handle an exception is also represented by unwinding through their
5442 The ``nextaction`` label indicates where control should transfer to if
5443 none of the ``catchpad`` instructions are suitable for catching the
5444 in-flight exception.
5446 If a ``nextaction`` label is not present, the instruction unwinds out of
5447 its parent function. The
5448 :ref:`personality function <personalityfn>` will continue processing
5449 exception handling actions in the caller.
5454 The instruction optionally takes a label, ``nextaction``, indicating
5455 where control should transfer to if none of the preceding
5456 ``catchpad`` instructions are suitable for the in-flight exception.
5461 When the call stack is being unwound due to an exception being thrown
5462 and none of the constituent ``catchpad`` instructions match, then
5463 control is transfered to ``nextaction`` if it is present. If it is not
5464 present, control is transfered to the caller.
5466 The ``catchendpad`` instruction has several restrictions:
5468 - A catch-end block is a basic block which is the unwind destination of
5469 an exceptional instruction.
5470 - A catch-end block must have a '``catchendpad``' instruction as its
5471 first non-PHI instruction.
5472 - There can be only one '``catchendpad``' instruction within the
5474 - A basic block that is not a catch-end block may not include a
5475 '``catchendpad``' instruction.
5476 - Exactly one catch block may unwind to a ``catchendpad``.
5477 - It is undefined behavior to execute a ``catchendpad`` if none of the
5478 '``catchpad``'s chained to it have been executed.
5479 - It is undefined behavior to execute a ``catchendpad`` twice without an
5480 intervening execution of one or more of the '``catchpad``'s chained to it.
5481 - It is undefined behavior to execute a ``catchendpad`` if, after the most
5482 recent execution of the normal successor edge of any ``catchpad`` chained
5483 to it, some ``catchret`` consuming that ``catchpad`` has already been
5485 - It is undefined behavior to execute a ``catchendpad`` if, after the most
5486 recent execution of the normal successor edge of any ``catchpad`` chained
5487 to it, any other ``catchpad`` or ``cleanuppad`` has been executed but has
5488 not had a corresponding
5489 ``catchret``/``cleanupret``/``catchendpad``/``cleanupendpad`` executed.
5494 .. code-block:: llvm
5496 catchendpad unwind label %terminate
5497 catchendpad unwind to caller
5501 '``catchret``' Instruction
5502 ^^^^^^^^^^^^^^^^^^^^^^^^^^
5509 catchret <value> to label <normal>
5514 The '``catchret``' instruction is a terminator instruction that has a
5521 The first argument to a '``catchret``' indicates which ``catchpad`` it
5522 exits. It must be a :ref:`catchpad <i_catchpad>`.
5523 The second argument to a '``catchret``' specifies where control will
5529 The '``catchret``' instruction ends the existing (in-flight) exception
5530 whose unwinding was interrupted with a
5531 :ref:`catchpad <i_catchpad>` instruction.
5532 The :ref:`personality function <personalityfn>` gets a chance to execute
5533 arbitrary code to, for example, run a C++ destructor.
5534 Control then transfers to ``normal``.
5535 It may be passed an optional, personality specific, value.
5537 It is undefined behavior to execute a ``catchret`` whose ``catchpad`` has
5540 It is undefined behavior to execute a ``catchret`` if, after the most recent
5541 execution of its ``catchpad``, some ``catchret`` or ``catchendpad`` linked
5542 to the same ``catchpad`` has already been executed.
5544 It is undefined behavior to execute a ``catchret`` if, after the most recent
5545 execution of its ``catchpad``, any other ``catchpad`` or ``cleanuppad`` has
5546 been executed but has not had a corresponding
5547 ``catchret``/``cleanupret``/``catchendpad``/``cleanupendpad`` executed.
5552 .. code-block:: llvm
5554 catchret %catch label %continue
5556 .. _i_cleanupendpad:
5558 '``cleanupendpad``' Instruction
5559 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5566 cleanupendpad <value> unwind label <nextaction>
5567 cleanupendpad <value> unwind to caller
5572 The '``cleanupendpad``' instruction is used by `LLVM's exception handling
5573 system <ExceptionHandling.html#overview>`_ to communicate to the
5574 :ref:`personality function <personalityfn>` which invokes are associated
5575 with a :ref:`cleanuppad <i_cleanuppad>` instructions; propagating an exception
5576 out of a cleanup is represented by unwinding through its ``cleanupendpad``.
5578 The ``nextaction`` label indicates where control should unwind to next, in the
5579 event that a cleanup is exited by means of an(other) exception being raised.
5581 If a ``nextaction`` label is not present, the instruction unwinds out of
5582 its parent function. The
5583 :ref:`personality function <personalityfn>` will continue processing
5584 exception handling actions in the caller.
5589 The '``cleanupendpad``' instruction requires one argument, which indicates
5590 which ``cleanuppad`` it exits, and must be a :ref:`cleanuppad <i_cleanuppad>`.
5591 It also has an optional successor, ``nextaction``, indicating where control
5597 When and exception propagates to a ``cleanupendpad``, control is transfered to
5598 ``nextaction`` if it is present. If it is not present, control is transfered to
5601 The ``cleanupendpad`` instruction has several restrictions:
5603 - A cleanup-end block is a basic block which is the unwind destination of
5604 an exceptional instruction.
5605 - A cleanup-end block must have a '``cleanupendpad``' instruction as its
5606 first non-PHI instruction.
5607 - There can be only one '``cleanupendpad``' instruction within the
5609 - A basic block that is not a cleanup-end block may not include a
5610 '``cleanupendpad``' instruction.
5611 - It is undefined behavior to execute a ``cleanupendpad`` whose ``cleanuppad``
5612 has not been executed.
5613 - It is undefined behavior to execute a ``cleanupendpad`` if, after the most
5614 recent execution of its ``cleanuppad``, some ``cleanupret`` or ``cleanupendpad``
5615 consuming the same ``cleanuppad`` has already been executed.
5616 - It is undefined behavior to execute a ``cleanupendpad`` if, after the most
5617 recent execution of its ``cleanuppad``, any other ``cleanuppad`` or
5618 ``catchpad`` has been executed but has not had a corresponding
5619 ``cleanupret``/``catchret``/``cleanupendpad``/``catchendpad`` executed.
5624 .. code-block:: llvm
5626 cleanupendpad %cleanup unwind label %terminate
5627 cleanupendpad %cleanup unwind to caller
5631 '``cleanupret``' Instruction
5632 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5639 cleanupret <value> unwind label <continue>
5640 cleanupret <value> unwind to caller
5645 The '``cleanupret``' instruction is a terminator instruction that has
5646 an optional successor.
5652 The '``cleanupret``' instruction requires one argument, which indicates
5653 which ``cleanuppad`` it exits, and must be a :ref:`cleanuppad <i_cleanuppad>`.
5654 It also has an optional successor, ``continue``.
5659 The '``cleanupret``' instruction indicates to the
5660 :ref:`personality function <personalityfn>` that one
5661 :ref:`cleanuppad <i_cleanuppad>` it transferred control to has ended.
5662 It transfers control to ``continue`` or unwinds out of the function.
5664 It is undefined behavior to execute a ``cleanupret`` whose ``cleanuppad`` has
5667 It is undefined behavior to execute a ``cleanupret`` if, after the most recent
5668 execution of its ``cleanuppad``, some ``cleanupret`` or ``cleanupendpad``
5669 consuming the same ``cleanuppad`` has already been executed.
5671 It is undefined behavior to execute a ``cleanupret`` if, after the most recent
5672 execution of its ``cleanuppad``, any other ``cleanuppad`` or ``catchpad`` has
5673 been executed but has not had a corresponding
5674 ``cleanupret``/``catchret``/``cleanupendpad``/``catchendpad`` executed.
5679 .. code-block:: llvm
5681 cleanupret %cleanup unwind to caller
5682 cleanupret %cleanup unwind label %continue
5686 '``terminatepad``' Instruction
5687 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5694 terminatepad [<args>*] unwind label <exception label>
5695 terminatepad [<args>*] unwind to caller
5700 The '``terminatepad``' instruction is used by `LLVM's exception handling
5701 system <ExceptionHandling.html#overview>`_ to specify that a basic block
5702 is a terminate block --- one where a personality routine may decide to
5703 terminate the program.
5704 The ``args`` correspond to whatever information the personality
5705 routine requires to know if this is an appropriate place to terminate the
5706 program. Control is transferred to the ``exception`` label if the
5707 personality routine decides not to terminate the program for the
5708 in-flight exception.
5713 The instruction takes a list of arbitrary values which are interpreted
5714 by the :ref:`personality function <personalityfn>`.
5716 The ``terminatepad`` may be given an ``exception`` label to
5717 transfer control to if the in-flight exception matches the ``args``.
5722 When the call stack is being unwound due to an exception being thrown,
5723 the exception is compared against the ``args``. If it matches,
5724 then control is transfered to the ``exception`` basic block. Otherwise,
5725 the program is terminated via personality-specific means. Typically,
5726 the first argument to ``terminatepad`` specifies what function the
5727 personality should defer to in order to terminate the program.
5729 The ``terminatepad`` instruction has several restrictions:
5731 - A terminate block is a basic block which is the unwind destination of
5732 an exceptional instruction.
5733 - A terminate block must have a '``terminatepad``' instruction as its
5734 first non-PHI instruction.
5735 - There can be only one '``terminatepad``' instruction within the
5737 - A basic block that is not a terminate block may not include a
5738 '``terminatepad``' instruction.
5743 .. code-block:: llvm
5745 ;; A terminate block which only permits integers.
5746 terminatepad [i8** @_ZTIi] unwind label %continue
5750 '``unreachable``' Instruction
5751 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5763 The '``unreachable``' instruction has no defined semantics. This
5764 instruction is used to inform the optimizer that a particular portion of
5765 the code is not reachable. This can be used to indicate that the code
5766 after a no-return function cannot be reached, and other facts.
5771 The '``unreachable``' instruction has no defined semantics.
5778 Binary operators are used to do most of the computation in a program.
5779 They require two operands of the same type, execute an operation on
5780 them, and produce a single value. The operands might represent multiple
5781 data, as is the case with the :ref:`vector <t_vector>` data type. The
5782 result value has the same type as its operands.
5784 There are several different binary operators:
5788 '``add``' Instruction
5789 ^^^^^^^^^^^^^^^^^^^^^
5796 <result> = add <ty> <op1>, <op2> ; yields ty:result
5797 <result> = add nuw <ty> <op1>, <op2> ; yields ty:result
5798 <result> = add nsw <ty> <op1>, <op2> ; yields ty:result
5799 <result> = add nuw nsw <ty> <op1>, <op2> ; yields ty:result
5804 The '``add``' instruction returns the sum of its two operands.
5809 The two arguments to the '``add``' instruction must be
5810 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5811 arguments must have identical types.
5816 The value produced is the integer sum of the two operands.
5818 If the sum has unsigned overflow, the result returned is the
5819 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
5822 Because LLVM integers use a two's complement representation, this
5823 instruction is appropriate for both signed and unsigned integers.
5825 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
5826 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
5827 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
5828 unsigned and/or signed overflow, respectively, occurs.
5833 .. code-block:: llvm
5835 <result> = add i32 4, %var ; yields i32:result = 4 + %var
5839 '``fadd``' Instruction
5840 ^^^^^^^^^^^^^^^^^^^^^^
5847 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
5852 The '``fadd``' instruction returns the sum of its two operands.
5857 The two arguments to the '``fadd``' instruction must be :ref:`floating
5858 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
5859 Both arguments must have identical types.
5864 The value produced is the floating point sum of the two operands. This
5865 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
5866 which are optimization hints to enable otherwise unsafe floating point
5872 .. code-block:: llvm
5874 <result> = fadd float 4.0, %var ; yields float:result = 4.0 + %var
5876 '``sub``' Instruction
5877 ^^^^^^^^^^^^^^^^^^^^^
5884 <result> = sub <ty> <op1>, <op2> ; yields ty:result
5885 <result> = sub nuw <ty> <op1>, <op2> ; yields ty:result
5886 <result> = sub nsw <ty> <op1>, <op2> ; yields ty:result
5887 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields ty:result
5892 The '``sub``' instruction returns the difference of its two operands.
5894 Note that the '``sub``' instruction is used to represent the '``neg``'
5895 instruction present in most other intermediate representations.
5900 The two arguments to the '``sub``' instruction must be
5901 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5902 arguments must have identical types.
5907 The value produced is the integer difference of the two operands.
5909 If the difference has unsigned overflow, the result returned is the
5910 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
5913 Because LLVM integers use a two's complement representation, this
5914 instruction is appropriate for both signed and unsigned integers.
5916 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
5917 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
5918 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
5919 unsigned and/or signed overflow, respectively, occurs.
5924 .. code-block:: llvm
5926 <result> = sub i32 4, %var ; yields i32:result = 4 - %var
5927 <result> = sub i32 0, %val ; yields i32:result = -%var
5931 '``fsub``' Instruction
5932 ^^^^^^^^^^^^^^^^^^^^^^
5939 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
5944 The '``fsub``' instruction returns the difference of its two operands.
5946 Note that the '``fsub``' instruction is used to represent the '``fneg``'
5947 instruction present in most other intermediate representations.
5952 The two arguments to the '``fsub``' instruction must be :ref:`floating
5953 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
5954 Both arguments must have identical types.
5959 The value produced is the floating point difference of the two operands.
5960 This instruction can also take any number of :ref:`fast-math
5961 flags <fastmath>`, which are optimization hints to enable otherwise
5962 unsafe floating point optimizations:
5967 .. code-block:: llvm
5969 <result> = fsub float 4.0, %var ; yields float:result = 4.0 - %var
5970 <result> = fsub float -0.0, %val ; yields float:result = -%var
5972 '``mul``' Instruction
5973 ^^^^^^^^^^^^^^^^^^^^^
5980 <result> = mul <ty> <op1>, <op2> ; yields ty:result
5981 <result> = mul nuw <ty> <op1>, <op2> ; yields ty:result
5982 <result> = mul nsw <ty> <op1>, <op2> ; yields ty:result
5983 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields ty:result
5988 The '``mul``' instruction returns the product of its two operands.
5993 The two arguments to the '``mul``' instruction must be
5994 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5995 arguments must have identical types.
6000 The value produced is the integer product of the two operands.
6002 If the result of the multiplication has unsigned overflow, the result
6003 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
6004 bit width of the result.
6006 Because LLVM integers use a two's complement representation, and the
6007 result is the same width as the operands, this instruction returns the
6008 correct result for both signed and unsigned integers. If a full product
6009 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
6010 sign-extended or zero-extended as appropriate to the width of the full
6013 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
6014 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
6015 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
6016 unsigned and/or signed overflow, respectively, occurs.
6021 .. code-block:: llvm
6023 <result> = mul i32 4, %var ; yields i32:result = 4 * %var
6027 '``fmul``' Instruction
6028 ^^^^^^^^^^^^^^^^^^^^^^
6035 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
6040 The '``fmul``' instruction returns the product of its two operands.
6045 The two arguments to the '``fmul``' instruction must be :ref:`floating
6046 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
6047 Both arguments must have identical types.
6052 The value produced is the floating point product of the two operands.
6053 This instruction can also take any number of :ref:`fast-math
6054 flags <fastmath>`, which are optimization hints to enable otherwise
6055 unsafe floating point optimizations:
6060 .. code-block:: llvm
6062 <result> = fmul float 4.0, %var ; yields float:result = 4.0 * %var
6064 '``udiv``' Instruction
6065 ^^^^^^^^^^^^^^^^^^^^^^
6072 <result> = udiv <ty> <op1>, <op2> ; yields ty:result
6073 <result> = udiv exact <ty> <op1>, <op2> ; yields ty:result
6078 The '``udiv``' instruction returns the quotient of its two operands.
6083 The two arguments to the '``udiv``' instruction must be
6084 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6085 arguments must have identical types.
6090 The value produced is the unsigned integer quotient of the two operands.
6092 Note that unsigned integer division and signed integer division are
6093 distinct operations; for signed integer division, use '``sdiv``'.
6095 Division by zero leads to undefined behavior.
6097 If the ``exact`` keyword is present, the result value of the ``udiv`` is
6098 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
6099 such, "((a udiv exact b) mul b) == a").
6104 .. code-block:: llvm
6106 <result> = udiv i32 4, %var ; yields i32:result = 4 / %var
6108 '``sdiv``' Instruction
6109 ^^^^^^^^^^^^^^^^^^^^^^
6116 <result> = sdiv <ty> <op1>, <op2> ; yields ty:result
6117 <result> = sdiv exact <ty> <op1>, <op2> ; yields ty:result
6122 The '``sdiv``' instruction returns the quotient of its two operands.
6127 The two arguments to the '``sdiv``' instruction must be
6128 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6129 arguments must have identical types.
6134 The value produced is the signed integer quotient of the two operands
6135 rounded towards zero.
6137 Note that signed integer division and unsigned integer division are
6138 distinct operations; for unsigned integer division, use '``udiv``'.
6140 Division by zero leads to undefined behavior. Overflow also leads to
6141 undefined behavior; this is a rare case, but can occur, for example, by
6142 doing a 32-bit division of -2147483648 by -1.
6144 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
6145 a :ref:`poison value <poisonvalues>` if the result would be rounded.
6150 .. code-block:: llvm
6152 <result> = sdiv i32 4, %var ; yields i32:result = 4 / %var
6156 '``fdiv``' Instruction
6157 ^^^^^^^^^^^^^^^^^^^^^^
6164 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
6169 The '``fdiv``' instruction returns the quotient of its two operands.
6174 The two arguments to the '``fdiv``' instruction must be :ref:`floating
6175 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
6176 Both arguments must have identical types.
6181 The value produced is the floating point quotient of the two operands.
6182 This instruction can also take any number of :ref:`fast-math
6183 flags <fastmath>`, which are optimization hints to enable otherwise
6184 unsafe floating point optimizations:
6189 .. code-block:: llvm
6191 <result> = fdiv float 4.0, %var ; yields float:result = 4.0 / %var
6193 '``urem``' Instruction
6194 ^^^^^^^^^^^^^^^^^^^^^^
6201 <result> = urem <ty> <op1>, <op2> ; yields ty:result
6206 The '``urem``' instruction returns the remainder from the unsigned
6207 division of its two arguments.
6212 The two arguments to the '``urem``' instruction must be
6213 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6214 arguments must have identical types.
6219 This instruction returns the unsigned integer *remainder* of a division.
6220 This instruction always performs an unsigned division to get the
6223 Note that unsigned integer remainder and signed integer remainder are
6224 distinct operations; for signed integer remainder, use '``srem``'.
6226 Taking the remainder of a division by zero leads to undefined behavior.
6231 .. code-block:: llvm
6233 <result> = urem i32 4, %var ; yields i32:result = 4 % %var
6235 '``srem``' Instruction
6236 ^^^^^^^^^^^^^^^^^^^^^^
6243 <result> = srem <ty> <op1>, <op2> ; yields ty:result
6248 The '``srem``' instruction returns the remainder from the signed
6249 division of its two operands. This instruction can also take
6250 :ref:`vector <t_vector>` versions of the values in which case the elements
6256 The two arguments to the '``srem``' instruction must be
6257 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6258 arguments must have identical types.
6263 This instruction returns the *remainder* of a division (where the result
6264 is either zero or has the same sign as the dividend, ``op1``), not the
6265 *modulo* operator (where the result is either zero or has the same sign
6266 as the divisor, ``op2``) of a value. For more information about the
6267 difference, see `The Math
6268 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
6269 table of how this is implemented in various languages, please see
6271 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
6273 Note that signed integer remainder and unsigned integer remainder are
6274 distinct operations; for unsigned integer remainder, use '``urem``'.
6276 Taking the remainder of a division by zero leads to undefined behavior.
6277 Overflow also leads to undefined behavior; this is a rare case, but can
6278 occur, for example, by taking the remainder of a 32-bit division of
6279 -2147483648 by -1. (The remainder doesn't actually overflow, but this
6280 rule lets srem be implemented using instructions that return both the
6281 result of the division and the remainder.)
