1 ==============================
2 LLVM Language Reference Manual
3 ==============================
12 This document is a reference manual for the LLVM assembly language. LLVM
13 is a Static Single Assignment (SSA) based representation that provides
14 type safety, low-level operations, flexibility, and the capability of
15 representing 'all' high-level languages cleanly. It is the common code
16 representation used throughout all phases of the LLVM compilation
22 The LLVM code representation is designed to be used in three different
23 forms: as an in-memory compiler IR, as an on-disk bitcode representation
24 (suitable for fast loading by a Just-In-Time compiler), and as a human
25 readable assembly language representation. This allows LLVM to provide a
26 powerful intermediate representation for efficient compiler
27 transformations and analysis, while providing a natural means to debug
28 and visualize the transformations. The three different forms of LLVM are
29 all equivalent. This document describes the human readable
30 representation and notation.
32 The LLVM representation aims to be light-weight and low-level while
33 being expressive, typed, and extensible at the same time. It aims to be
34 a "universal IR" of sorts, by being at a low enough level that
35 high-level ideas may be cleanly mapped to it (similar to how
36 microprocessors are "universal IR's", allowing many source languages to
37 be mapped to them). By providing type information, LLVM can be used as
38 the target of optimizations: for example, through pointer analysis, it
39 can be proven that a C automatic variable is never accessed outside of
40 the current function, allowing it to be promoted to a simple SSA value
41 instead of a memory location.
48 It is important to note that this document describes 'well formed' LLVM
49 assembly language. There is a difference between what the parser accepts
50 and what is considered 'well formed'. For example, the following
51 instruction is syntactically okay, but not well formed:
57 because the definition of ``%x`` does not dominate all of its uses. The
58 LLVM infrastructure provides a verification pass that may be used to
59 verify that an LLVM module is well formed. This pass is automatically
60 run by the parser after parsing input assembly and by the optimizer
61 before it outputs bitcode. The violations pointed out by the verifier
62 pass indicate bugs in transformation passes or input to the parser.
69 LLVM identifiers come in two basic types: global and local. Global
70 identifiers (functions, global variables) begin with the ``'@'``
71 character. Local identifiers (register names, types) begin with the
72 ``'%'`` character. Additionally, there are three different formats for
73 identifiers, for different purposes:
75 #. Named values are represented as a string of characters with their
76 prefix. For example, ``%foo``, ``@DivisionByZero``,
77 ``%a.really.long.identifier``. The actual regular expression used is
78 '``[%@][-a-zA-Z$._][-a-zA-Z$._0-9]*``'. Identifiers that require other
79 characters in their names can be surrounded with quotes. Special
80 characters may be escaped using ``"\xx"`` where ``xx`` is the ASCII
81 code for the character in hexadecimal. In this way, any character can
82 be used in a name value, even quotes themselves. The ``"\01"`` prefix
83 can be used on global variables to suppress mangling.
84 #. Unnamed values are represented as an unsigned numeric value with
85 their prefix. For example, ``%12``, ``@2``, ``%44``.
86 #. Constants, which are described in the section Constants_ below.
88 LLVM requires that values start with a prefix for two reasons: Compilers
89 don't need to worry about name clashes with reserved words, and the set
90 of reserved words may be expanded in the future without penalty.
91 Additionally, unnamed identifiers allow a compiler to quickly come up
92 with a temporary variable without having to avoid symbol table
95 Reserved words in LLVM are very similar to reserved words in other
96 languages. There are keywords for different opcodes ('``add``',
97 '``bitcast``', '``ret``', etc...), for primitive type names ('``void``',
98 '``i32``', etc...), and others. These reserved words cannot conflict
99 with variable names, because none of them start with a prefix character
100 (``'%'`` or ``'@'``).
102 Here is an example of LLVM code to multiply the integer variable
109 %result = mul i32 %X, 8
111 After strength reduction:
115 %result = shl i32 %X, 3
121 %0 = add i32 %X, %X ; yields i32:%0
122 %1 = add i32 %0, %0 ; yields i32:%1
123 %result = add i32 %1, %1
125 This last way of multiplying ``%X`` by 8 illustrates several important
126 lexical features of LLVM:
128 #. Comments are delimited with a '``;``' and go until the end of line.
129 #. Unnamed temporaries are created when the result of a computation is
130 not assigned to a named value.
131 #. Unnamed temporaries are numbered sequentially (using a per-function
132 incrementing counter, starting with 0). Note that basic blocks and unnamed
133 function parameters are included in this numbering. For example, if the
134 entry basic block is not given a label name and all function parameters are
135 named, then it will get number 0.
137 It also shows a convention that we follow in this document. When
138 demonstrating instructions, we will follow an instruction with a comment
139 that defines the type and name of value produced.
147 LLVM programs are composed of ``Module``'s, each of which is a
148 translation unit of the input programs. Each module consists of
149 functions, global variables, and symbol table entries. Modules may be
150 combined together with the LLVM linker, which merges function (and
151 global variable) definitions, resolves forward declarations, and merges
152 symbol table entries. Here is an example of the "hello world" module:
156 ; Declare the string constant as a global constant.