6286 .. code-block:: llvm
6288 <result> = srem i32 4, %var ; yields i32:result = 4 % %var
6292 '``frem``' Instruction
6293 ^^^^^^^^^^^^^^^^^^^^^^
6300 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
6305 The '``frem``' instruction returns the remainder from the division of
6311 The two arguments to the '``frem``' instruction must be :ref:`floating
6312 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
6313 Both arguments must have identical types.
6318 This instruction returns the *remainder* of a division. The remainder
6319 has the same sign as the dividend. This instruction can also take any
6320 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
6321 to enable otherwise unsafe floating point optimizations:
6326 .. code-block:: llvm
6328 <result> = frem float 4.0, %var ; yields float:result = 4.0 % %var
6332 Bitwise Binary Operations
6333 -------------------------
6335 Bitwise binary operators are used to do various forms of bit-twiddling
6336 in a program. They are generally very efficient instructions and can
6337 commonly be strength reduced from other instructions. They require two
6338 operands of the same type, execute an operation on them, and produce a
6339 single value. The resulting value is the same type as its operands.
6341 '``shl``' Instruction
6342 ^^^^^^^^^^^^^^^^^^^^^
6349 <result> = shl <ty> <op1>, <op2> ; yields ty:result
6350 <result> = shl nuw <ty> <op1>, <op2> ; yields ty:result
6351 <result> = shl nsw <ty> <op1>, <op2> ; yields ty:result
6352 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields ty:result
6357 The '``shl``' instruction returns the first operand shifted to the left
6358 a specified number of bits.
6363 Both arguments to the '``shl``' instruction must be the same
6364 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
6365 '``op2``' is treated as an unsigned value.
6370 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
6371 where ``n`` is the width of the result. If ``op2`` is (statically or
6372 dynamically) equal to or larger than the number of bits in
6373 ``op1``, the result is undefined. If the arguments are vectors, each
6374 vector element of ``op1`` is shifted by the corresponding shift amount
6377 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
6378 value <poisonvalues>` if it shifts out any non-zero bits. If the
6379 ``nsw`` keyword is present, then the shift produces a :ref:`poison
6380 value <poisonvalues>` if it shifts out any bits that disagree with the
6381 resultant sign bit. As such, NUW/NSW have the same semantics as they
6382 would if the shift were expressed as a mul instruction with the same
6383 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
6388 .. code-block:: llvm
6390 <result> = shl i32 4, %var ; yields i32: 4 << %var
6391 <result> = shl i32 4, 2 ; yields i32: 16
6392 <result> = shl i32 1, 10 ; yields i32: 1024
6393 <result> = shl i32 1, 32 ; undefined
6394 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
6396 '``lshr``' Instruction
6397 ^^^^^^^^^^^^^^^^^^^^^^
6404 <result> = lshr <ty> <op1>, <op2> ; yields ty:result
6405 <result> = lshr exact <ty> <op1>, <op2> ; yields ty:result
6410 The '``lshr``' instruction (logical shift right) returns the first
6411 operand shifted to the right a specified number of bits with zero fill.
6416 Both arguments to the '``lshr``' instruction must be the same
6417 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
6418 '``op2``' is treated as an unsigned value.
6423 This instruction always performs a logical shift right operation. The
6424 most significant bits of the result will be filled with zero bits after
6425 the shift. If ``op2`` is (statically or dynamically) equal to or larger
6426 than the number of bits in ``op1``, the result is undefined. If the
6427 arguments are vectors, each vector element of ``op1`` is shifted by the
6428 corresponding shift amount in ``op2``.
6430 If the ``exact`` keyword is present, the result value of the ``lshr`` is
6431 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
6437 .. code-block:: llvm
6439 <result> = lshr i32 4, 1 ; yields i32:result = 2
6440 <result> = lshr i32 4, 2 ; yields i32:result = 1
6441 <result> = lshr i8 4, 3 ; yields i8:result = 0
6442 <result> = lshr i8 -2, 1 ; yields i8:result = 0x7F
6443 <result> = lshr i32 1, 32 ; undefined
6444 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
6446 '``ashr``' Instruction
6447 ^^^^^^^^^^^^^^^^^^^^^^
6454 <result> = ashr <ty> <op1>, <op2> ; yields ty:result
6455 <result> = ashr exact <ty> <op1>, <op2> ; yields ty:result
6460 The '``ashr``' instruction (arithmetic shift right) returns the first
6461 operand shifted to the right a specified number of bits with sign
6467 Both arguments to the '``ashr``' instruction must be the same
6468 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
6469 '``op2``' is treated as an unsigned value.
6474 This instruction always performs an arithmetic shift right operation,
6475 The most significant bits of the result will be filled with the sign bit
6476 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
6477 than the number of bits in ``op1``, the result is undefined. If the
6478 arguments are vectors, each vector element of ``op1`` is shifted by the
6479 corresponding shift amount in ``op2``.
6481 If the ``exact`` keyword is present, the result value of the ``ashr`` is
6482 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
6488 .. code-block:: llvm
6490 <result> = ashr i32 4, 1 ; yields i32:result = 2
6491 <result> = ashr i32 4, 2 ; yields i32:result = 1
6492 <result> = ashr i8 4, 3 ; yields i8:result = 0
6493 <result> = ashr i8 -2, 1 ; yields i8:result = -1
6494 <result> = ashr i32 1, 32 ; undefined
6495 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
6497 '``and``' Instruction
6498 ^^^^^^^^^^^^^^^^^^^^^
6505 <result> = and <ty> <op1>, <op2> ; yields ty:result
6510 The '``and``' instruction returns the bitwise logical and of its two
6516 The two arguments to the '``and``' instruction must be
6517 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6518 arguments must have identical types.
6523 The truth table used for the '``and``' instruction is:
6540 .. code-block:: llvm
6542 <result> = and i32 4, %var ; yields i32:result = 4 & %var
6543 <result> = and i32 15, 40 ; yields i32:result = 8
6544 <result> = and i32 4, 8 ; yields i32:result = 0
6546 '``or``' Instruction
6547 ^^^^^^^^^^^^^^^^^^^^
6554 <result> = or <ty> <op1>, <op2> ; yields ty:result
6559 The '``or``' instruction returns the bitwise logical inclusive or of its
6565 The two arguments to the '``or``' instruction must be
6566 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6567 arguments must have identical types.
6572 The truth table used for the '``or``' instruction is:
6591 <result> = or i32 4, %var ; yields i32:result = 4 | %var
6592 <result> = or i32 15, 40 ; yields i32:result = 47
6593 <result> = or i32 4, 8 ; yields i32:result = 12
6595 '``xor``' Instruction
6596 ^^^^^^^^^^^^^^^^^^^^^
6603 <result> = xor <ty> <op1>, <op2> ; yields ty:result
6608 The '``xor``' instruction returns the bitwise logical exclusive or of
6609 its two operands. The ``xor`` is used to implement the "one's
6610 complement" operation, which is the "~" operator in C.
6615 The two arguments to the '``xor``' instruction must be
6616 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6617 arguments must have identical types.
6622 The truth table used for the '``xor``' instruction is:
6639 .. code-block:: llvm
6641 <result> = xor i32 4, %var ; yields i32:result = 4 ^ %var
6642 <result> = xor i32 15, 40 ; yields i32:result = 39
6643 <result> = xor i32 4, 8 ; yields i32:result = 12
6644 <result> = xor i32 %V, -1 ; yields i32:result = ~%V
6649 LLVM supports several instructions to represent vector operations in a
6650 target-independent manner. These instructions cover the element-access
6651 and vector-specific operations needed to process vectors effectively.
6652 While LLVM does directly support these vector operations, many
6653 sophisticated algorithms will want to use target-specific intrinsics to
6654 take full advantage of a specific target.
6656 .. _i_extractelement:
6658 '``extractelement``' Instruction
6659 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6666 <result> = extractelement <n x <ty>> <val>, <ty2> <idx> ; yields <ty>
6671 The '``extractelement``' instruction extracts a single scalar element
6672 from a vector at a specified index.
6677 The first operand of an '``extractelement``' instruction is a value of
6678 :ref:`vector <t_vector>` type. The second operand is an index indicating
6679 the position from which to extract the element. The index may be a
6680 variable of any integer type.
6685 The result is a scalar of the same type as the element type of ``val``.
6686 Its value is the value at position ``idx`` of ``val``. If ``idx``
6687 exceeds the length of ``val``, the results are undefined.
6692 .. code-block:: llvm
6694 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
6696 .. _i_insertelement:
6698 '``insertelement``' Instruction
6699 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6706 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, <ty2> <idx> ; yields <n x <ty>>
6711 The '``insertelement``' instruction inserts a scalar element into a
6712 vector at a specified index.
6717 The first operand of an '``insertelement``' instruction is a value of
6718 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
6719 type must equal the element type of the first operand. The third operand
6720 is an index indicating the position at which to insert the value. The
6721 index may be a variable of any integer type.
6726 The result is a vector of the same type as ``val``. Its element values
6727 are those of ``val`` except at position ``idx``, where it gets the value
6728 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
6734 .. code-block:: llvm
6736 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
6738 .. _i_shufflevector:
6740 '``shufflevector``' Instruction
6741 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6748 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
6753 The '``shufflevector``' instruction constructs a permutation of elements
6754 from two input vectors, returning a vector with the same element type as
6755 the input and length that is the same as the shuffle mask.
6760 The first two operands of a '``shufflevector``' instruction are vectors
6761 with the same type. The third argument is a shuffle mask whose element
6762 type is always 'i32'. The result of the instruction is a vector whose
6763 length is the same as the shuffle mask and whose element type is the
6764 same as the element type of the first two operands.
6766 The shuffle mask operand is required to be a constant vector with either
6767 constant integer or undef values.
6772 The elements of the two input vectors are numbered from left to right
6773 across both of the vectors. The shuffle mask operand specifies, for each
6774 element of the result vector, which element of the two input vectors the
6775 result element gets. The element selector may be undef (meaning "don't
6776 care") and the second operand may be undef if performing a shuffle from
6782 .. code-block:: llvm
6784 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
6785 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
6786 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
6787 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
6788 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
6789 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
6790 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
6791 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
6793 Aggregate Operations
6794 --------------------
6796 LLVM supports several instructions for working with
6797 :ref:`aggregate <t_aggregate>` values.
6801 '``extractvalue``' Instruction
6802 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6809 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
6814 The '``extractvalue``' instruction extracts the value of a member field
6815 from an :ref:`aggregate <t_aggregate>` value.
6820 The first operand of an '``extractvalue``' instruction is a value of
6821 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The other operands are
6822 constant indices to specify which value to extract in a similar manner
6823 as indices in a '``getelementptr``' instruction.
6825 The major differences to ``getelementptr`` indexing are:
6827 - Since the value being indexed is not a pointer, the first index is
6828 omitted and assumed to be zero.
6829 - At least one index must be specified.
6830 - Not only struct indices but also array indices must be in bounds.
6835 The result is the value at the position in the aggregate specified by
6841 .. code-block:: llvm
6843 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
6847 '``insertvalue``' Instruction
6848 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6855 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
6860 The '``insertvalue``' instruction inserts a value into a member field in
6861 an :ref:`aggregate <t_aggregate>` value.
6866 The first operand of an '``insertvalue``' instruction is a value of
6867 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
6868 a first-class value to insert. The following operands are constant
6869 indices indicating the position at which to insert the value in a
6870 similar manner as indices in a '``extractvalue``' instruction. The value
6871 to insert must have the same type as the value identified by the
6877 The result is an aggregate of the same type as ``val``. Its value is
6878 that of ``val`` except that the value at the position specified by the
6879 indices is that of ``elt``.
6884 .. code-block:: llvm
6886 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
6887 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
6888 %agg3 = insertvalue {i32, {float}} undef, float %val, 1, 0 ; yields {i32 undef, {float %val}}
6892 Memory Access and Addressing Operations
6893 ---------------------------------------
6895 A key design point of an SSA-based representation is how it represents
6896 memory. In LLVM, no memory locations are in SSA form, which makes things
6897 very simple. This section describes how to read, write, and allocate
6902 '``alloca``' Instruction
6903 ^^^^^^^^^^^^^^^^^^^^^^^^
6910 <result> = alloca [inalloca] <type> [, <ty> <NumElements>] [, align <alignment>] ; yields type*:result
6915 The '``alloca``' instruction allocates memory on the stack frame of the
6916 currently executing function, to be automatically released when this
6917 function returns to its caller. The object is always allocated in the
6918 generic address space (address space zero).
6923 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
6924 bytes of memory on the runtime stack, returning a pointer of the
6925 appropriate type to the program. If "NumElements" is specified, it is
6926 the number of elements allocated, otherwise "NumElements" is defaulted
6927 to be one. If a constant alignment is specified, the value result of the
6928 allocation is guaranteed to be aligned to at least that boundary. The
6929 alignment may not be greater than ``1 << 29``. If not specified, or if
6930 zero, the target can choose to align the allocation on any convenient
6931 boundary compatible with the type.
6933 '``type``' may be any sized type.
6938 Memory is allocated; a pointer is returned. The operation is undefined
6939 if there is insufficient stack space for the allocation. '``alloca``'d
6940 memory is automatically released when the function returns. The
6941 '``alloca``' instruction is commonly used to represent automatic
6942 variables that must have an address available. When the function returns
6943 (either with the ``ret`` or ``resume`` instructions), the memory is
6944 reclaimed. Allocating zero bytes is legal, but the result is undefined.
6945 The order in which memory is allocated (ie., which way the stack grows)
6951 .. code-block:: llvm
6953 %ptr = alloca i32 ; yields i32*:ptr
6954 %ptr = alloca i32, i32 4 ; yields i32*:ptr
6955 %ptr = alloca i32, i32 4, align 1024 ; yields i32*:ptr
6956 %ptr = alloca i32, align 1024 ; yields i32*:ptr
6960 '``load``' Instruction
6961 ^^^^^^^^^^^^^^^^^^^^^^
6968 <result> = load [volatile] <ty>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>][, !invariant.group !<index>][, !nonnull !<index>][, !dereferenceable !<deref_bytes_node>][, !dereferenceable_or_null !<deref_bytes_node>][, !align !<align_node>]
6969 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment> [, !invariant.group !<index>]
6970 !<index> = !{ i32 1 }
6971 !<deref_bytes_node> = !{i64 <dereferenceable_bytes>}
6972 !<align_node> = !{ i64 <value_alignment> }
6977 The '``load``' instruction is used to read from memory.
6982 The argument to the ``load`` instruction specifies the memory address
6983 from which to load. The type specified must be a :ref:`first
6984 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
6985 then the optimizer is not allowed to modify the number or order of
6986 execution of this ``load`` with other :ref:`volatile
6987 operations <volatile>`.
6989 If the ``load`` is marked as ``atomic``, it takes an extra
6990 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
6991 ``release`` and ``acq_rel`` orderings are not valid on ``load``
6992 instructions. Atomic loads produce :ref:`defined <memmodel>` results
6993 when they may see multiple atomic stores. The type of the pointee must
6994 be an integer type whose bit width is a power of two greater than or
6995 equal to eight and less than or equal to a target-specific size limit.
6996 ``align`` must be explicitly specified on atomic loads, and the load has
6997 undefined behavior if the alignment is not set to a value which is at
6998 least the size in bytes of the pointee. ``!nontemporal`` does not have
6999 any defined semantics for atomic loads.
7001 The optional constant ``align`` argument specifies the alignment of the
7002 operation (that is, the alignment of the memory address). A value of 0
7003 or an omitted ``align`` argument means that the operation has the ABI
7004 alignment for the target. It is the responsibility of the code emitter
7005 to ensure that the alignment information is correct. Overestimating the
7006 alignment results in undefined behavior. Underestimating the alignment
7007 may produce less efficient code. An alignment of 1 is always safe. The
7008 maximum possible alignment is ``1 << 29``.
7010 The optional ``!nontemporal`` metadata must reference a single
7011 metadata name ``<index>`` corresponding to a metadata node with one
7012 ``i32`` entry of value 1. The existence of the ``!nontemporal``
7013 metadata on the instruction tells the optimizer and code generator
7014 that this load is not expected to be reused in the cache. The code
7015 generator may select special instructions to save cache bandwidth, such
7016 as the ``MOVNT`` instruction on x86.
7018 The optional ``!invariant.load`` metadata must reference a single
7019 metadata name ``<index>`` corresponding to a metadata node with no
7020 entries. The existence of the ``!invariant.load`` metadata on the
7021 instruction tells the optimizer and code generator that the address
7022 operand to this load points to memory which can be assumed unchanged.
7023 Being invariant does not imply that a location is dereferenceable,
7024 but it does imply that once the location is known dereferenceable
7025 its value is henceforth unchanging.
7027 The optional ``!invariant.group`` metadata must reference a single metadata name
7028 ``<index>`` corresponding to a metadata node. See ``invariant.group`` metadata.
7030 The optional ``!nonnull`` metadata must reference a single
7031 metadata name ``<index>`` corresponding to a metadata node with no
7032 entries. The existence of the ``!nonnull`` metadata on the
7033 instruction tells the optimizer that the value loaded is known to
7034 never be null. This is analogous to the ``nonnull`` attribute
7035 on parameters and return values. This metadata can only be applied
7036 to loads of a pointer type.
7038 The optional ``!dereferenceable`` metadata must reference a single metadata
7039 name ``<deref_bytes_node>`` corresponding to a metadata node with one ``i64``
7040 entry. The existence of the ``!dereferenceable`` metadata on the instruction
7041 tells the optimizer that the value loaded is known to be dereferenceable.
7042 The number of bytes known to be dereferenceable is specified by the integer
7043 value in the metadata node. This is analogous to the ''dereferenceable''
7044 attribute on parameters and return values. This metadata can only be applied
7045 to loads of a pointer type.
7047 The optional ``!dereferenceable_or_null`` metadata must reference a single
7048 metadata name ``<deref_bytes_node>`` corresponding to a metadata node with one
7049 ``i64`` entry. The existence of the ``!dereferenceable_or_null`` metadata on the
7050 instruction tells the optimizer that the value loaded is known to be either
7051 dereferenceable or null.
7052 The number of bytes known to be dereferenceable is specified by the integer
7053 value in the metadata node. This is analogous to the ''dereferenceable_or_null''
7054 attribute on parameters and return values. This metadata can only be applied
7055 to loads of a pointer type.
7057 The optional ``!align`` metadata must reference a single metadata name
7058 ``<align_node>`` corresponding to a metadata node with one ``i64`` entry.
7059 The existence of the ``!align`` metadata on the instruction tells the
7060 optimizer that the value loaded is known to be aligned to a boundary specified
7061 by the integer value in the metadata node. The alignment must be a power of 2.
7062 This is analogous to the ''align'' attribute on parameters and return values.
7063 This metadata can only be applied to loads of a pointer type.
7068 The location of memory pointed to is loaded. If the value being loaded
7069 is of scalar type then the number of bytes read does not exceed the
7070 minimum number of bytes needed to hold all bits of the type. For
7071 example, loading an ``i24`` reads at most three bytes. When loading a
7072 value of a type like ``i20`` with a size that is not an integral number
7073 of bytes, the result is undefined if the value was not originally
7074 written using a store of the same type.
7079 .. code-block:: llvm
7081 %ptr = alloca i32 ; yields i32*:ptr
7082 store i32 3, i32* %ptr ; yields void
7083 %val = load i32, i32* %ptr ; yields i32:val = i32 3
7087 '``store``' Instruction
7088 ^^^^^^^^^^^^^^^^^^^^^^^
7095 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.group !<index>] ; yields void
7096 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> [, !invariant.group !<index>] ; yields void
7101 The '``store``' instruction is used to write to memory.
7106 There are two arguments to the ``store`` instruction: a value to store
7107 and an address at which to store it. The type of the ``<pointer>``
7108 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
7109 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
7110 then the optimizer is not allowed to modify the number or order of
7111 execution of this ``store`` with other :ref:`volatile
7112 operations <volatile>`.
7114 If the ``store`` is marked as ``atomic``, it takes an extra
7115 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
7116 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
7117 instructions. Atomic loads produce :ref:`defined <memmodel>` results
7118 when they may see multiple atomic stores. The type of the pointee must
7119 be an integer type whose bit width is a power of two greater than or
7120 equal to eight and less than or equal to a target-specific size limit.
7121 ``align`` must be explicitly specified on atomic stores, and the store
7122 has undefined behavior if the alignment is not set to a value which is
7123 at least the size in bytes of the pointee. ``!nontemporal`` does not
7124 have any defined semantics for atomic stores.
7126 The optional constant ``align`` argument specifies the alignment of the
7127 operation (that is, the alignment of the memory address). A value of 0
7128 or an omitted ``align`` argument means that the operation has the ABI
7129 alignment for the target. It is the responsibility of the code emitter
7130 to ensure that the alignment information is correct. Overestimating the
7131 alignment results in undefined behavior. Underestimating the
7132 alignment may produce less efficient code. An alignment of 1 is always
7133 safe. The maximum possible alignment is ``1 << 29``.
7135 The optional ``!nontemporal`` metadata must reference a single metadata
7136 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
7137 value 1. The existence of the ``!nontemporal`` metadata on the instruction
7138 tells the optimizer and code generator that this load is not expected to
7139 be reused in the cache. The code generator may select special
7140 instructions to save cache bandwidth, such as the MOVNT instruction on
7143 The optional ``!invariant.group`` metadata must reference a
7144 single metadata name ``<index>``. See ``invariant.group`` metadata.
7149 The contents of memory are updated to contain ``<value>`` at the
7150 location specified by the ``<pointer>`` operand. If ``<value>`` is
7151 of scalar type then the number of bytes written does not exceed the
7152 minimum number of bytes needed to hold all bits of the type. For
7153 example, storing an ``i24`` writes at most three bytes. When writing a
7154 value of a type like ``i20`` with a size that is not an integral number
7155 of bytes, it is unspecified what happens to the extra bits that do not
7156 belong to the type, but they will typically be overwritten.
7161 .. code-block:: llvm
7163 %ptr = alloca i32 ; yields i32*:ptr
7164 store i32 3, i32* %ptr ; yields void
7165 %val = load i32, i32* %ptr ; yields i32:val = i32 3
7169 '``fence``' Instruction
7170 ^^^^^^^^^^^^^^^^^^^^^^^
7177 fence [singlethread] <ordering> ; yields void
7182 The '``fence``' instruction is used to introduce happens-before edges
7188 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
7189 defines what *synchronizes-with* edges they add. They can only be given
7190 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
7195 A fence A which has (at least) ``release`` ordering semantics
7196 *synchronizes with* a fence B with (at least) ``acquire`` ordering
7197 semantics if and only if there exist atomic operations X and Y, both
7198 operating on some atomic object M, such that A is sequenced before X, X
7199 modifies M (either directly or through some side effect of a sequence
7200 headed by X), Y is sequenced before B, and Y observes M. This provides a
7201 *happens-before* dependency between A and B. Rather than an explicit
7202 ``fence``, one (but not both) of the atomic operations X or Y might
7203 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
7204 still *synchronize-with* the explicit ``fence`` and establish the
7205 *happens-before* edge.
7207 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
7208 ``acquire`` and ``release`` semantics specified above, participates in
7209 the global program order of other ``seq_cst`` operations and/or fences.
7211 The optional ":ref:`singlethread <singlethread>`" argument specifies
7212 that the fence only synchronizes with other fences in the same thread.
7213 (This is useful for interacting with signal handlers.)
7218 .. code-block:: llvm
7220 fence acquire ; yields void
7221 fence singlethread seq_cst ; yields void
7225 '``cmpxchg``' Instruction
7226 ^^^^^^^^^^^^^^^^^^^^^^^^^
7233 cmpxchg [weak] [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <success ordering> <failure ordering> ; yields { ty, i1 }
7238 The '``cmpxchg``' instruction is used to atomically modify memory. It
7239 loads a value in memory and compares it to a given value. If they are
7240 equal, it tries to store a new value into the memory.
7245 There are three arguments to the '``cmpxchg``' instruction: an address
7246 to operate on, a value to compare to the value currently be at that
7247 address, and a new value to place at that address if the compared values
7248 are equal. The type of '<cmp>' must be an integer type whose bit width
7249 is a power of two greater than or equal to eight and less than or equal
7250 to a target-specific size limit. '<cmp>' and '<new>' must have the same
7251 type, and the type of '<pointer>' must be a pointer to that type. If the
7252 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
7253 to modify the number or order of execution of this ``cmpxchg`` with
7254 other :ref:`volatile operations <volatile>`.
7256 The success and failure :ref:`ordering <ordering>` arguments specify how this
7257 ``cmpxchg`` synchronizes with other atomic operations. Both ordering parameters
7258 must be at least ``monotonic``, the ordering constraint on failure must be no
7259 stronger than that on success, and the failure ordering cannot be either
7260 ``release`` or ``acq_rel``.
7262 The optional "``singlethread``" argument declares that the ``cmpxchg``
7263 is only atomic with respect to code (usually signal handlers) running in
7264 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
7265 respect to all other code in the system.
7267 The pointer passed into cmpxchg must have alignment greater than or
7268 equal to the size in memory of the operand.
7273 The contents of memory at the location specified by the '``<pointer>``' operand
7274 is read and compared to '``<cmp>``'; if the read value is the equal, the
7275 '``<new>``' is written. The original value at the location is returned, together
7276 with a flag indicating success (true) or failure (false).
7278 If the cmpxchg operation is marked as ``weak`` then a spurious failure is
7279 permitted: the operation may not write ``<new>`` even if the comparison
7282 If the cmpxchg operation is strong (the default), the i1 value is 1 if and only
7283 if the value loaded equals ``cmp``.
7285 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose of
7286 identifying release sequences. A failed ``cmpxchg`` is equivalent to an atomic
7287 load with an ordering parameter determined the second ordering parameter.
7292 .. code-block:: llvm
7295 %orig = atomic load i32, i32* %ptr unordered ; yields i32
7299 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
7300 %squared = mul i32 %cmp, %cmp
7301 %val_success = cmpxchg i32* %ptr, i32 %cmp, i32 %squared acq_rel monotonic ; yields { i32, i1 }
7302 %value_loaded = extractvalue { i32, i1 } %val_success, 0
7303 %success = extractvalue { i32, i1 } %val_success, 1
7304 br i1 %success, label %done, label %loop
7311 '``atomicrmw``' Instruction
7312 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7319 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields ty
7324 The '``atomicrmw``' instruction is used to atomically modify memory.
7329 There are three arguments to the '``atomicrmw``' instruction: an
7330 operation to apply, an address whose value to modify, an argument to the
7331 operation. The operation must be one of the following keywords:
7345 The type of '<value>' must be an integer type whose bit width is a power
7346 of two greater than or equal to eight and less than or equal to a
7347 target-specific size limit. The type of the '``<pointer>``' operand must
7348 be a pointer to that type. If the ``atomicrmw`` is marked as
7349 ``volatile``, then the optimizer is not allowed to modify the number or
7350 order of execution of this ``atomicrmw`` with other :ref:`volatile
7351 operations <volatile>`.
7356 The contents of memory at the location specified by the '``<pointer>``'
7357 operand are atomically read, modified, and written back. The original
7358 value at the location is returned. The modification is specified by the
7361 - xchg: ``*ptr = val``
7362 - add: ``*ptr = *ptr + val``
7363 - sub: ``*ptr = *ptr - val``
7364 - and: ``*ptr = *ptr & val``
7365 - nand: ``*ptr = ~(*ptr & val)``
7366 - or: ``*ptr = *ptr | val``
7367 - xor: ``*ptr = *ptr ^ val``
7368 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
7369 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
7370 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
7372 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
7378 .. code-block:: llvm
7380 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields i32
7382 .. _i_getelementptr:
7384 '``getelementptr``' Instruction
7385 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7392 <result> = getelementptr <ty>, <ty>* <ptrval>{, <ty> <idx>}*
7393 <result> = getelementptr inbounds <ty>, <ty>* <ptrval>{, <ty> <idx>}*
7394 <result> = getelementptr <ty>, <ptr vector> <ptrval>, <vector index type> <idx>
7399 The '``getelementptr``' instruction is used to get the address of a
7400 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
7401 address calculation only and does not access memory. The instruction can also
7402 be used to calculate a vector of such addresses.
7407 The first argument is always a type used as the basis for the calculations.
7408 The second argument is always a pointer or a vector of pointers, and is the
7409 base address to start from. The remaining arguments are indices
7410 that indicate which of the elements of the aggregate object are indexed.
7411 The interpretation of each index is dependent on the type being indexed
7412 into. The first index always indexes the pointer value given as the
7413 first argument, the second index indexes a value of the type pointed to
7414 (not necessarily the value directly pointed to, since the first index
7415 can be non-zero), etc. The first type indexed into must be a pointer
7416 value, subsequent types can be arrays, vectors, and structs. Note that
7417 subsequent types being indexed into can never be pointers, since that
7418 would require loading the pointer before continuing calculation.
7420 The type of each index argument depends on the type it is indexing into.
7421 When indexing into a (optionally packed) structure, only ``i32`` integer
7422 **constants** are allowed (when using a vector of indices they must all
7423 be the **same** ``i32`` integer constant). When indexing into an array,
7424 pointer or vector, integers of any width are allowed, and they are not
7425 required to be constant. These integers are treated as signed values
7428 For example, let's consider a C code fragment and how it gets compiled
7444 int *foo(struct ST *s) {
7445 return &s[1].Z.B[5][13];
7448 The LLVM code generated by Clang is:
7450 .. code-block:: llvm
7452 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
7453 %struct.ST = type { i32, double, %struct.RT }
7455 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
7457 %arrayidx = getelementptr inbounds %struct.ST, %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
7464 In the example above, the first index is indexing into the
7465 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
7466 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
7467 indexes into the third element of the structure, yielding a
7468 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
7469 structure. The third index indexes into the second element of the
7470 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
7471 dimensions of the array are subscripted into, yielding an '``i32``'
7472 type. The '``getelementptr``' instruction returns a pointer to this
7473 element, thus computing a value of '``i32*``' type.
7475 Note that it is perfectly legal to index partially through a structure,
7476 returning a pointer to an inner element. Because of this, the LLVM code
7477 for the given testcase is equivalent to:
7479 .. code-block:: llvm
7481 define i32* @foo(%struct.ST* %s) {
7482 %t1 = getelementptr %struct.ST, %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
7483 %t2 = getelementptr %struct.ST, %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
7484 %t3 = getelementptr %struct.RT, %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
7485 %t4 = getelementptr [10 x [20 x i32]], [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
7486 %t5 = getelementptr [20 x i32], [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
7490 If the ``inbounds`` keyword is present, the result value of the
7491 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
7492 pointer is not an *in bounds* address of an allocated object, or if any
7493 of the addresses that would be formed by successive addition of the
7494 offsets implied by the indices to the base address with infinitely
7495 precise signed arithmetic are not an *in bounds* address of that
7496 allocated object. The *in bounds* addresses for an allocated object are
7497 all the addresses that point into the object, plus the address one byte
7498 past the end. In cases where the base is a vector of pointers the
7499 ``inbounds`` keyword applies to each of the computations element-wise.
7501 If the ``inbounds`` keyword is not present, the offsets are added to the
7502 base address with silently-wrapping two's complement arithmetic. If the
7503 offsets have a different width from the pointer, they are sign-extended
7504 or truncated to the width of the pointer. The result value of the
7505 ``getelementptr`` may be outside the object pointed to by the base
7506 pointer. The result value may not necessarily be used to access memory
7507 though, even if it happens to point into allocated storage. See the
7508 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
7511 The getelementptr instruction is often confusing. For some more insight
7512 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
7517 .. code-block:: llvm
7519 ; yields [12 x i8]*:aptr
7520 %aptr = getelementptr {i32, [12 x i8]}, {i32, [12 x i8]}* %saptr, i64 0, i32 1
7522 %vptr = getelementptr {i32, <2 x i8>}, {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
7524 %eptr = getelementptr [12 x i8], [12 x i8]* %aptr, i64 0, i32 1
7526 %iptr = getelementptr [10 x i32], [10 x i32]* @arr, i16 0, i16 0
7531 The ``getelementptr`` returns a vector of pointers, instead of a single address,
7532 when one or more of its arguments is a vector. In such cases, all vector
7533 arguments should have the same number of elements, and every scalar argument
7534 will be effectively broadcast into a vector during address calculation.
7536 .. code-block:: llvm
7538 ; All arguments are vectors:
7539 ; A[i] = ptrs[i] + offsets[i]*sizeof(i8)
7540 %A = getelementptr i8, <4 x i8*> %ptrs, <4 x i64> %offsets
7542 ; Add the same scalar offset to each pointer of a vector:
7543 ; A[i] = ptrs[i] + offset*sizeof(i8)
7544 %A = getelementptr i8, <4 x i8*> %ptrs, i64 %offset
7546 ; Add distinct offsets to the same pointer:
7547 ; A[i] = ptr + offsets[i]*sizeof(i8)
7548 %A = getelementptr i8, i8* %ptr, <4 x i64> %offsets
7550 ; In all cases described above the type of the result is <4 x i8*>
7552 The two following instructions are equivalent:
7554 .. code-block:: llvm
7556 getelementptr %struct.ST, <4 x %struct.ST*> %s, <4 x i64> %ind1,
7557 <4 x i32> <i32 2, i32 2, i32 2, i32 2>,
7558 <4 x i32> <i32 1, i32 1, i32 1, i32 1>,
7560 <4 x i64> <i64 13, i64 13, i64 13, i64 13>
7562 getelementptr %struct.ST, <4 x %struct.ST*> %s, <4 x i64> %ind1,
7563 i32 2, i32 1, <4 x i32> %ind4, i64 13
7565 Let's look at the C code, where the vector version of ``getelementptr``
7570 // Let's assume that we vectorize the following loop:
7571 double *A, B; int *C;
7572 for (int i = 0; i < size; ++i) {
7576 .. code-block:: llvm
7578 ; get pointers for 8 elements from array B
7579 %ptrs = getelementptr double, double* %B, <8 x i32> %C
7580 ; load 8 elements from array B into A
7581 %A = call <8 x double> @llvm.masked.gather.v8f64(<8 x double*> %ptrs,
7582 i32 8, <8 x i1> %mask, <8 x double> %passthru)
7584 Conversion Operations
7585 ---------------------
7587 The instructions in this category are the conversion instructions
7588 (casting) which all take a single operand and a type. They perform
7589 various bit conversions on the operand.
7591 '``trunc .. to``' Instruction
7592 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7599 <result> = trunc <ty> <value> to <ty2> ; yields ty2
7604 The '``trunc``' instruction truncates its operand to the type ``ty2``.
7609 The '``trunc``' instruction takes a value to trunc, and a type to trunc
7610 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
7611 of the same number of integers. The bit size of the ``value`` must be
7612 larger than the bit size of the destination type, ``ty2``. Equal sized
7613 types are not allowed.
7618 The '``trunc``' instruction truncates the high order bits in ``value``
7619 and converts the remaining bits to ``ty2``. Since the source size must
7620 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
7621 It will always truncate bits.
7626 .. code-block:: llvm
7628 %X = trunc i32 257 to i8 ; yields i8:1
7629 %Y = trunc i32 123 to i1 ; yields i1:true
7630 %Z = trunc i32 122 to i1 ; yields i1:false
7631 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
7633 '``zext .. to``' Instruction
7634 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7641 <result> = zext <ty> <value> to <ty2> ; yields ty2
7646 The '``zext``' instruction zero extends its operand to type ``ty2``.
7651 The '``zext``' instruction takes a value to cast, and a type to cast it
7652 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
7653 the same number of integers. The bit size of the ``value`` must be
7654 smaller than the bit size of the destination type, ``ty2``.
7659 The ``zext`` fills the high order bits of the ``value`` with zero bits
7660 until it reaches the size of the destination type, ``ty2``.
7662 When zero extending from i1, the result will always be either 0 or 1.
7667 .. code-block:: llvm
7669 %X = zext i32 257 to i64 ; yields i64:257
7670 %Y = zext i1 true to i32 ; yields i32:1
7671 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
7673 '``sext .. to``' Instruction
7674 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7681 <result> = sext <ty> <value> to <ty2> ; yields ty2
7686 The '``sext``' sign extends ``value`` to the type ``ty2``.
7691 The '``sext``' instruction takes a value to cast, and a type to cast it
7692 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
7693 the same number of integers. The bit size of the ``value`` must be
7694 smaller than the bit size of the destination type, ``ty2``.
7699 The '``sext``' instruction performs a sign extension by copying the sign
7700 bit (highest order bit) of the ``value`` until it reaches the bit size
7701 of the type ``ty2``.
7703 When sign extending from i1, the extension always results in -1 or 0.
7708 .. code-block:: llvm
7710 %X = sext i8 -1 to i16 ; yields i16 :65535
7711 %Y = sext i1 true to i32 ; yields i32:-1
7712 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
7714 '``fptrunc .. to``' Instruction
7715 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7722 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
7727 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
7732 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
7733 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
7734 The size of ``value`` must be larger than the size of ``ty2``. This
7735 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
7740 The '``fptrunc``' instruction casts a ``value`` from a larger
7741 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
7742 point <t_floating>` type. If the value cannot fit (i.e. overflows) within the
7743 destination type, ``ty2``, then the results are undefined. If the cast produces
7744 an inexact result, how rounding is performed (e.g. truncation, also known as
7745 round to zero) is undefined.
7750 .. code-block:: llvm
7752 %X = fptrunc double 123.0 to float ; yields float:123.0
7753 %Y = fptrunc double 1.0E+300 to float ; yields undefined
7755 '``fpext .. to``' Instruction
7756 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7763 <result> = fpext <ty> <value> to <ty2> ; yields ty2
7768 The '``fpext``' extends a floating point ``value`` to a larger floating
7774 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
7775 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
7776 to. The source type must be smaller than the destination type.
7781 The '``fpext``' instruction extends the ``value`` from a smaller
7782 :ref:`floating point <t_floating>` type to a larger :ref:`floating
7783 point <t_floating>` type. The ``fpext`` cannot be used to make a
7784 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
7785 *no-op cast* for a floating point cast.
7790 .. code-block:: llvm
7792 %X = fpext float 3.125 to double ; yields double:3.125000e+00
7793 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
7795 '``fptoui .. to``' Instruction
7796 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7803 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
7808 The '``fptoui``' converts a floating point ``value`` to its unsigned
7809 integer equivalent of type ``ty2``.
7814 The '``fptoui``' instruction takes a value to cast, which must be a
7815 scalar or vector :ref:`floating point <t_floating>` value, and a type to
7816 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
7817 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
7818 type with the same number of elements as ``ty``
7823 The '``fptoui``' instruction converts its :ref:`floating
7824 point <t_floating>` operand into the nearest (rounding towards zero)
7825 unsigned integer value. If the value cannot fit in ``ty2``, the results
7831 .. code-block:: llvm
7833 %X = fptoui double 123.0 to i32 ; yields i32:123
7834 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
7835 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
7837 '``fptosi .. to``' Instruction
7838 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7845 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
7850 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
7851 ``value`` to type ``ty2``.
7856 The '``fptosi``' instruction takes a value to cast, which must be a
7857 scalar or vector :ref:`floating point <t_floating>` value, and a type to
7858 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
7859 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
7860 type with the same number of elements as ``ty``
7865 The '``fptosi``' instruction converts its :ref:`floating
7866 point <t_floating>` operand into the nearest (rounding towards zero)
7867 signed integer value. If the value cannot fit in ``ty2``, the results
7873 .. code-block:: llvm
7875 %X = fptosi double -123.0 to i32 ; yields i32:-123
7876 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
7877 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
7879 '``uitofp .. to``' Instruction
7880 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7887 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
7892 The '``uitofp``' instruction regards ``value`` as an unsigned integer
7893 and converts that value to the ``ty2`` type.
7898 The '``uitofp``' instruction takes a value to cast, which must be a
7899 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
7900 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
7901 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
7902 type with the same number of elements as ``ty``
7907 The '``uitofp``' instruction interprets its operand as an unsigned
7908 integer quantity and converts it to the corresponding floating point
7909 value. If the value cannot fit in the floating point value, the results
7915 .. code-block:: llvm
7917 %X = uitofp i32 257 to float ; yields float:257.0
7918 %Y = uitofp i8 -1 to double ; yields double:255.0
7920 '``sitofp .. to``' Instruction
7921 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7928 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
7933 The '``sitofp``' instruction regards ``value`` as a signed integer and
7934 converts that value to the ``ty2`` type.
7939 The '``sitofp``' instruction takes a value to cast, which must be a
7940 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
7941 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
7942 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
7943 type with the same number of elements as ``ty``
7948 The '``sitofp``' instruction interprets its operand as a signed integer
7949 quantity and converts it to the corresponding floating point value. If
7950 the value cannot fit in the floating point value, the results are
7956 .. code-block:: llvm
7958 %X = sitofp i32 257 to float ; yields float:257.0
7959 %Y = sitofp i8 -1 to double ; yields double:-1.0
7963 '``ptrtoint .. to``' Instruction
7964 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7971 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
7976 The '``ptrtoint``' instruction converts the pointer or a vector of
7977 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
7982 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
7983 a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
7984 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
7985 a vector of integers type.
7990 The '``ptrtoint``' instruction converts ``value`` to integer type
7991 ``ty2`` by interpreting the pointer value as an integer and either
7992 truncating or zero extending that value to the size of the integer type.
7993 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
7994 ``value`` is larger than ``ty2`` then a truncation is done. If they are
7995 the same size, then nothing is done (*no-op cast*) other than a type
8001 .. code-block:: llvm
8003 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
8004 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
8005 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
8009 '``inttoptr .. to``' Instruction
8010 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8017 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
8022 The '``inttoptr``' instruction converts an integer ``value`` to a
8023 pointer type, ``ty2``.
8028 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
8029 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
8035 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
8036 applying either a zero extension or a truncation depending on the size
8037 of the integer ``value``. If ``value`` is larger than the size of a
8038 pointer then a truncation is done. If ``value`` is smaller than the size
8039 of a pointer then a zero extension is done. If they are the same size,
8040 nothing is done (*no-op cast*).
8045 .. code-block:: llvm
8047 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
8048 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
8049 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
8050 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
8054 '``bitcast .. to``' Instruction
8055 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8062 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
8067 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
8073 The '``bitcast``' instruction takes a value to cast, which must be a
8074 non-aggregate first class value, and a type to cast it to, which must
8075 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
8076 bit sizes of ``value`` and the destination type, ``ty2``, must be
8077 identical. If the source type is a pointer, the destination type must
8078 also be a pointer of the same size. This instruction supports bitwise
8079 conversion of vectors to integers and to vectors of other types (as
8080 long as they have the same size).
8085 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
8086 is always a *no-op cast* because no bits change with this
8087 conversion. The conversion is done as if the ``value`` had been stored
8088 to memory and read back as type ``ty2``. Pointer (or vector of
8089 pointers) types may only be converted to other pointer (or vector of
8090 pointers) types with the same address space through this instruction.
8091 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
8092 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
8097 .. code-block:: llvm
8099 %X = bitcast i8 255 to i8 ; yields i8 :-1
8100 %Y = bitcast i32* %x to sint* ; yields sint*:%x
8101 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
8102 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
8104 .. _i_addrspacecast:
8106 '``addrspacecast .. to``' Instruction
8107 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8114 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
8119 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
8120 address space ``n`` to type ``pty2`` in address space ``m``.
8125 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
8126 to cast and a pointer type to cast it to, which must have a different
8132 The '``addrspacecast``' instruction converts the pointer value
8133 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
8134 value modification, depending on the target and the address space
8135 pair. Pointer conversions within the same address space must be
8136 performed with the ``bitcast`` instruction. Note that if the address space
8137 conversion is legal then both result and operand refer to the same memory
8143 .. code-block:: llvm
8145 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
8146 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
8147 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
8154 The instructions in this category are the "miscellaneous" instructions,
8155 which defy better classification.
8159 '``icmp``' Instruction
8160 ^^^^^^^^^^^^^^^^^^^^^^
8167 <result> = icmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
8172 The '``icmp``' instruction returns a boolean value or a vector of
8173 boolean values based on comparison of its two integer, integer vector,
8174 pointer, or pointer vector operands.
8179 The '``icmp``' instruction takes three operands. The first operand is
8180 the condition code indicating the kind of comparison to perform. It is
8181 not a value, just a keyword. The possible condition code are:
8184 #. ``ne``: not equal
8185 #. ``ugt``: unsigned greater than
8186 #. ``uge``: unsigned greater or equal
8187 #. ``ult``: unsigned less than
8188 #. ``ule``: unsigned less or equal
8189 #. ``sgt``: signed greater than
8190 #. ``sge``: signed greater or equal
8191 #. ``slt``: signed less than
8192 #. ``sle``: signed less or equal
8194 The remaining two arguments must be :ref:`integer <t_integer>` or
8195 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
8196 must also be identical types.
8201 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
8202 code given as ``cond``. The comparison performed always yields either an
8203 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
8205 #. ``eq``: yields ``true`` if the operands are equal, ``false``
8206 otherwise. No sign interpretation is necessary or performed.
8207 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
8208 otherwise. No sign interpretation is necessary or performed.
8209 #. ``ugt``: interprets the operands as unsigned values and yields
8210 ``true`` if ``op1`` is greater than ``op2``.
8211 #. ``uge``: interprets the operands as unsigned values and yields
8212 ``true`` if ``op1`` is greater than or equal to ``op2``.
8213 #. ``ult``: interprets the operands as unsigned values and yields
8214 ``true`` if ``op1`` is less than ``op2``.
8215 #. ``ule``: interprets the operands as unsigned values and yields
8216 ``true`` if ``op1`` is less than or equal to ``op2``.
8217 #. ``sgt``: interprets the operands as signed values and yields ``true``
8218 if ``op1`` is greater than ``op2``.
8219 #. ``sge``: interprets the operands as signed values and yields ``true``
8220 if ``op1`` is greater than or equal to ``op2``.
8221 #. ``slt``: interprets the operands as signed values and yields ``true``
8222 if ``op1`` is less than ``op2``.
8223 #. ``sle``: interprets the operands as signed values and yields ``true``
8224 if ``op1`` is less than or equal to ``op2``.
8226 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
8227 are compared as if they were integers.
8229 If the operands are integer vectors, then they are compared element by
8230 element. The result is an ``i1`` vector with the same number of elements
8231 as the values being compared. Otherwise, the result is an ``i1``.
8236 .. code-block:: llvm
8238 <result> = icmp eq i32 4, 5 ; yields: result=false
8239 <result> = icmp ne float* %X, %X ; yields: result=false
8240 <result> = icmp ult i16 4, 5 ; yields: result=true
8241 <result> = icmp sgt i16 4, 5 ; yields: result=false
8242 <result> = icmp ule i16 -4, 5 ; yields: result=false
8243 <result> = icmp sge i16 4, 5 ; yields: result=false
8245 Note that the code generator does not yet support vector types with the
8246 ``icmp`` instruction.
8250 '``fcmp``' Instruction
8251 ^^^^^^^^^^^^^^^^^^^^^^
8258 <result> = fcmp [fast-math flags]* <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
8263 The '``fcmp``' instruction returns a boolean value or vector of boolean
8264 values based on comparison of its operands.
8266 If the operands are floating point scalars, then the result type is a
8267 boolean (:ref:`i1 <t_integer>`).
8269 If the operands are floating point vectors, then the result type is a
8270 vector of boolean with the same number of elements as the operands being
8276 The '``fcmp``' instruction takes three operands. The first operand is
8277 the condition code indicating the kind of comparison to perform. It is
8278 not a value, just a keyword. The possible condition code are:
8280 #. ``false``: no comparison, always returns false
8281 #. ``oeq``: ordered and equal
8282 #. ``ogt``: ordered and greater than
8283 #. ``oge``: ordered and greater than or equal
8284 #. ``olt``: ordered and less than
8285 #. ``ole``: ordered and less than or equal
8286 #. ``one``: ordered and not equal
8287 #. ``ord``: ordered (no nans)
8288 #. ``ueq``: unordered or equal
8289 #. ``ugt``: unordered or greater than
8290 #. ``uge``: unordered or greater than or equal
8291 #. ``ult``: unordered or less than
8292 #. ``ule``: unordered or less than or equal
8293 #. ``une``: unordered or not equal
8294 #. ``uno``: unordered (either nans)
8295 #. ``true``: no comparison, always returns true
8297 *Ordered* means that neither operand is a QNAN while *unordered* means
8298 that either operand may be a QNAN.
8300 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
8301 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
8302 type. They must have identical types.
8307 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
8308 condition code given as ``cond``. If the operands are vectors, then the
8309 vectors are compared element by element. Each comparison performed
8310 always yields an :ref:`i1 <t_integer>` result, as follows:
8312 #. ``false``: always yields ``false``, regardless of operands.
8313 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
8314 is equal to ``op2``.
8315 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
8316 is greater than ``op2``.
8317 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
8318 is greater than or equal to ``op2``.
8319 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
8320 is less than ``op2``.
8321 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
8322 is less than or equal to ``op2``.
8323 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
8324 is not equal to ``op2``.
8325 #. ``ord``: yields ``true`` if both operands are not a QNAN.
8326 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
8328 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
8329 greater than ``op2``.
8330 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
8331 greater than or equal to ``op2``.
8332 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
8334 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
8335 less than or equal to ``op2``.
8336 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
8337 not equal to ``op2``.
8338 #. ``uno``: yields ``true`` if either operand is a QNAN.
8339 #. ``true``: always yields ``true``, regardless of operands.
8341 The ``fcmp`` instruction can also optionally take any number of
8342 :ref:`fast-math flags <fastmath>`, which are optimization hints to enable
8343 otherwise unsafe floating point optimizations.
8345 Any set of fast-math flags are legal on an ``fcmp`` instruction, but the
8346 only flags that have any effect on its semantics are those that allow
8347 assumptions to be made about the values of input arguments; namely
8348 ``nnan``, ``ninf``, and ``nsz``. See :ref:`fastmath` for more information.
8353 .. code-block:: llvm
8355 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
8356 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
8357 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
8358 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
8360 Note that the code generator does not yet support vector types with the
8361 ``fcmp`` instruction.
8365 '``phi``' Instruction
8366 ^^^^^^^^^^^^^^^^^^^^^
8373 <result> = phi <ty> [ <val0>, <label0>], ...
8378 The '``phi``' instruction is used to implement the φ node in the SSA
8379 graph representing the function.
8384 The type of the incoming values is specified with the first type field.
8385 After this, the '``phi``' instruction takes a list of pairs as
8386 arguments, with one pair for each predecessor basic block of the current
8387 block. Only values of :ref:`first class <t_firstclass>` type may be used as
8388 the value arguments to the PHI node. Only labels may be used as the
8391 There must be no non-phi instructions between the start of a basic block
8392 and the PHI instructions: i.e. PHI instructions must be first in a basic
8395 For the purposes of the SSA form, the use of each incoming value is
8396 deemed to occur on the edge from the corresponding predecessor block to
8397 the current block (but after any definition of an '``invoke``'
8398 instruction's return value on the same edge).
8403 At runtime, the '``phi``' instruction logically takes on the value
8404 specified by the pair corresponding to the predecessor basic block that
8405 executed just prior to the current block.
8410 .. code-block:: llvm
8412 Loop: ; Infinite loop that counts from 0 on up...
8413 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
8414 %nextindvar = add i32 %indvar, 1
8419 '``select``' Instruction
8420 ^^^^^^^^^^^^^^^^^^^^^^^^
8427 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
8429 selty is either i1 or {<N x i1>}
8434 The '``select``' instruction is used to choose one value based on a
8435 condition, without IR-level branching.
8440 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
8441 values indicating the condition, and two values of the same :ref:`first
8442 class <t_firstclass>` type.
8447 If the condition is an i1 and it evaluates to 1, the instruction returns
8448 the first value argument; otherwise, it returns the second value
8451 If the condition is a vector of i1, then the value arguments must be
8452 vectors of the same size, and the selection is done element by element.
8454 If the condition is an i1 and the value arguments are vectors of the
8455 same size, then an entire vector is selected.
8460 .. code-block:: llvm
8462 %X = select i1 true, i8 17, i8 42 ; yields i8:17
8466 '``call``' Instruction
8467 ^^^^^^^^^^^^^^^^^^^^^^
8474 <result> = [tail | musttail | notail ] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
8480 The '``call``' instruction represents a simple function call.
8485 This instruction requires several arguments:
8487 #. The optional ``tail`` and ``musttail`` markers indicate that the optimizers
8488 should perform tail call optimization. The ``tail`` marker is a hint that
8489 `can be ignored <CodeGenerator.html#sibcallopt>`_. The ``musttail`` marker
8490 means that the call must be tail call optimized in order for the program to
8491 be correct. The ``musttail`` marker provides these guarantees:
8493 #. The call will not cause unbounded stack growth if it is part of a
8494 recursive cycle in the call graph.
8495 #. Arguments with the :ref:`inalloca <attr_inalloca>` attribute are
8498 Both markers imply that the callee does not access allocas or varargs from
8499 the caller. Calls marked ``musttail`` must obey the following additional
8502 - The call must immediately precede a :ref:`ret <i_ret>` instruction,
8503 or a pointer bitcast followed by a ret instruction.
8504 - The ret instruction must return the (possibly bitcasted) value
8505 produced by the call or void.
8506 - The caller and callee prototypes must match. Pointer types of
8507 parameters or return types may differ in pointee type, but not
8509 - The calling conventions of the caller and callee must match.
8510 - All ABI-impacting function attributes, such as sret, byval, inreg,
8511 returned, and inalloca, must match.
8512 - The callee must be varargs iff the caller is varargs. Bitcasting a
8513 non-varargs function to the appropriate varargs type is legal so
8514 long as the non-varargs prefixes obey the other rules.
8516 Tail call optimization for calls marked ``tail`` is guaranteed to occur if
8517 the following conditions are met:
8519 - Caller and callee both have the calling convention ``fastcc``.
8520 - The call is in tail position (ret immediately follows call and ret
8521 uses value of call or is void).
8522 - Option ``-tailcallopt`` is enabled, or
8523 ``llvm::GuaranteedTailCallOpt`` is ``true``.
8524 - `Platform-specific constraints are
8525 met. <CodeGenerator.html#tailcallopt>`_
8527 #. The optional ``notail`` marker indicates that the optimizers should not add
8528 ``tail`` or ``musttail`` markers to the call. It is used to prevent tail
8529 call optimization from being performed on the call.
8531 #. The optional "cconv" marker indicates which :ref:`calling
8532 convention <callingconv>` the call should use. If none is
8533 specified, the call defaults to using C calling conventions. The
8534 calling convention of the call must match the calling convention of
8535 the target function, or else the behavior is undefined.
8536 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
8537 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
8539 #. '``ty``': the type of the call instruction itself which is also the
8540 type of the return value. Functions that return no value are marked
8542 #. '``fnty``': shall be the signature of the pointer to function value
8543 being invoked. The argument types must match the types implied by
8544 this signature. This type can be omitted if the function is not
8545 varargs and if the function type does not return a pointer to a
8547 #. '``fnptrval``': An LLVM value containing a pointer to a function to
8548 be invoked. In most cases, this is a direct function invocation, but
8549 indirect ``call``'s are just as possible, calling an arbitrary pointer
8551 #. '``function args``': argument list whose types match the function
8552 signature argument types and parameter attributes. All arguments must
8553 be of :ref:`first class <t_firstclass>` type. If the function signature
8554 indicates the function accepts a variable number of arguments, the
8555 extra arguments can be specified.
8556 #. The optional :ref:`function attributes <fnattrs>` list. Only
8557 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
8558 attributes are valid here.
8559 #. The optional :ref:`operand bundles <opbundles>` list.
8564 The '``call``' instruction is used to cause control flow to transfer to
8565 a specified function, with its incoming arguments bound to the specified
8566 values. Upon a '``ret``' instruction in the called function, control
8567 flow continues with the instruction after the function call, and the
8568 return value of the function is bound to the result argument.
8573 .. code-block:: llvm
8575 %retval = call i32 @test(i32 %argc)
8576 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
8577 %X = tail call i32 @foo() ; yields i32
8578 %Y = tail call fastcc i32 @foo() ; yields i32
8579 call void %foo(i8 97 signext)
8581 %struct.A = type { i32, i8 }
8582 %r = call %struct.A @foo() ; yields { i32, i8 }
8583 %gr = extractvalue %struct.A %r, 0 ; yields i32
8584 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
8585 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
8586 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
8588 llvm treats calls to some functions with names and arguments that match
8589 the standard C99 library as being the C99 library functions, and may
8590 perform optimizations or generate code for them under that assumption.
8591 This is something we'd like to change in the future to provide better
8592 support for freestanding environments and non-C-based languages.
8596 '``va_arg``' Instruction
8597 ^^^^^^^^^^^^^^^^^^^^^^^^
8604 <resultval> = va_arg <va_list*> <arglist>, <argty>
8609 The '``va_arg``' instruction is used to access arguments passed through
8610 the "variable argument" area of a function call. It is used to implement
8611 the ``va_arg`` macro in C.
8616 This instruction takes a ``va_list*`` value and the type of the
8617 argument. It returns a value of the specified argument type and
8618 increments the ``va_list`` to point to the next argument. The actual
8619 type of ``va_list`` is target specific.
8624 The '``va_arg``' instruction loads an argument of the specified type
8625 from the specified ``va_list`` and causes the ``va_list`` to point to
8626 the next argument. For more information, see the variable argument
8627 handling :ref:`Intrinsic Functions <int_varargs>`.
8629 It is legal for this instruction to be called in a function which does
8630 not take a variable number of arguments, for example, the ``vfprintf``
8633 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
8634 function <intrinsics>` because it takes a type as an argument.
8639 See the :ref:`variable argument processing <int_varargs>` section.
8641 Note that the code generator does not yet fully support va\_arg on many
8642 targets. Also, it does not currently support va\_arg with aggregate
8643 types on any target.
8647 '``landingpad``' Instruction
8648 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8655 <resultval> = landingpad <resultty> <clause>+
8656 <resultval> = landingpad <resultty> cleanup <clause>*
8658 <clause> := catch <type> <value>
8659 <clause> := filter <array constant type> <array constant>
8664 The '``landingpad``' instruction is used by `LLVM's exception handling
8665 system <ExceptionHandling.html#overview>`_ to specify that a basic block
8666 is a landing pad --- one where the exception lands, and corresponds to the
8667 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
8668 defines values supplied by the :ref:`personality function <personalityfn>` upon
8669 re-entry to the function. The ``resultval`` has the type ``resultty``.
8675 ``cleanup`` flag indicates that the landing pad block is a cleanup.
8677 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
8678 contains the global variable representing the "type" that may be caught
8679 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
8680 clause takes an array constant as its argument. Use
8681 "``[0 x i8**] undef``" for a filter which cannot throw. The
8682 '``landingpad``' instruction must contain *at least* one ``clause`` or
8683 the ``cleanup`` flag.
8688 The '``landingpad``' instruction defines the values which are set by the
8689 :ref:`personality function <personalityfn>` upon re-entry to the function, and
8690 therefore the "result type" of the ``landingpad`` instruction. As with
8691 calling conventions, how the personality function results are
8692 represented in LLVM IR is target specific.
8694 The clauses are applied in order from top to bottom. If two
8695 ``landingpad`` instructions are merged together through inlining, the
8696 clauses from the calling function are appended to the list of clauses.
8697 When the call stack is being unwound due to an exception being thrown,
8698 the exception is compared against each ``clause`` in turn. If it doesn't
8699 match any of the clauses, and the ``cleanup`` flag is not set, then
8700 unwinding continues further up the call stack.
8702 The ``landingpad`` instruction has several restrictions:
8704 - A landing pad block is a basic block which is the unwind destination
8705 of an '``invoke``' instruction.
8706 - A landing pad block must have a '``landingpad``' instruction as its
8707 first non-PHI instruction.
8708 - There can be only one '``landingpad``' instruction within the landing
8710 - A basic block that is not a landing pad block may not include a
8711 '``landingpad``' instruction.
8716 .. code-block:: llvm
8718 ;; A landing pad which can catch an integer.
8719 %res = landingpad { i8*, i32 }
8721 ;; A landing pad that is a cleanup.
8722 %res = landingpad { i8*, i32 }
8724 ;; A landing pad which can catch an integer and can only throw a double.
8725 %res = landingpad { i8*, i32 }
8727 filter [1 x i8**] [@_ZTId]
8731 '``cleanuppad``' Instruction
8732 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8739 <resultval> = cleanuppad [<args>*]
8744 The '``cleanuppad``' instruction is used by `LLVM's exception handling
8745 system <ExceptionHandling.html#overview>`_ to specify that a basic block
8746 is a cleanup block --- one where a personality routine attempts to
8747 transfer control to run cleanup actions.
8748 The ``args`` correspond to whatever additional
8749 information the :ref:`personality function <personalityfn>` requires to
8750 execute the cleanup.
8751 The ``resultval`` has the type :ref:`token <t_token>` and is used to
8752 match the ``cleanuppad`` to corresponding :ref:`cleanuprets <i_cleanupret>`
8753 and :ref:`cleanupendpads <i_cleanupendpad>`.
8758 The instruction takes a list of arbitrary values which are interpreted
8759 by the :ref:`personality function <personalityfn>`.
8764 When the call stack is being unwound due to an exception being thrown,
8765 the :ref:`personality function <personalityfn>` transfers control to the
8766 ``cleanuppad`` with the aid of the personality-specific arguments.
8767 As with calling conventions, how the personality function results are
8768 represented in LLVM IR is target specific.
8770 The ``cleanuppad`` instruction has several restrictions:
8772 - A cleanup block is a basic block which is the unwind destination of
8773 an exceptional instruction.
8774 - A cleanup block must have a '``cleanuppad``' instruction as its
8775 first non-PHI instruction.
8776 - There can be only one '``cleanuppad``' instruction within the
8778 - A basic block that is not a cleanup block may not include a
8779 '``cleanuppad``' instruction.
8780 - All '``cleanupret``'s and '``cleanupendpad``'s which consume a ``cleanuppad``
8781 must have the same exceptional successor.
8782 - It is undefined behavior for control to transfer from a ``cleanuppad`` to a
8783 ``ret`` without first executing a ``cleanupret`` or ``cleanupendpad`` that
8784 consumes the ``cleanuppad``.
8785 - It is undefined behavior for control to transfer from a ``cleanuppad`` to
8786 itself without first executing a ``cleanupret`` or ``cleanupendpad`` that
8787 consumes the ``cleanuppad``.
8792 .. code-block:: llvm
8794 %tok = cleanuppad []
8801 LLVM supports the notion of an "intrinsic function". These functions
8802 have well known names and semantics and are required to follow certain
8803 restrictions. Overall, these intrinsics represent an extension mechanism
8804 for the LLVM language that does not require changing all of the
8805 transformations in LLVM when adding to the language (or the bitcode
8806 reader/writer, the parser, etc...).
8808 Intrinsic function names must all start with an "``llvm.``" prefix. This
8809 prefix is reserved in LLVM for intrinsic names; thus, function names may
8810 not begin with this prefix. Intrinsic functions must always be external
8811 functions: you cannot define the body of intrinsic functions. Intrinsic
8812 functions may only be used in call or invoke instructions: it is illegal
8813 to take the address of an intrinsic function. Additionally, because
8814 intrinsic functions are part of the LLVM language, it is required if any
8815 are added that they be documented here.
8817 Some intrinsic functions can be overloaded, i.e., the intrinsic
8818 represents a family of functions that perform the same operation but on
8819 different data types. Because LLVM can represent over 8 million
8820 different integer types, overloading is used commonly to allow an
8821 intrinsic function to operate on any integer type. One or more of the
8822 argument types or the result type can be overloaded to accept any
8823 integer type. Argument types may also be defined as exactly matching a
8824 previous argument's type or the result type. This allows an intrinsic
8825 function which accepts multiple arguments, but needs all of them to be
8826 of the same type, to only be overloaded with respect to a single
8827 argument or the result.
8829 Overloaded intrinsics will have the names of its overloaded argument
8830 types encoded into its function name, each preceded by a period. Only
8831 those types which are overloaded result in a name suffix. Arguments
8832 whose type is matched against another type do not. For example, the
8833 ``llvm.ctpop`` function can take an integer of any width and returns an
8834 integer of exactly the same integer width. This leads to a family of
8835 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
8836 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
8837 overloaded, and only one type suffix is required. Because the argument's
8838 type is matched against the return type, it does not require its own
8841 To learn how to add an intrinsic function, please see the `Extending
8842 LLVM Guide <ExtendingLLVM.html>`_.
8846 Variable Argument Handling Intrinsics
8847 -------------------------------------
8849 Variable argument support is defined in LLVM with the
8850 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
8851 functions. These functions are related to the similarly named macros
8852 defined in the ``<stdarg.h>`` header file.
8854 All of these functions operate on arguments that use a target-specific
8855 value type "``va_list``". The LLVM assembly language reference manual
8856 does not define what this type is, so all transformations should be
8857 prepared to handle these functions regardless of the type used.
8859 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
8860 variable argument handling intrinsic functions are used.
8862 .. code-block:: llvm
8864 ; This struct is different for every platform. For most platforms,
8865 ; it is merely an i8*.
8866 %struct.va_list = type { i8* }
8868 ; For Unix x86_64 platforms, va_list is the following struct:
8869 ; %struct.va_list = type { i32, i32, i8*, i8* }
8871 define i32 @test(i32 %X, ...) {
8872 ; Initialize variable argument processing
8873 %ap = alloca %struct.va_list
8874 %ap2 = bitcast %struct.va_list* %ap to i8*
8875 call void @llvm.va_start(i8* %ap2)
8877 ; Read a single integer argument
8878 %tmp = va_arg i8* %ap2, i32
8880 ; Demonstrate usage of llvm.va_copy and llvm.va_end
8882 %aq2 = bitcast i8** %aq to i8*
8883 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
8884 call void @llvm.va_end(i8* %aq2)
8886 ; Stop processing of arguments.
8887 call void @llvm.va_end(i8* %ap2)
8891 declare void @llvm.va_start(i8*)
8892 declare void @llvm.va_copy(i8*, i8*)
8893 declare void @llvm.va_end(i8*)
8897 '``llvm.va_start``' Intrinsic
8898 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8905 declare void @llvm.va_start(i8* <arglist>)
8910 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
8911 subsequent use by ``va_arg``.
8916 The argument is a pointer to a ``va_list`` element to initialize.
8921 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
8922 available in C. In a target-dependent way, it initializes the
8923 ``va_list`` element to which the argument points, so that the next call
8924 to ``va_arg`` will produce the first variable argument passed to the
8925 function. Unlike the C ``va_start`` macro, this intrinsic does not need
8926 to know the last argument of the function as the compiler can figure
8929 '``llvm.va_end``' Intrinsic
8930 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8937 declare void @llvm.va_end(i8* <arglist>)
8942 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
8943 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
8948 The argument is a pointer to a ``va_list`` to destroy.
8953 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
8954 available in C. In a target-dependent way, it destroys the ``va_list``
8955 element to which the argument points. Calls to
8956 :ref:`llvm.va_start <int_va_start>` and
8957 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
8962 '``llvm.va_copy``' Intrinsic
8963 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8970 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
8975 The '``llvm.va_copy``' intrinsic copies the current argument position
8976 from the source argument list to the destination argument list.
8981 The first argument is a pointer to a ``va_list`` element to initialize.
8982 The second argument is a pointer to a ``va_list`` element to copy from.
8987 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
8988 available in C. In a target-dependent way, it copies the source
8989 ``va_list`` element into the destination ``va_list`` element. This
8990 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
8991 arbitrarily complex and require, for example, memory allocation.
8993 Accurate Garbage Collection Intrinsics
8994 --------------------------------------
8996 LLVM's support for `Accurate Garbage Collection <GarbageCollection.html>`_
8997 (GC) requires the frontend to generate code containing appropriate intrinsic
8998 calls and select an appropriate GC strategy which knows how to lower these
8999 intrinsics in a manner which is appropriate for the target collector.
9001 These intrinsics allow identification of :ref:`GC roots on the
9002 stack <int_gcroot>`, as well as garbage collector implementations that
9003 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
9004 Frontends for type-safe garbage collected languages should generate
9005 these intrinsics to make use of the LLVM garbage collectors. For more
9006 details, see `Garbage Collection with LLVM <GarbageCollection.html>`_.
9008 Experimental Statepoint Intrinsics
9009 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9011 LLVM provides an second experimental set of intrinsics for describing garbage
9012 collection safepoints in compiled code. These intrinsics are an alternative
9013 to the ``llvm.gcroot`` intrinsics, but are compatible with the ones for
9014 :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers. The
9015 differences in approach are covered in the `Garbage Collection with LLVM
9016 <GarbageCollection.html>`_ documentation. The intrinsics themselves are
9017 described in :doc:`Statepoints`.
9021 '``llvm.gcroot``' Intrinsic
9022 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9029 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
9034 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
9035 the code generator, and allows some metadata to be associated with it.
9040 The first argument specifies the address of a stack object that contains
9041 the root pointer. The second pointer (which must be either a constant or
9042 a global value address) contains the meta-data to be associated with the
9048 At runtime, a call to this intrinsic stores a null pointer into the
9049 "ptrloc" location. At compile-time, the code generator generates
9050 information to allow the runtime to find the pointer at GC safe points.
9051 The '``llvm.gcroot``' intrinsic may only be used in a function which
9052 :ref:`specifies a GC algorithm <gc>`.
9056 '``llvm.gcread``' Intrinsic
9057 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9064 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
9069 The '``llvm.gcread``' intrinsic identifies reads of references from heap
9070 locations, allowing garbage collector implementations that require read
9076 The second argument is the address to read from, which should be an
9077 address allocated from the garbage collector. The first object is a
9078 pointer to the start of the referenced object, if needed by the language
9079 runtime (otherwise null).
9084 The '``llvm.gcread``' intrinsic has the same semantics as a load
9085 instruction, but may be replaced with substantially more complex code by
9086 the garbage collector runtime, as needed. The '``llvm.gcread``'
9087 intrinsic may only be used in a function which :ref:`specifies a GC
9092 '``llvm.gcwrite``' Intrinsic
9093 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9100 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
9105 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
9106 locations, allowing garbage collector implementations that require write
9107 barriers (such as generational or reference counting collectors).
9112 The first argument is the reference to store, the second is the start of
9113 the object to store it to, and the third is the address of the field of
9114 Obj to store to. If the runtime does not require a pointer to the
9115 object, Obj may be null.
9120 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
9121 instruction, but may be replaced with substantially more complex code by
9122 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
9123 intrinsic may only be used in a function which :ref:`specifies a GC
9126 Code Generator Intrinsics
9127 -------------------------
9129 These intrinsics are provided by LLVM to expose special features that
9130 may only be implemented with code generator support.
9132 '``llvm.returnaddress``' Intrinsic
9133 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9140 declare i8 *@llvm.returnaddress(i32 <level>)
9145 The '``llvm.returnaddress``' intrinsic attempts to compute a
9146 target-specific value indicating the return address of the current
9147 function or one of its callers.
9152 The argument to this intrinsic indicates which function to return the
9153 address for. Zero indicates the calling function, one indicates its
9154 caller, etc. The argument is **required** to be a constant integer
9160 The '``llvm.returnaddress``' intrinsic either returns a pointer
9161 indicating the return address of the specified call frame, or zero if it
9162 cannot be identified. The value returned by this intrinsic is likely to
9163 be incorrect or 0 for arguments other than zero, so it should only be
9164 used for debugging purposes.
9166 Note that calling this intrinsic does not prevent function inlining or
9167 other aggressive transformations, so the value returned may not be that
9168 of the obvious source-language caller.
9170 '``llvm.frameaddress``' Intrinsic
9171 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9178 declare i8* @llvm.frameaddress(i32 <level>)
9183 The '``llvm.frameaddress``' intrinsic attempts to return the
9184 target-specific frame pointer value for the specified stack frame.
9189 The argument to this intrinsic indicates which function to return the
9190 frame pointer for. Zero indicates the calling function, one indicates
9191 its caller, etc. The argument is **required** to be a constant integer
9197 The '``llvm.frameaddress``' intrinsic either returns a pointer
9198 indicating the frame address of the specified call frame, or zero if it
9199 cannot be identified. The value returned by this intrinsic is likely to
9200 be incorrect or 0 for arguments other than zero, so it should only be
9201 used for debugging purposes.
9203 Note that calling this intrinsic does not prevent function inlining or
9204 other aggressive transformations, so the value returned may not be that
9205 of the obvious source-language caller.
9207 '``llvm.localescape``' and '``llvm.localrecover``' Intrinsics
9208 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9215 declare void @llvm.localescape(...)
9216 declare i8* @llvm.localrecover(i8* %func, i8* %fp, i32 %idx)
9221 The '``llvm.localescape``' intrinsic escapes offsets of a collection of static
9222 allocas, and the '``llvm.localrecover``' intrinsic applies those offsets to a
9223 live frame pointer to recover the address of the allocation. The offset is
9224 computed during frame layout of the caller of ``llvm.localescape``.
9229 All arguments to '``llvm.localescape``' must be pointers to static allocas or
9230 casts of static allocas. Each function can only call '``llvm.localescape``'
9231 once, and it can only do so from the entry block.
9233 The ``func`` argument to '``llvm.localrecover``' must be a constant
9234 bitcasted pointer to a function defined in the current module. The code
9235 generator cannot determine the frame allocation offset of functions defined in
9238 The ``fp`` argument to '``llvm.localrecover``' must be a frame pointer of a
9239 call frame that is currently live. The return value of '``llvm.localaddress``'
9240 is one way to produce such a value, but various runtimes also expose a suitable
9241 pointer in platform-specific ways.
9243 The ``idx`` argument to '``llvm.localrecover``' indicates which alloca passed to
9244 '``llvm.localescape``' to recover. It is zero-indexed.
9249 These intrinsics allow a group of functions to share access to a set of local
9250 stack allocations of a one parent function. The parent function may call the
9251 '``llvm.localescape``' intrinsic once from the function entry block, and the
9252 child functions can use '``llvm.localrecover``' to access the escaped allocas.
9253 The '``llvm.localescape``' intrinsic blocks inlining, as inlining changes where
9254 the escaped allocas are allocated, which would break attempts to use
9255 '``llvm.localrecover``'.
9257 .. _int_read_register:
9258 .. _int_write_register:
9260 '``llvm.read_register``' and '``llvm.write_register``' Intrinsics
9261 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9268 declare i32 @llvm.read_register.i32(metadata)
9269 declare i64 @llvm.read_register.i64(metadata)
9270 declare void @llvm.write_register.i32(metadata, i32 @value)
9271 declare void @llvm.write_register.i64(metadata, i64 @value)
9277 The '``llvm.read_register``' and '``llvm.write_register``' intrinsics
9278 provides access to the named register. The register must be valid on
9279 the architecture being compiled to. The type needs to be compatible
9280 with the register being read.
9285 The '``llvm.read_register``' intrinsic returns the current value of the
9286 register, where possible. The '``llvm.write_register``' intrinsic sets
9287 the current value of the register, where possible.
9289 This is useful to implement named register global variables that need
9290 to always be mapped to a specific register, as is common practice on
9291 bare-metal programs including OS kernels.
9293 The compiler doesn't check for register availability or use of the used
9294 register in surrounding code, including inline assembly. Because of that,
9295 allocatable registers are not supported.
9297 Warning: So far it only works with the stack pointer on selected
9298 architectures (ARM, AArch64, PowerPC and x86_64). Significant amount of
9299 work is needed to support other registers and even more so, allocatable
9304 '``llvm.stacksave``' Intrinsic
9305 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9312 declare i8* @llvm.stacksave()
9317 The '``llvm.stacksave``' intrinsic is used to remember the current state
9318 of the function stack, for use with
9319 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
9320 implementing language features like scoped automatic variable sized
9326 This intrinsic returns a opaque pointer value that can be passed to
9327 :ref:`llvm.stackrestore <int_stackrestore>`. When an
9328 ``llvm.stackrestore`` intrinsic is executed with a value saved from
9329 ``llvm.stacksave``, it effectively restores the state of the stack to
9330 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
9331 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
9332 were allocated after the ``llvm.stacksave`` was executed.
9334 .. _int_stackrestore:
9336 '``llvm.stackrestore``' Intrinsic
9337 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9344 declare void @llvm.stackrestore(i8* %ptr)
9349 The '``llvm.stackrestore``' intrinsic is used to restore the state of
9350 the function stack to the state it was in when the corresponding
9351 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
9352 useful for implementing language features like scoped automatic variable
9353 sized arrays in C99.
9358 See the description for :ref:`llvm.stacksave <int_stacksave>`.
9360 .. _int_get_dynamic_area_offset:
9362 '``llvm.get.dynamic.area.offset``' Intrinsic
9363 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9370 declare i32 @llvm.get.dynamic.area.offset.i32()
9371 declare i64 @llvm.get.dynamic.area.offset.i64()
9376 The '``llvm.get.dynamic.area.offset.*``' intrinsic family is used to
9377 get the offset from native stack pointer to the address of the most
9378 recent dynamic alloca on the caller's stack. These intrinsics are
9379 intendend for use in combination with
9380 :ref:`llvm.stacksave <int_stacksave>` to get a
9381 pointer to the most recent dynamic alloca. This is useful, for example,
9382 for AddressSanitizer's stack unpoisoning routines.
9387 These intrinsics return a non-negative integer value that can be used to
9388 get the address of the most recent dynamic alloca, allocated by :ref:`alloca <i_alloca>`
9389 on the caller's stack. In particular, for targets where stack grows downwards,
9390 adding this offset to the native stack pointer would get the address of the most
9391 recent dynamic alloca. For targets where stack grows upwards, the situation is a bit more
9392 complicated, because substracting this value from stack pointer would get the address
9393 one past the end of the most recent dynamic alloca.
9395 Although for most targets `llvm.get.dynamic.area.offset <int_get_dynamic_area_offset>`
9396 returns just a zero, for others, such as PowerPC and PowerPC64, it returns a
9397 compile-time-known constant value.
9399 The return value type of :ref:`llvm.get.dynamic.area.offset <int_get_dynamic_area_offset>`
9400 must match the target's generic address space's (address space 0) pointer type.
9402 '``llvm.prefetch``' Intrinsic
9403 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9410 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
9415 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
9416 insert a prefetch instruction if supported; otherwise, it is a noop.
9417 Prefetches have no effect on the behavior of the program but can change
9418 its performance characteristics.
9423 ``address`` is the address to be prefetched, ``rw`` is the specifier
9424 determining if the fetch should be for a read (0) or write (1), and
9425 ``locality`` is a temporal locality specifier ranging from (0) - no
9426 locality, to (3) - extremely local keep in cache. The ``cache type``
9427 specifies whether the prefetch is performed on the data (1) or
9428 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
9429 arguments must be constant integers.
9434 This intrinsic does not modify the behavior of the program. In
9435 particular, prefetches cannot trap and do not produce a value. On
9436 targets that support this intrinsic, the prefetch can provide hints to
9437 the processor cache for better performance.
9439 '``llvm.pcmarker``' Intrinsic
9440 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9447 declare void @llvm.pcmarker(i32 <id>)
9452 The '``llvm.pcmarker``' intrinsic is a method to export a Program
9453 Counter (PC) in a region of code to simulators and other tools. The
9454 method is target specific, but it is expected that the marker will use
9455 exported symbols to transmit the PC of the marker. The marker makes no
9456 guarantees that it will remain with any specific instruction after
9457 optimizations. It is possible that the presence of a marker will inhibit
9458 optimizations. The intended use is to be inserted after optimizations to
9459 allow correlations of simulation runs.
9464 ``id`` is a numerical id identifying the marker.
9469 This intrinsic does not modify the behavior of the program. Backends
9470 that do not support this intrinsic may ignore it.
9472 '``llvm.readcyclecounter``' Intrinsic
9473 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9480 declare i64 @llvm.readcyclecounter()
9485 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
9486 counter register (or similar low latency, high accuracy clocks) on those
9487 targets that support it. On X86, it should map to RDTSC. On Alpha, it
9488 should map to RPCC. As the backing counters overflow quickly (on the
9489 order of 9 seconds on alpha), this should only be used for small
9495 When directly supported, reading the cycle counter should not modify any
9496 memory. Implementations are allowed to either return a application
9497 specific value or a system wide value. On backends without support, this
9498 is lowered to a constant 0.
9500 Note that runtime support may be conditional on the privilege-level code is
9501 running at and the host platform.
9503 '``llvm.clear_cache``' Intrinsic
9504 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9511 declare void @llvm.clear_cache(i8*, i8*)
9516 The '``llvm.clear_cache``' intrinsic ensures visibility of modifications
9517 in the specified range to the execution unit of the processor. On
9518 targets with non-unified instruction and data cache, the implementation
9519 flushes the instruction cache.
9524 On platforms with coherent instruction and data caches (e.g. x86), this
9525 intrinsic is a nop. On platforms with non-coherent instruction and data
9526 cache (e.g. ARM, MIPS), the intrinsic is lowered either to appropriate
9527 instructions or a system call, if cache flushing requires special
9530 The default behavior is to emit a call to ``__clear_cache`` from the run
9533 This instrinsic does *not* empty the instruction pipeline. Modifications
9534 of the current function are outside the scope of the intrinsic.
9536 '``llvm.instrprof_increment``' Intrinsic
9537 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9544 declare void @llvm.instrprof_increment(i8* <name>, i64 <hash>,
9545 i32 <num-counters>, i32 <index>)
9550 The '``llvm.instrprof_increment``' intrinsic can be emitted by a
9551 frontend for use with instrumentation based profiling. These will be
9552 lowered by the ``-instrprof`` pass to generate execution counts of a
9558 The first argument is a pointer to a global variable containing the
9559 name of the entity being instrumented. This should generally be the
9560 (mangled) function name for a set of counters.
9562 The second argument is a hash value that can be used by the consumer
9563 of the profile data to detect changes to the instrumented source, and
9564 the third is the number of counters associated with ``name``. It is an
9565 error if ``hash`` or ``num-counters`` differ between two instances of
9566 ``instrprof_increment`` that refer to the same name.
9568 The last argument refers to which of the counters for ``name`` should
9569 be incremented. It should be a value between 0 and ``num-counters``.
9574 This intrinsic represents an increment of a profiling counter. It will
9575 cause the ``-instrprof`` pass to generate the appropriate data
9576 structures and the code to increment the appropriate value, in a
9577 format that can be written out by a compiler runtime and consumed via
9578 the ``llvm-profdata`` tool.
9580 '``llvm.instrprof_value_profile``' Intrinsic
9581 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9588 declare void @llvm.instrprof_value_profile(i8* <name>, i64 <hash>,
9589 i64 <value>, i32 <value_kind>,
9595 The '``llvm.instrprof_value_profile``' intrinsic can be emitted by a
9596 frontend for use with instrumentation based profiling. This will be
9597 lowered by the ``-instrprof`` pass to find out the target values,
9598 instrumented expressions take in a program at runtime.
9603 The first argument is a pointer to a global variable containing the
9604 name of the entity being instrumented. ``name`` should generally be the
9605 (mangled) function name for a set of counters.
9607 The second argument is a hash value that can be used by the consumer
9608 of the profile data to detect changes to the instrumented source. It
9609 is an error if ``hash`` differs between two instances of
9610 ``llvm.instrprof_*`` that refer to the same name.
9612 The third argument is the value of the expression being profiled. The profiled
9613 expression's value should be representable as an unsigned 64-bit value. The
9614 fourth argument represents the kind of value profiling that is being done. The
9615 supported value profiling kinds are enumerated through the
9616 ``InstrProfValueKind`` type declared in the
9617 ``<include/llvm/ProfileData/InstrProf.h>`` header file. The last argument is the
9618 index of the instrumented expression within ``name``. It should be >= 0.
9623 This intrinsic represents the point where a call to a runtime routine
9624 should be inserted for value profiling of target expressions. ``-instrprof``
9625 pass will generate the appropriate data structures and replace the
9626 ``llvm.instrprof_value_profile`` intrinsic with the call to the profile
9627 runtime library with proper arguments.
9629 Standard C Library Intrinsics
9630 -----------------------------
9632 LLVM provides intrinsics for a few important standard C library
9633 functions. These intrinsics allow source-language front-ends to pass
9634 information about the alignment of the pointer arguments to the code
9635 generator, providing opportunity for more efficient code generation.
9639 '``llvm.memcpy``' Intrinsic
9640 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9645 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
9646 integer bit width and for different address spaces. Not all targets
9647 support all bit widths however.
9651 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
9652 i32 <len>, i32 <align>, i1 <isvolatile>)
9653 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
9654 i64 <len>, i32 <align>, i1 <isvolatile>)
9659 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
9660 source location to the destination location.
9662 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
9663 intrinsics do not return a value, takes extra alignment/isvolatile
9664 arguments and the pointers can be in specified address spaces.
9669 The first argument is a pointer to the destination, the second is a
9670 pointer to the source. The third argument is an integer argument
9671 specifying the number of bytes to copy, the fourth argument is the
9672 alignment of the source and destination locations, and the fifth is a
9673 boolean indicating a volatile access.
9675 If the call to this intrinsic has an alignment value that is not 0 or 1,
9676 then the caller guarantees that both the source and destination pointers
9677 are aligned to that boundary.
9679 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
9680 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
9681 very cleanly specified and it is unwise to depend on it.
9686 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
9687 source location to the destination location, which are not allowed to
9688 overlap. It copies "len" bytes of memory over. If the argument is known
9689 to be aligned to some boundary, this can be specified as the fourth
9690 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
9692 '``llvm.memmove``' Intrinsic
9693 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9698 This is an overloaded intrinsic. You can use llvm.memmove on any integer
9699 bit width and for different address space. Not all targets support all
9704 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
9705 i32 <len>, i32 <align>, i1 <isvolatile>)
9706 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
9707 i64 <len>, i32 <align>, i1 <isvolatile>)
9712 The '``llvm.memmove.*``' intrinsics move a block of memory from the
9713 source location to the destination location. It is similar to the
9714 '``llvm.memcpy``' intrinsic but allows the two memory locations to
9717 Note that, unlike the standard libc function, the ``llvm.memmove.*``
9718 intrinsics do not return a value, takes extra alignment/isvolatile
9719 arguments and the pointers can be in specified address spaces.
9724 The first argument is a pointer to the destination, the second is a
9725 pointer to the source. The third argument is an integer argument
9726 specifying the number of bytes to copy, the fourth argument is the
9727 alignment of the source and destination locations, and the fifth is a
9728 boolean indicating a volatile access.
9730 If the call to this intrinsic has an alignment value that is not 0 or 1,
9731 then the caller guarantees that the source and destination pointers are
9732 aligned to that boundary.
9734 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
9735 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
9736 not very cleanly specified and it is unwise to depend on it.
9741 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
9742 source location to the destination location, which may overlap. It
9743 copies "len" bytes of memory over. If the argument is known to be
9744 aligned to some boundary, this can be specified as the fourth argument,
9745 otherwise it should be set to 0 or 1 (both meaning no alignment).
9747 '``llvm.memset.*``' Intrinsics
9748 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9753 This is an overloaded intrinsic. You can use llvm.memset on any integer
9754 bit width and for different address spaces. However, not all targets
9755 support all bit widths.
9759 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
9760 i32 <len>, i32 <align>, i1 <isvolatile>)
9761 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
9762 i64 <len>, i32 <align>, i1 <isvolatile>)
9767 The '``llvm.memset.*``' intrinsics fill a block of memory with a
9768 particular byte value.
9770 Note that, unlike the standard libc function, the ``llvm.memset``
9771 intrinsic does not return a value and takes extra alignment/volatile
9772 arguments. Also, the destination can be in an arbitrary address space.
9777 The first argument is a pointer to the destination to fill, the second
9778 is the byte value with which to fill it, the third argument is an
9779 integer argument specifying the number of bytes to fill, and the fourth
9780 argument is the known alignment of the destination location.
9782 If the call to this intrinsic has an alignment value that is not 0 or 1,
9783 then the caller guarantees that the destination pointer is aligned to
9786 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
9787 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
9788 very cleanly specified and it is unwise to depend on it.
9793 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
9794 at the destination location. If the argument is known to be aligned to
9795 some boundary, this can be specified as the fourth argument, otherwise
9796 it should be set to 0 or 1 (both meaning no alignment).
9798 '``llvm.sqrt.*``' Intrinsic
9799 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9804 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
9805 floating point or vector of floating point type. Not all targets support
9810 declare float @llvm.sqrt.f32(float %Val)
9811 declare double @llvm.sqrt.f64(double %Val)
9812 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
9813 declare fp128 @llvm.sqrt.f128(fp128 %Val)
9814 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
9819 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
9820 returning the same value as the libm '``sqrt``' functions would. Unlike
9821 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
9822 negative numbers other than -0.0 (which allows for better optimization,
9823 because there is no need to worry about errno being set).
9824 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
9829 The argument and return value are floating point numbers of the same
9835 This function returns the sqrt of the specified operand if it is a
9836 nonnegative floating point number.
9838 '``llvm.powi.*``' Intrinsic
9839 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9844 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
9845 floating point or vector of floating point type. Not all targets support
9850 declare float @llvm.powi.f32(float %Val, i32 %power)
9851 declare double @llvm.powi.f64(double %Val, i32 %power)
9852 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
9853 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
9854 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
9859 The '``llvm.powi.*``' intrinsics return the first operand raised to the
9860 specified (positive or negative) power. The order of evaluation of
9861 multiplications is not defined. When a vector of floating point type is
9862 used, the second argument remains a scalar integer value.
9867 The second argument is an integer power, and the first is a value to
9868 raise to that power.
9873 This function returns the first value raised to the second power with an
9874 unspecified sequence of rounding operations.
9876 '``llvm.sin.*``' Intrinsic
9877 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9882 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
9883 floating point or vector of floating point type. Not all targets support
9888 declare float @llvm.sin.f32(float %Val)
9889 declare double @llvm.sin.f64(double %Val)
9890 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
9891 declare fp128 @llvm.sin.f128(fp128 %Val)
9892 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
9897 The '``llvm.sin.*``' intrinsics return the sine of the operand.
9902 The argument and return value are floating point numbers of the same
9908 This function returns the sine of the specified operand, returning the
9909 same values as the libm ``sin`` functions would, and handles error
9910 conditions in the same way.
9912 '``llvm.cos.*``' Intrinsic
9913 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9918 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
9919 floating point or vector of floating point type. Not all targets support
9924 declare float @llvm.cos.f32(float %Val)
9925 declare double @llvm.cos.f64(double %Val)
9926 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
9927 declare fp128 @llvm.cos.f128(fp128 %Val)
9928 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
9933 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
9938 The argument and return value are floating point numbers of the same
9944 This function returns the cosine of the specified operand, returning the
9945 same values as the libm ``cos`` functions would, and handles error
9946 conditions in the same way.
9948 '``llvm.pow.*``' Intrinsic
9949 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9954 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
9955 floating point or vector of floating point type. Not all targets support
9960 declare float @llvm.pow.f32(float %Val, float %Power)
9961 declare double @llvm.pow.f64(double %Val, double %Power)
9962 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
9963 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
9964 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
9969 The '``llvm.pow.*``' intrinsics return the first operand raised to the
9970 specified (positive or negative) power.
9975 The second argument is a floating point power, and the first is a value
9976 to raise to that power.
9981 This function returns the first value raised to the second power,
9982 returning the same values as the libm ``pow`` functions would, and
9983 handles error conditions in the same way.
9985 '``llvm.exp.*``' Intrinsic
9986 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9991 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
9992 floating point or vector of floating point type. Not all targets support
9997 declare float @llvm.exp.f32(float %Val)
9998 declare double @llvm.exp.f64(double %Val)
9999 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
10000 declare fp128 @llvm.exp.f128(fp128 %Val)
10001 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
10006 The '``llvm.exp.*``' intrinsics perform the exp function.
10011 The argument and return value are floating point numbers of the same
10017 This function returns the same values as the libm ``exp`` functions
10018 would, and handles error conditions in the same way.
10020 '``llvm.exp2.*``' Intrinsic
10021 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10026 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
10027 floating point or vector of floating point type. Not all targets support
10032 declare float @llvm.exp2.f32(float %Val)
10033 declare double @llvm.exp2.f64(double %Val)
10034 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
10035 declare fp128 @llvm.exp2.f128(fp128 %Val)
10036 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
10041 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
10046 The argument and return value are floating point numbers of the same
10052 This function returns the same values as the libm ``exp2`` functions
10053 would, and handles error conditions in the same way.
10055 '``llvm.log.*``' Intrinsic
10056 ^^^^^^^^^^^^^^^^^^^^^^^^^^
10061 This is an overloaded intrinsic. You can use ``llvm.log`` on any
10062 floating point or vector of floating point type. Not all targets support
10067 declare float @llvm.log.f32(float %Val)
10068 declare double @llvm.log.f64(double %Val)
10069 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
10070 declare fp128 @llvm.log.f128(fp128 %Val)
10071 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
10076 The '``llvm.log.*``' intrinsics perform the log function.
10081 The argument and return value are floating point numbers of the same
10087 This function returns the same values as the libm ``log`` functions
10088 would, and handles error conditions in the same way.
10090 '``llvm.log10.*``' Intrinsic
10091 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10096 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
10097 floating point or vector of floating point type. Not all targets support
10102 declare float @llvm.log10.f32(float %Val)
10103 declare double @llvm.log10.f64(double %Val)
10104 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
10105 declare fp128 @llvm.log10.f128(fp128 %Val)
10106 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
10111 The '``llvm.log10.*``' intrinsics perform the log10 function.
10116 The argument and return value are floating point numbers of the same
10122 This function returns the same values as the libm ``log10`` functions
10123 would, and handles error conditions in the same way.
10125 '``llvm.log2.*``' Intrinsic
10126 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10131 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
10132 floating point or vector of floating point type. Not all targets support
10137 declare float @llvm.log2.f32(float %Val)
10138 declare double @llvm.log2.f64(double %Val)
10139 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
10140 declare fp128 @llvm.log2.f128(fp128 %Val)
10141 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
10146 The '``llvm.log2.*``' intrinsics perform the log2 function.
10151 The argument and return value are floating point numbers of the same
10157 This function returns the same values as the libm ``log2`` functions
10158 would, and handles error conditions in the same way.
10160 '``llvm.fma.*``' Intrinsic
10161 ^^^^^^^^^^^^^^^^^^^^^^^^^^
10166 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
10167 floating point or vector of floating point type. Not all targets support
10172 declare float @llvm.fma.f32(float %a, float %b, float %c)
10173 declare double @llvm.fma.f64(double %a, double %b, double %c)
10174 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
10175 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
10176 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
10181 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
10187 The argument and return value are floating point numbers of the same
10193 This function returns the same values as the libm ``fma`` functions
10194 would, and does not set errno.
10196 '``llvm.fabs.*``' Intrinsic
10197 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10202 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
10203 floating point or vector of floating point type. Not all targets support
10208 declare float @llvm.fabs.f32(float %Val)
10209 declare double @llvm.fabs.f64(double %Val)
10210 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
10211 declare fp128 @llvm.fabs.f128(fp128 %Val)
10212 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
10217 The '``llvm.fabs.*``' intrinsics return the absolute value of the
10223 The argument and return value are floating point numbers of the same
10229 This function returns the same values as the libm ``fabs`` functions
10230 would, and handles error conditions in the same way.
10232 '``llvm.minnum.*``' Intrinsic
10233 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10238 This is an overloaded intrinsic. You can use ``llvm.minnum`` on any
10239 floating point or vector of floating point type. Not all targets support
10244 declare float @llvm.minnum.f32(float %Val0, float %Val1)
10245 declare double @llvm.minnum.f64(double %Val0, double %Val1)
10246 declare x86_fp80 @llvm.minnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
10247 declare fp128 @llvm.minnum.f128(fp128 %Val0, fp128 %Val1)
10248 declare ppc_fp128 @llvm.minnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
10253 The '``llvm.minnum.*``' intrinsics return the minimum of the two
10260 The arguments and return value are floating point numbers of the same
10266 Follows the IEEE-754 semantics for minNum, which also match for libm's
10269 If either operand is a NaN, returns the other non-NaN operand. Returns
10270 NaN only if both operands are NaN. If the operands compare equal,
10271 returns a value that compares equal to both operands. This means that
10272 fmin(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
10274 '``llvm.maxnum.*``' Intrinsic
10275 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10280 This is an overloaded intrinsic. You can use ``llvm.maxnum`` on any
10281 floating point or vector of floating point type. Not all targets support
10286 declare float @llvm.maxnum.f32(float %Val0, float %Val1l)
10287 declare double @llvm.maxnum.f64(double %Val0, double %Val1)
10288 declare x86_fp80 @llvm.maxnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
10289 declare fp128 @llvm.maxnum.f128(fp128 %Val0, fp128 %Val1)
10290 declare ppc_fp128 @llvm.maxnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
10295 The '``llvm.maxnum.*``' intrinsics return the maximum of the two
10302 The arguments and return value are floating point numbers of the same
10307 Follows the IEEE-754 semantics for maxNum, which also match for libm's
10310 If either operand is a NaN, returns the other non-NaN operand. Returns
10311 NaN only if both operands are NaN. If the operands compare equal,
10312 returns a value that compares equal to both operands. This means that
10313 fmax(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
10315 '``llvm.copysign.*``' Intrinsic
10316 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10321 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
10322 floating point or vector of floating point type. Not all targets support
10327 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
10328 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
10329 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
10330 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
10331 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
10336 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
10337 first operand and the sign of the second operand.
10342 The arguments and return value are floating point numbers of the same
10348 This function returns the same values as the libm ``copysign``
10349 functions would, and handles error conditions in the same way.
10351 '``llvm.floor.*``' Intrinsic
10352 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10357 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
10358 floating point or vector of floating point type. Not all targets support
10363 declare float @llvm.floor.f32(float %Val)
10364 declare double @llvm.floor.f64(double %Val)
10365 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
10366 declare fp128 @llvm.floor.f128(fp128 %Val)
10367 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
10372 The '``llvm.floor.*``' intrinsics return the floor of the operand.
10377 The argument and return value are floating point numbers of the same
10383 This function returns the same values as the libm ``floor`` functions
10384 would, and handles error conditions in the same way.
10386 '``llvm.ceil.*``' Intrinsic
10387 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10392 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
10393 floating point or vector of floating point type. Not all targets support
10398 declare float @llvm.ceil.f32(float %Val)
10399 declare double @llvm.ceil.f64(double %Val)
10400 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
10401 declare fp128 @llvm.ceil.f128(fp128 %Val)
10402 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
10407 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
10412 The argument and return value are floating point numbers of the same
10418 This function returns the same values as the libm ``ceil`` functions
10419 would, and handles error conditions in the same way.
10421 '``llvm.trunc.*``' Intrinsic
10422 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10427 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
10428 floating point or vector of floating point type. Not all targets support
10433 declare float @llvm.trunc.f32(float %Val)
10434 declare double @llvm.trunc.f64(double %Val)
10435 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
10436 declare fp128 @llvm.trunc.f128(fp128 %Val)
10437 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
10442 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
10443 nearest integer not larger in magnitude than the operand.
10448 The argument and return value are floating point numbers of the same
10454 This function returns the same values as the libm ``trunc`` functions
10455 would, and handles error conditions in the same way.
10457 '``llvm.rint.*``' Intrinsic
10458 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10463 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
10464 floating point or vector of floating point type. Not all targets support
10469 declare float @llvm.rint.f32(float %Val)
10470 declare double @llvm.rint.f64(double %Val)
10471 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
10472 declare fp128 @llvm.rint.f128(fp128 %Val)
10473 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
10478 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
10479 nearest integer. It may raise an inexact floating-point exception if the
10480 operand isn't an integer.
10485 The argument and return value are floating point numbers of the same
10491 This function returns the same values as the libm ``rint`` functions
10492 would, and handles error conditions in the same way.
10494 '``llvm.nearbyint.*``' Intrinsic
10495 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10500 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
10501 floating point or vector of floating point type. Not all targets support
10506 declare float @llvm.nearbyint.f32(float %Val)
10507 declare double @llvm.nearbyint.f64(double %Val)
10508 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
10509 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
10510 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
10515 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
10521 The argument and return value are floating point numbers of the same
10527 This function returns the same values as the libm ``nearbyint``
10528 functions would, and handles error conditions in the same way.
10530 '``llvm.round.*``' Intrinsic
10531 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10536 This is an overloaded intrinsic. You can use ``llvm.round`` on any
10537 floating point or vector of floating point type. Not all targets support
10542 declare float @llvm.round.f32(float %Val)
10543 declare double @llvm.round.f64(double %Val)
10544 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
10545 declare fp128 @llvm.round.f128(fp128 %Val)
10546 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
10551 The '``llvm.round.*``' intrinsics returns the operand rounded to the
10557 The argument and return value are floating point numbers of the same
10563 This function returns the same values as the libm ``round``
10564 functions would, and handles error conditions in the same way.
10566 Bit Manipulation Intrinsics
10567 ---------------------------
10569 LLVM provides intrinsics for a few important bit manipulation
10570 operations. These allow efficient code generation for some algorithms.
10572 '``llvm.bitreverse.*``' Intrinsics
10573 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10578 This is an overloaded intrinsic function. You can use bitreverse on any
10583 declare i16 @llvm.bitreverse.i16(i16 <id>)
10584 declare i32 @llvm.bitreverse.i32(i32 <id>)
10585 declare i64 @llvm.bitreverse.i64(i64 <id>)
10590 The '``llvm.bitreverse``' family of intrinsics is used to reverse the
10591 bitpattern of an integer value; for example ``0b1234567`` becomes
10597 The ``llvm.bitreverse.iN`` intrinsic returns an i16 value that has bit
10598 ``M`` in the input moved to bit ``N-M`` in the output.
10600 '``llvm.bswap.*``' Intrinsics
10601 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10606 This is an overloaded intrinsic function. You can use bswap on any
10607 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
10611 declare i16 @llvm.bswap.i16(i16 <id>)
10612 declare i32 @llvm.bswap.i32(i32 <id>)
10613 declare i64 @llvm.bswap.i64(i64 <id>)
10618 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
10619 values with an even number of bytes (positive multiple of 16 bits).
10620 These are useful for performing operations on data that is not in the
10621 target's native byte order.
10626 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
10627 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
10628 intrinsic returns an i32 value that has the four bytes of the input i32
10629 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
10630 returned i32 will have its bytes in 3, 2, 1, 0 order. The
10631 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
10632 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
10635 '``llvm.ctpop.*``' Intrinsic
10636 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10641 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
10642 bit width, or on any vector with integer elements. Not all targets
10643 support all bit widths or vector types, however.
10647 declare i8 @llvm.ctpop.i8(i8 <src>)
10648 declare i16 @llvm.ctpop.i16(i16 <src>)
10649 declare i32 @llvm.ctpop.i32(i32 <src>)
10650 declare i64 @llvm.ctpop.i64(i64 <src>)
10651 declare i256 @llvm.ctpop.i256(i256 <src>)
10652 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
10657 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
10663 The only argument is the value to be counted. The argument may be of any
10664 integer type, or a vector with integer elements. The return type must
10665 match the argument type.
10670 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
10671 each element of a vector.
10673 '``llvm.ctlz.*``' Intrinsic
10674 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10679 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
10680 integer bit width, or any vector whose elements are integers. Not all
10681 targets support all bit widths or vector types, however.
10685 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
10686 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
10687 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
10688 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
10689 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
10690 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
10695 The '``llvm.ctlz``' family of intrinsic functions counts the number of
10696 leading zeros in a variable.
10701 The first argument is the value to be counted. This argument may be of
10702 any integer type, or a vector with integer element type. The return
10703 type must match the first argument type.
10705 The second argument must be a constant and is a flag to indicate whether
10706 the intrinsic should ensure that a zero as the first argument produces a
10707 defined result. Historically some architectures did not provide a
10708 defined result for zero values as efficiently, and many algorithms are
10709 now predicated on avoiding zero-value inputs.
10714 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
10715 zeros in a variable, or within each element of the vector. If
10716 ``src == 0`` then the result is the size in bits of the type of ``src``
10717 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
10718 ``llvm.ctlz(i32 2) = 30``.
10720 '``llvm.cttz.*``' Intrinsic
10721 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10726 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
10727 integer bit width, or any vector of integer elements. Not all targets
10728 support all bit widths or vector types, however.
10732 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
10733 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
10734 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
10735 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
10736 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
10737 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
10742 The '``llvm.cttz``' family of intrinsic functions counts the number of
10748 The first argument is the value to be counted. This argument may be of
10749 any integer type, or a vector with integer element type. The return
10750 type must match the first argument type.
10752 The second argument must be a constant and is a flag to indicate whether
10753 the intrinsic should ensure that a zero as the first argument produces a
10754 defined result. Historically some architectures did not provide a
10755 defined result for zero values as efficiently, and many algorithms are
10756 now predicated on avoiding zero-value inputs.
10761 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
10762 zeros in a variable, or within each element of a vector. If ``src == 0``
10763 then the result is the size in bits of the type of ``src`` if
10764 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
10765 ``llvm.cttz(2) = 1``.
10769 Arithmetic with Overflow Intrinsics
10770 -----------------------------------
10772 LLVM provides intrinsics for some arithmetic with overflow operations.
10774 '``llvm.sadd.with.overflow.*``' Intrinsics
10775 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10780 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
10781 on any integer bit width.
10785 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
10786 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
10787 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
10792 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
10793 a signed addition of the two arguments, and indicate whether an overflow
10794 occurred during the signed summation.
10799 The arguments (%a and %b) and the first element of the result structure
10800 may be of integer types of any bit width, but they must have the same
10801 bit width. The second element of the result structure must be of type
10802 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
10808 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
10809 a signed addition of the two variables. They return a structure --- the
10810 first element of which is the signed summation, and the second element
10811 of which is a bit specifying if the signed summation resulted in an
10817 .. code-block:: llvm
10819 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
10820 %sum = extractvalue {i32, i1} %res, 0
10821 %obit = extractvalue {i32, i1} %res, 1
10822 br i1 %obit, label %overflow, label %normal
10824 '``llvm.uadd.with.overflow.*``' Intrinsics
10825 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10830 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
10831 on any integer bit width.
10835 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
10836 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
10837 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
10842 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
10843 an unsigned addition of the two arguments, and indicate whether a carry
10844 occurred during the unsigned summation.
10849 The arguments (%a and %b) and the first element of the result structure
10850 may be of integer types of any bit width, but they must have the same
10851 bit width. The second element of the result structure must be of type
10852 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
10858 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
10859 an unsigned addition of the two arguments. They return a structure --- the
10860 first element of which is the sum, and the second element of which is a
10861 bit specifying if the unsigned summation resulted in a carry.
10866 .. code-block:: llvm
10868 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
10869 %sum = extractvalue {i32, i1} %res, 0
10870 %obit = extractvalue {i32, i1} %res, 1
10871 br i1 %obit, label %carry, label %normal
10873 '``llvm.ssub.with.overflow.*``' Intrinsics
10874 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10879 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
10880 on any integer bit width.
10884 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
10885 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
10886 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
10891 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
10892 a signed subtraction of the two arguments, and indicate whether an
10893 overflow occurred during the signed subtraction.
10898 The arguments (%a and %b) and the first element of the result structure
10899 may be of integer types of any bit width, but they must have the same
10900 bit width. The second element of the result structure must be of type
10901 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
10907 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
10908 a signed subtraction of the two arguments. They return a structure --- the
10909 first element of which is the subtraction, and the second element of
10910 which is a bit specifying if the signed subtraction resulted in an
10916 .. code-block:: llvm
10918 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
10919 %sum = extractvalue {i32, i1} %res, 0
10920 %obit = extractvalue {i32, i1} %res, 1
10921 br i1 %obit, label %overflow, label %normal
10923 '``llvm.usub.with.overflow.*``' Intrinsics
10924 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10929 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
10930 on any integer bit width.
10934 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
10935 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
10936 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
10941 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
10942 an unsigned subtraction of the two arguments, and indicate whether an
10943 overflow occurred during the unsigned subtraction.
10948 The arguments (%a and %b) and the first element of the result structure
10949 may be of integer types of any bit width, but they must have the same
10950 bit width. The second element of the result structure must be of type
10951 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
10957 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
10958 an unsigned subtraction of the two arguments. They return a structure ---
10959 the first element of which is the subtraction, and the second element of
10960 which is a bit specifying if the unsigned subtraction resulted in an
10966 .. code-block:: llvm
10968 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
10969 %sum = extractvalue {i32, i1} %res, 0
10970 %obit = extractvalue {i32, i1} %res, 1
10971 br i1 %obit, label %overflow, label %normal
10973 '``llvm.smul.with.overflow.*``' Intrinsics
10974 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10979 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
10980 on any integer bit width.
10984 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
10985 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
10986 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
10991 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
10992 a signed multiplication of the two arguments, and indicate whether an
10993 overflow occurred during the signed multiplication.
10998 The arguments (%a and %b) and the first element of the result structure
10999 may be of integer types of any bit width, but they must have the same
11000 bit width. The second element of the result structure must be of type
11001 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
11007 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
11008 a signed multiplication of the two arguments. They return a structure ---
11009 the first element of which is the multiplication, and the second element
11010 of which is a bit specifying if the signed multiplication resulted in an
11016 .. code-block:: llvm
11018 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
11019 %sum = extractvalue {i32, i1} %res, 0
11020 %obit = extractvalue {i32, i1} %res, 1
11021 br i1 %obit, label %overflow, label %normal
11023 '``llvm.umul.with.overflow.*``' Intrinsics
11024 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11029 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
11030 on any integer bit width.
11034 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
11035 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
11036 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
11041 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
11042 a unsigned multiplication of the two arguments, and indicate whether an
11043 overflow occurred during the unsigned multiplication.
11048 The arguments (%a and %b) and the first element of the result structure
11049 may be of integer types of any bit width, but they must have the same
11050 bit width. The second element of the result structure must be of type
11051 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
11057 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
11058 an unsigned multiplication of the two arguments. They return a structure ---
11059 the first element of which is the multiplication, and the second
11060 element of which is a bit specifying if the unsigned multiplication
11061 resulted in an overflow.
11066 .. code-block:: llvm
11068 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
11069 %sum = extractvalue {i32, i1} %res, 0
11070 %obit = extractvalue {i32, i1} %res, 1
11071 br i1 %obit, label %overflow, label %normal
11073 Specialised Arithmetic Intrinsics
11074 ---------------------------------
11076 '``llvm.canonicalize.*``' Intrinsic
11077 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11084 declare float @llvm.canonicalize.f32(float %a)
11085 declare double @llvm.canonicalize.f64(double %b)
11090 The '``llvm.canonicalize.*``' intrinsic returns the platform specific canonical
11091 encoding of a floating point number. This canonicalization is useful for
11092 implementing certain numeric primitives such as frexp. The canonical encoding is
11093 defined by IEEE-754-2008 to be:
11097 2.1.8 canonical encoding: The preferred encoding of a floating-point
11098 representation in a format. Applied to declets, significands of finite
11099 numbers, infinities, and NaNs, especially in decimal formats.
11101 This operation can also be considered equivalent to the IEEE-754-2008
11102 conversion of a floating-point value to the same format. NaNs are handled
11103 according to section 6.2.
11105 Examples of non-canonical encodings:
11107 - x87 pseudo denormals, pseudo NaNs, pseudo Infinity, Unnormals. These are
11108 converted to a canonical representation per hardware-specific protocol.
11109 - Many normal decimal floating point numbers have non-canonical alternative
11111 - Some machines, like GPUs or ARMv7 NEON, do not support subnormal values.
11112 These are treated as non-canonical encodings of zero and with be flushed to
11113 a zero of the same sign by this operation.
11115 Note that per IEEE-754-2008 6.2, systems that support signaling NaNs with
11116 default exception handling must signal an invalid exception, and produce a
11119 This function should always be implementable as multiplication by 1.0, provided
11120 that the compiler does not constant fold the operation. Likewise, division by
11121 1.0 and ``llvm.minnum(x, x)`` are possible implementations. Addition with
11122 -0.0 is also sufficient provided that the rounding mode is not -Infinity.
11124 ``@llvm.canonicalize`` must preserve the equality relation. That is:
11126 - ``(@llvm.canonicalize(x) == x)`` is equivalent to ``(x == x)``
11127 - ``(@llvm.canonicalize(x) == @llvm.canonicalize(y))`` is equivalent to
11130 Additionally, the sign of zero must be conserved:
11131 ``@llvm.canonicalize(-0.0) = -0.0`` and ``@llvm.canonicalize(+0.0) = +0.0``
11133 The payload bits of a NaN must be conserved, with two exceptions.
11134 First, environments which use only a single canonical representation of NaN
11135 must perform said canonicalization. Second, SNaNs must be quieted per the
11138 The canonicalization operation may be optimized away if:
11140 - The input is known to be canonical. For example, it was produced by a
11141 floating-point operation that is required by the standard to be canonical.
11142 - The result is consumed only by (or fused with) other floating-point
11143 operations. That is, the bits of the floating point value are not examined.
11145 '``llvm.fmuladd.*``' Intrinsic
11146 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11153 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
11154 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
11159 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
11160 expressions that can be fused if the code generator determines that (a) the
11161 target instruction set has support for a fused operation, and (b) that the
11162 fused operation is more efficient than the equivalent, separate pair of mul
11163 and add instructions.
11168 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
11169 multiplicands, a and b, and an addend c.
11178 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
11180 is equivalent to the expression a \* b + c, except that rounding will
11181 not be performed between the multiplication and addition steps if the
11182 code generator fuses the operations. Fusion is not guaranteed, even if
11183 the target platform supports it. If a fused multiply-add is required the
11184 corresponding llvm.fma.\* intrinsic function should be used
11185 instead. This never sets errno, just as '``llvm.fma.*``'.
11190 .. code-block:: llvm
11192 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields float:r2 = (a * b) + c
11195 '``llvm.uabsdiff.*``' and '``llvm.sabsdiff.*``' Intrinsics
11196 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11200 This is an overloaded intrinsic. The loaded data is a vector of any integer bit width.
11202 .. code-block:: llvm
11204 declare <4 x integer> @llvm.uabsdiff.v4i32(<4 x integer> %a, <4 x integer> %b)
11210 The ``llvm.uabsdiff`` intrinsic returns a vector result of the absolute difference
11211 of the two operands, treating them both as unsigned integers. The intermediate
11212 calculations are computed using infinitely precise unsigned arithmetic. The final
11213 result will be truncated to the given type.
11215 The ``llvm.sabsdiff`` intrinsic returns a vector result of the absolute difference of
11216 the two operands, treating them both as signed integers. If the result overflows, the
11217 behavior is undefined.
11221 These intrinsics are primarily used during the code generation stage of compilation.
11222 They are generated by compiler passes such as the Loop and SLP vectorizers. It is not
11223 recommended for users to create them manually.
11228 Both intrinsics take two integer of the same bitwidth.
11235 call <4 x i32> @llvm.uabsdiff.v4i32(<4 x i32> %a, <4 x i32> %b)
11239 %1 = zext <4 x i32> %a to <4 x i64>
11240 %2 = zext <4 x i32> %b to <4 x i64>
11241 %sub = sub <4 x i64> %1, %2
11242 %trunc = trunc <4 x i64> to <4 x i32>
11244 and the expression::
11246 call <4 x i32> @llvm.sabsdiff.v4i32(<4 x i32> %a, <4 x i32> %b)
11250 %sub = sub nsw <4 x i32> %a, %b
11251 %ispos = icmp sge <4 x i32> %sub, zeroinitializer
11252 %neg = sub nsw <4 x i32> zeroinitializer, %sub
11253 %1 = select <4 x i1> %ispos, <4 x i32> %sub, <4 x i32> %neg
11256 Half Precision Floating Point Intrinsics
11257 ----------------------------------------
11259 For most target platforms, half precision floating point is a
11260 storage-only format. This means that it is a dense encoding (in memory)
11261 but does not support computation in the format.
11263 This means that code must first load the half-precision floating point
11264 value as an i16, then convert it to float with
11265 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
11266 then be performed on the float value (including extending to double
11267 etc). To store the value back to memory, it is first converted to float
11268 if needed, then converted to i16 with
11269 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
11272 .. _int_convert_to_fp16:
11274 '``llvm.convert.to.fp16``' Intrinsic
11275 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11282 declare i16 @llvm.convert.to.fp16.f32(float %a)
11283 declare i16 @llvm.convert.to.fp16.f64(double %a)
11288 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
11289 conventional floating point type to half precision floating point format.
11294 The intrinsic function contains single argument - the value to be
11300 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
11301 conventional floating point format to half precision floating point format. The
11302 return value is an ``i16`` which contains the converted number.
11307 .. code-block:: llvm
11309 %res = call i16 @llvm.convert.to.fp16.f32(float %a)
11310 store i16 %res, i16* @x, align 2
11312 .. _int_convert_from_fp16:
11314 '``llvm.convert.from.fp16``' Intrinsic
11315 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11322 declare float @llvm.convert.from.fp16.f32(i16 %a)
11323 declare double @llvm.convert.from.fp16.f64(i16 %a)
11328 The '``llvm.convert.from.fp16``' intrinsic function performs a
11329 conversion from half precision floating point format to single precision
11330 floating point format.
11335 The intrinsic function contains single argument - the value to be
11341 The '``llvm.convert.from.fp16``' intrinsic function performs a
11342 conversion from half single precision floating point format to single
11343 precision floating point format. The input half-float value is
11344 represented by an ``i16`` value.
11349 .. code-block:: llvm
11351 %a = load i16, i16* @x, align 2
11352 %res = call float @llvm.convert.from.fp16(i16 %a)
11354 .. _dbg_intrinsics:
11356 Debugger Intrinsics
11357 -------------------
11359 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
11360 prefix), are described in the `LLVM Source Level
11361 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
11364 Exception Handling Intrinsics
11365 -----------------------------
11367 The LLVM exception handling intrinsics (which all start with
11368 ``llvm.eh.`` prefix), are described in the `LLVM Exception
11369 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
11371 .. _int_trampoline:
11373 Trampoline Intrinsics
11374 ---------------------
11376 These intrinsics make it possible to excise one parameter, marked with
11377 the :ref:`nest <nest>` attribute, from a function. The result is a
11378 callable function pointer lacking the nest parameter - the caller does
11379 not need to provide a value for it. Instead, the value to use is stored
11380 in advance in a "trampoline", a block of memory usually allocated on the
11381 stack, which also contains code to splice the nest value into the
11382 argument list. This is used to implement the GCC nested function address
11385 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
11386 then the resulting function pointer has signature ``i32 (i32, i32)*``.
11387 It can be created as follows:
11389 .. code-block:: llvm
11391 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
11392 %tramp1 = getelementptr [10 x i8], [10 x i8]* %tramp, i32 0, i32 0
11393 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
11394 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
11395 %fp = bitcast i8* %p to i32 (i32, i32)*
11397 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
11398 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
11402 '``llvm.init.trampoline``' Intrinsic
11403 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11410 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
11415 This fills the memory pointed to by ``tramp`` with executable code,
11416 turning it into a trampoline.
11421 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
11422 pointers. The ``tramp`` argument must point to a sufficiently large and
11423 sufficiently aligned block of memory; this memory is written to by the
11424 intrinsic. Note that the size and the alignment are target-specific -
11425 LLVM currently provides no portable way of determining them, so a
11426 front-end that generates this intrinsic needs to have some
11427 target-specific knowledge. The ``func`` argument must hold a function
11428 bitcast to an ``i8*``.
11433 The block of memory pointed to by ``tramp`` is filled with target
11434 dependent code, turning it into a function. Then ``tramp`` needs to be
11435 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
11436 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
11437 function's signature is the same as that of ``func`` with any arguments
11438 marked with the ``nest`` attribute removed. At most one such ``nest``
11439 argument is allowed, and it must be of pointer type. Calling the new
11440 function is equivalent to calling ``func`` with the same argument list,
11441 but with ``nval`` used for the missing ``nest`` argument. If, after
11442 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
11443 modified, then the effect of any later call to the returned function
11444 pointer is undefined.
11448 '``llvm.adjust.trampoline``' Intrinsic
11449 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11456 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
11461 This performs any required machine-specific adjustment to the address of
11462 a trampoline (passed as ``tramp``).
11467 ``tramp`` must point to a block of memory which already has trampoline
11468 code filled in by a previous call to
11469 :ref:`llvm.init.trampoline <int_it>`.
11474 On some architectures the address of the code to be executed needs to be
11475 different than the address where the trampoline is actually stored. This
11476 intrinsic returns the executable address corresponding to ``tramp``
11477 after performing the required machine specific adjustments. The pointer
11478 returned can then be :ref:`bitcast and executed <int_trampoline>`.
11480 .. _int_mload_mstore:
11482 Masked Vector Load and Store Intrinsics
11483 ---------------------------------------
11485 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.
11489 '``llvm.masked.load.*``' Intrinsics
11490 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11494 This is an overloaded intrinsic. The loaded data is a vector of any integer, floating point or pointer data type.
11498 declare <16 x float> @llvm.masked.load.v16f32 (<16 x float>* <ptr>, i32 <alignment>, <16 x i1> <mask>, <16 x float> <passthru>)
11499 declare <2 x double> @llvm.masked.load.v2f64 (<2 x double>* <ptr>, i32 <alignment>, <2 x i1> <mask>, <2 x double> <passthru>)
11500 ;; The data is a vector of pointers to double
11501 declare <8 x double*> @llvm.masked.load.v8p0f64 (<8 x double*>* <ptr>, i32 <alignment>, <8 x i1> <mask>, <8 x double*> <passthru>)
11502 ;; The data is a vector of function pointers
11503 declare <8 x i32 ()*> @llvm.masked.load.v8p0f_i32f (<8 x i32 ()*>* <ptr>, i32 <alignment>, <8 x i1> <mask>, <8 x i32 ()*> <passthru>)
11508 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 of the '``passthru``' operand.
11514 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 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 the '``passthru``' operand are the same vector types.
11520 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.
11521 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.
11526 %res = call <16 x float> @llvm.masked.load.v16f32 (<16 x float>* %ptr, i32 4, <16 x i1>%mask, <16 x float> %passthru)
11528 ;; The result of the two following instructions is identical aside from potential memory access exception
11529 %loadlal = load <16 x float>, <16 x float>* %ptr, align 4
11530 %res = select <16 x i1> %mask, <16 x float> %loadlal, <16 x float> %passthru
11534 '``llvm.masked.store.*``' Intrinsics
11535 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11539 This is an overloaded intrinsic. The data stored in memory is a vector of any integer, floating point or pointer data type.
11543 declare void @llvm.masked.store.v8i32 (<8 x i32> <value>, <8 x i32>* <ptr>, i32 <alignment>, <8 x i1> <mask>)
11544 declare void @llvm.masked.store.v16f32 (<16 x float> <value>, <16 x float>* <ptr>, i32 <alignment>, <16 x i1> <mask>)
11545 ;; The data is a vector of pointers to double
11546 declare void @llvm.masked.store.v8p0f64 (<8 x double*> <value>, <8 x double*>* <ptr>, i32 <alignment>, <8 x i1> <mask>)
11547 ;; The data is a vector of function pointers
11548 declare void @llvm.masked.store.v4p0f_i32f (<4 x i32 ()*> <value>, <4 x i32 ()*>* <ptr>, i32 <alignment>, <4 x i1> <mask>)
11553 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.
11558 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.
11564 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.
11565 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.
11569 call void @llvm.masked.store.v16f32(<16 x float> %value, <16 x float>* %ptr, i32 4, <16 x i1> %mask)
11571 ;; The result of the following instructions is identical aside from potential data races and memory access exceptions
11572 %oldval = load <16 x float>, <16 x float>* %ptr, align 4
11573 %res = select <16 x i1> %mask, <16 x float> %value, <16 x float> %oldval
11574 store <16 x float> %res, <16 x float>* %ptr, align 4
11577 Masked Vector Gather and Scatter Intrinsics
11578 -------------------------------------------
11580 LLVM provides intrinsics for vector gather and scatter operations. They are similar to :ref:`Masked Vector Load and Store <int_mload_mstore>`, except they are designed for arbitrary memory accesses, rather than sequential memory accesses. Gather and scatter also employ 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 are off, no memory is accessed.
11584 '``llvm.masked.gather.*``' Intrinsics
11585 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11589 This is an overloaded intrinsic. The loaded data are multiple scalar values of any integer, floating point or pointer data type gathered together into one vector.
11593 declare <16 x float> @llvm.masked.gather.v16f32 (<16 x float*> <ptrs>, i32 <alignment>, <16 x i1> <mask>, <16 x float> <passthru>)
11594 declare <2 x double> @llvm.masked.gather.v2f64 (<2 x double*> <ptrs>, i32 <alignment>, <2 x i1> <mask>, <2 x double> <passthru>)
11595 declare <8 x float*> @llvm.masked.gather.v8p0f32 (<8 x float**> <ptrs>, i32 <alignment>, <8 x i1> <mask>, <8 x float*> <passthru>)
11600 Reads scalar values from arbitrary memory locations and gathers them into one vector. The memory locations are provided in the vector of pointers '``ptrs``'. The memory is accessed 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 of the '``passthru``' operand.
11606 The first operand is a vector of pointers which holds all memory addresses to read. The second operand is an alignment of the source addresses. It must be a constant integer value. The third operand, mask, is a vector of boolean 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 vector of pointers and the type of the '``passthru``' operand are the same vector types.
11612 The '``llvm.masked.gather``' intrinsic is designed for conditional reading of multiple scalar values from arbitrary memory locations in a single IR operation. It is useful for targets that support vector masked gathers and allows vectorizing basic blocks with data and control divergence. Other targets may support this intrinsic differently, for example by lowering it into a sequence of scalar load operations.
11613 The semantics of this operation are equivalent to a sequence of conditional scalar loads with subsequent gathering all loaded values into a single vector. The mask restricts memory access to certain lanes and facilitates vectorization of predicated basic blocks.
11618 %res = call <4 x double> @llvm.masked.gather.v4f64 (<4 x double*> %ptrs, i32 8, <4 x i1>%mask, <4 x double> <true, true, true, true>)
11620 ;; The gather with all-true mask is equivalent to the following instruction sequence
11621 %ptr0 = extractelement <4 x double*> %ptrs, i32 0
11622 %ptr1 = extractelement <4 x double*> %ptrs, i32 1
11623 %ptr2 = extractelement <4 x double*> %ptrs, i32 2
11624 %ptr3 = extractelement <4 x double*> %ptrs, i32 3
11626 %val0 = load double, double* %ptr0, align 8
11627 %val1 = load double, double* %ptr1, align 8
11628 %val2 = load double, double* %ptr2, align 8
11629 %val3 = load double, double* %ptr3, align 8
11631 %vec0 = insertelement <4 x double>undef, %val0, 0
11632 %vec01 = insertelement <4 x double>%vec0, %val1, 1
11633 %vec012 = insertelement <4 x double>%vec01, %val2, 2
11634 %vec0123 = insertelement <4 x double>%vec012, %val3, 3
11638 '``llvm.masked.scatter.*``' Intrinsics
11639 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11643 This is an overloaded intrinsic. The data stored in memory is a vector of any integer, floating point or pointer data type. Each vector element is stored in an arbitrary memory address. Scatter with overlapping addresses is guaranteed to be ordered from least-significant to most-significant element.
11647 declare void @llvm.masked.scatter.v8i32 (<8 x i32> <value>, <8 x i32*> <ptrs>, i32 <alignment>, <8 x i1> <mask>)
11648 declare void @llvm.masked.scatter.v16f32 (<16 x float> <value>, <16 x float*> <ptrs>, i32 <alignment>, <16 x i1> <mask>)
11649 declare void @llvm.masked.scatter.v4p0f64 (<4 x double*> <value>, <4 x double**> <ptrs>, i32 <alignment>, <4 x i1> <mask>)
11654 Writes each element from the value vector to the corresponding memory address. The memory addresses are represented as a vector of pointers. Writing is done 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.
11659 The first operand is a vector value to be written to memory. The second operand is a vector of pointers, pointing to where the value elements should be stored. It has the same underlying type as the value operand. The third operand is an alignment of the destination addresses. 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.
11665 The '``llvm.masked.scatter``' intrinsics is designed for writing selected vector elements to arbitrary memory addresses in a single IR operation. The operation may be conditional, when not all bits in the mask are switched on. It is useful for targets that support vector masked scatter and allows vectorizing basic blocks with data and control divergence. Other targets may support this intrinsic differently, for example by lowering it into a sequence of branches that guard scalar store operations.
11669 ;; This instruction unconditionaly stores data vector in multiple addresses
11670 call @llvm.masked.scatter.v8i32 (<8 x i32> %value, <8 x i32*> %ptrs, i32 4, <8 x i1> <true, true, .. true>)
11672 ;; It is equivalent to a list of scalar stores
11673 %val0 = extractelement <8 x i32> %value, i32 0
11674 %val1 = extractelement <8 x i32> %value, i32 1
11676 %val7 = extractelement <8 x i32> %value, i32 7
11677 %ptr0 = extractelement <8 x i32*> %ptrs, i32 0
11678 %ptr1 = extractelement <8 x i32*> %ptrs, i32 1
11680 %ptr7 = extractelement <8 x i32*> %ptrs, i32 7
11681 ;; Note: the order of the following stores is important when they overlap:
11682 store i32 %val0, i32* %ptr0, align 4
11683 store i32 %val1, i32* %ptr1, align 4
11685 store i32 %val7, i32* %ptr7, align 4
11691 This class of intrinsics provides information about the lifetime of
11692 memory objects and ranges where variables are immutable.
11696 '``llvm.lifetime.start``' Intrinsic
11697 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11704 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
11709 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
11715 The first argument is a constant integer representing the size of the
11716 object, or -1 if it is variable sized. The second argument is a pointer
11722 This intrinsic indicates that before this point in the code, the value
11723 of the memory pointed to by ``ptr`` is dead. This means that it is known
11724 to never be used and has an undefined value. A load from the pointer
11725 that precedes this intrinsic can be replaced with ``'undef'``.
11729 '``llvm.lifetime.end``' Intrinsic
11730 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11737 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
11742 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
11748 The first argument is a constant integer representing the size of the
11749 object, or -1 if it is variable sized. The second argument is a pointer
11755 This intrinsic indicates that after this point in the code, the value of
11756 the memory pointed to by ``ptr`` is dead. This means that it is known to
11757 never be used and has an undefined value. Any stores into the memory
11758 object following this intrinsic may be removed as dead.
11760 '``llvm.invariant.start``' Intrinsic
11761 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11768 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
11773 The '``llvm.invariant.start``' intrinsic specifies that the contents of
11774 a memory object will not change.
11779 The first argument is a constant integer representing the size of the
11780 object, or -1 if it is variable sized. The second argument is a pointer
11786 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
11787 the return value, the referenced memory location is constant and
11790 '``llvm.invariant.end``' Intrinsic
11791 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11798 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
11803 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
11804 memory object are mutable.
11809 The first argument is the matching ``llvm.invariant.start`` intrinsic.
11810 The second argument is a constant integer representing the size of the
11811 object, or -1 if it is variable sized and the third argument is a
11812 pointer to the object.
11817 This intrinsic indicates that the memory is mutable again.
11819 '``llvm.invariant.group.barrier``' Intrinsic
11820 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11827 declare i8* @llvm.invariant.group.barrier(i8* <ptr>)
11832 The '``llvm.invariant.group.barrier``' intrinsic can be used when an invariant
11833 established by invariant.group metadata no longer holds, to obtain a new pointer
11834 value that does not carry the invariant information.
11840 The ``llvm.invariant.group.barrier`` takes only one argument, which is
11841 the pointer to the memory for which the ``invariant.group`` no longer holds.
11846 Returns another pointer that aliases its argument but which is considered different
11847 for the purposes of ``load``/``store`` ``invariant.group`` metadata.
11852 This class of intrinsics is designed to be generic and has no specific
11855 '``llvm.var.annotation``' Intrinsic
11856 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11863 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
11868 The '``llvm.var.annotation``' intrinsic.
11873 The first argument is a pointer to a value, the second is a pointer to a
11874 global string, the third is a pointer to a global string which is the
11875 source file name, and the last argument is the line number.
11880 This intrinsic allows annotation of local variables with arbitrary
11881 strings. This can be useful for special purpose optimizations that want
11882 to look for these annotations. These have no other defined use; they are
11883 ignored by code generation and optimization.
11885 '``llvm.ptr.annotation.*``' Intrinsic
11886 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11891 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
11892 pointer to an integer of any width. *NOTE* you must specify an address space for
11893 the pointer. The identifier for the default address space is the integer
11898 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
11899 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
11900 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
11901 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
11902 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
11907 The '``llvm.ptr.annotation``' intrinsic.
11912 The first argument is a pointer to an integer value of arbitrary bitwidth
11913 (result of some expression), the second is a pointer to a global string, the
11914 third is a pointer to a global string which is the source file name, and the
11915 last argument is the line number. It returns the value of the first argument.
11920 This intrinsic allows annotation of a pointer to an integer with arbitrary
11921 strings. This can be useful for special purpose optimizations that want to look
11922 for these annotations. These have no other defined use; they are ignored by code
11923 generation and optimization.
11925 '``llvm.annotation.*``' Intrinsic
11926 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11931 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
11932 any integer bit width.
11936 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
11937 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
11938 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
11939 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
11940 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
11945 The '``llvm.annotation``' intrinsic.
11950 The first argument is an integer value (result of some expression), the
11951 second is a pointer to a global string, the third is a pointer to a
11952 global string which is the source file name, and the last argument is
11953 the line number. It returns the value of the first argument.
11958 This intrinsic allows annotations to be put on arbitrary expressions
11959 with arbitrary strings. This can be useful for special purpose
11960 optimizations that want to look for these annotations. These have no
11961 other defined use; they are ignored by code generation and optimization.
11963 '``llvm.trap``' Intrinsic
11964 ^^^^^^^^^^^^^^^^^^^^^^^^^
11971 declare void @llvm.trap() noreturn nounwind
11976 The '``llvm.trap``' intrinsic.
11986 This intrinsic is lowered to the target dependent trap instruction. If
11987 the target does not have a trap instruction, this intrinsic will be
11988 lowered to a call of the ``abort()`` function.
11990 '``llvm.debugtrap``' Intrinsic
11991 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11998 declare void @llvm.debugtrap() nounwind
12003 The '``llvm.debugtrap``' intrinsic.
12013 This intrinsic is lowered to code which is intended to cause an
12014 execution trap with the intention of requesting the attention of a
12017 '``llvm.stackprotector``' Intrinsic
12018 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
12025 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
12030 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
12031 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
12032 is placed on the stack before local variables.
12037 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
12038 The first argument is the value loaded from the stack guard
12039 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
12040 enough space to hold the value of the guard.
12045 This intrinsic causes the prologue/epilogue inserter to force the position of
12046 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
12047 to ensure that if a local variable on the stack is overwritten, it will destroy
12048 the value of the guard. When the function exits, the guard on the stack is
12049 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
12050 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
12051 calling the ``__stack_chk_fail()`` function.
12053 '``llvm.stackprotectorcheck``' Intrinsic
12054 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
12061 declare void @llvm.stackprotectorcheck(i8** <guard>)
12066 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
12067 created stack protector and if they are not equal calls the
12068 ``__stack_chk_fail()`` function.
12073 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
12074 the variable ``@__stack_chk_guard``.
12079 This intrinsic is provided to perform the stack protector check by comparing
12080 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
12081 values do not match call the ``__stack_chk_fail()`` function.
12083 The reason to provide this as an IR level intrinsic instead of implementing it
12084 via other IR operations is that in order to perform this operation at the IR
12085 level without an intrinsic, one would need to create additional basic blocks to
12086 handle the success/failure cases. This makes it difficult to stop the stack
12087 protector check from disrupting sibling tail calls in Codegen. With this
12088 intrinsic, we are able to generate the stack protector basic blocks late in
12089 codegen after the tail call decision has occurred.
12091 '``llvm.objectsize``' Intrinsic
12092 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
12099 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
12100 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
12105 The ``llvm.objectsize`` intrinsic is designed to provide information to
12106 the optimizers to determine at compile time whether a) an operation
12107 (like memcpy) will overflow a buffer that corresponds to an object, or
12108 b) that a runtime check for overflow isn't necessary. An object in this
12109 context means an allocation of a specific class, structure, array, or
12115 The ``llvm.objectsize`` intrinsic takes two arguments. The first
12116 argument is a pointer to or into the ``object``. The second argument is
12117 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
12118 or -1 (if false) when the object size is unknown. The second argument
12119 only accepts constants.
12124 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
12125 the size of the object concerned. If the size cannot be determined at
12126 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
12127 on the ``min`` argument).
12129 '``llvm.expect``' Intrinsic
12130 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
12135 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
12140 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
12141 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
12142 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
12147 The ``llvm.expect`` intrinsic provides information about expected (the
12148 most probable) value of ``val``, which can be used by optimizers.
12153 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
12154 a value. The second argument is an expected value, this needs to be a
12155 constant value, variables are not allowed.
12160 This intrinsic is lowered to the ``val``.
12164 '``llvm.assume``' Intrinsic
12165 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
12172 declare void @llvm.assume(i1 %cond)
12177 The ``llvm.assume`` allows the optimizer to assume that the provided
12178 condition is true. This information can then be used in simplifying other parts
12184 The condition which the optimizer may assume is always true.
12189 The intrinsic allows the optimizer to assume that the provided condition is
12190 always true whenever the control flow reaches the intrinsic call. No code is
12191 generated for this intrinsic, and instructions that contribute only to the
12192 provided condition are not used for code generation. If the condition is
12193 violated during execution, the behavior is undefined.
12195 Note that the optimizer might limit the transformations performed on values
12196 used by the ``llvm.assume`` intrinsic in order to preserve the instructions
12197 only used to form the intrinsic's input argument. This might prove undesirable
12198 if the extra information provided by the ``llvm.assume`` intrinsic does not cause
12199 sufficient overall improvement in code quality. For this reason,
12200 ``llvm.assume`` should not be used to document basic mathematical invariants
12201 that the optimizer can otherwise deduce or facts that are of little use to the
12206 '``llvm.bitset.test``' Intrinsic
12207 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
12214 declare i1 @llvm.bitset.test(i8* %ptr, metadata %bitset) nounwind readnone
12220 The first argument is a pointer to be tested. The second argument is a
12221 metadata object representing an identifier for a :doc:`bitset <BitSets>`.
12226 The ``llvm.bitset.test`` intrinsic tests whether the given pointer is a
12227 member of the given bitset.
12229 '``llvm.donothing``' Intrinsic
12230 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
12237 declare void @llvm.donothing() nounwind readnone
12242 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's one of only
12243 two intrinsics (besides ``llvm.experimental.patchpoint``) that can be called
12244 with an invoke instruction.
12254 This intrinsic does nothing, and it's removed by optimizers and ignored
12257 Stack Map Intrinsics
12258 --------------------
12260 LLVM provides experimental intrinsics to support runtime patching
12261 mechanisms commonly desired in dynamic language JITs. These intrinsics
12262 are described in :doc:`StackMaps`.