157 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
159 ; External declaration of the puts function
160 declare i32 @puts(i8* nocapture) nounwind
162 ; Definition of main function
163 define i32 @main() { ; i32()*
164 ; Convert [13 x i8]* to i8 *...
165 %cast210 = getelementptr [13 x i8], [13 x i8]* @.str, i64 0, i64 0
167 ; Call puts function to write out the string to stdout.
168 call i32 @puts(i8* %cast210)
173 !0 = !{i32 42, null, !"string"}
176 This example is made up of a :ref:`global variable <globalvars>` named
177 "``.str``", an external declaration of the "``puts``" function, a
178 :ref:`function definition <functionstructure>` for "``main``" and
179 :ref:`named metadata <namedmetadatastructure>` "``foo``".
181 In general, a module is made up of a list of global values (where both
182 functions and global variables are global values). Global values are
183 represented by a pointer to a memory location (in this case, a pointer
184 to an array of char, and a pointer to a function), and have one of the
185 following :ref:`linkage types <linkage>`.
192 All Global Variables and Functions have one of the following types of
196 Global values with "``private``" linkage are only directly
197 accessible by objects in the current module. In particular, linking
198 code into a module with an private global value may cause the
199 private to be renamed as necessary to avoid collisions. Because the
200 symbol is private to the module, all references can be updated. This
201 doesn't show up in any symbol table in the object file.
203 Similar to private, but the value shows as a local symbol
204 (``STB_LOCAL`` in the case of ELF) in the object file. This
205 corresponds to the notion of the '``static``' keyword in C.
206 ``available_externally``
207 Globals with "``available_externally``" linkage are never emitted
208 into the object file corresponding to the LLVM module. They exist to
209 allow inlining and other optimizations to take place given knowledge
210 of the definition of the global, which is known to be somewhere
211 outside the module. Globals with ``available_externally`` linkage
212 are allowed to be discarded at will, and are otherwise the same as
213 ``linkonce_odr``. This linkage type is only allowed on definitions,
216 Globals with "``linkonce``" linkage are merged with other globals of
217 the same name when linkage occurs. This can be used to implement
218 some forms of inline functions, templates, or other code which must
219 be generated in each translation unit that uses it, but where the
220 body may be overridden with a more definitive definition later.
221 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
222 that ``linkonce`` linkage does not actually allow the optimizer to
223 inline the body of this function into callers because it doesn't
224 know if this definition of the function is the definitive definition
225 within the program or whether it will be overridden by a stronger
226 definition. To enable inlining and other optimizations, use
227 "``linkonce_odr``" linkage.
229 "``weak``" linkage has the same merging semantics as ``linkonce``
230 linkage, except that unreferenced globals with ``weak`` linkage may
231 not be discarded. This is used for globals that are declared "weak"
234 "``common``" linkage is most similar to "``weak``" linkage, but they
235 are used for tentative definitions in C, such as "``int X;``" at
236 global scope. Symbols with "``common``" linkage are merged in the
237 same way as ``weak symbols``, and they may not be deleted if
238 unreferenced. ``common`` symbols may not have an explicit section,
239 must have a zero initializer, and may not be marked
240 ':ref:`constant <globalvars>`'. Functions and aliases may not have
243 .. _linkage_appending:
246 "``appending``" linkage may only be applied to global variables of
247 pointer to array type. When two global variables with appending
248 linkage are linked together, the two global arrays are appended
249 together. This is the LLVM, typesafe, equivalent of having the
250 system linker append together "sections" with identical names when
253 The semantics of this linkage follow the ELF object file model: the
254 symbol is weak until linked, if not linked, the symbol becomes null
255 instead of being an undefined reference.
256 ``linkonce_odr``, ``weak_odr``
257 Some languages allow differing globals to be merged, such as two
258 functions with different semantics. Other languages, such as
259 ``C++``, ensure that only equivalent globals are ever merged (the
260 "one definition rule" --- "ODR"). Such languages can use the
261 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
262 global will only be merged with equivalent globals. These linkage
263 types are otherwise the same as their non-``odr`` versions.
265 If none of the above identifiers are used, the global is externally
266 visible, meaning that it participates in linkage and can be used to
267 resolve external symbol references.
269 It is illegal for a function *declaration* to have any linkage type
270 other than ``external`` or ``extern_weak``.
277 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
278 :ref:`invokes <i_invoke>` can all have an optional calling convention
279 specified for the call. The calling convention of any pair of dynamic
280 caller/callee must match, or the behavior of the program is undefined.
281 The following calling conventions are supported by LLVM, and more may be
284 "``ccc``" - The C calling convention
285 This calling convention (the default if no other calling convention
286 is specified) matches the target C calling conventions. This calling
287 convention supports varargs function calls and tolerates some
288 mismatch in the declared prototype and implemented declaration of
289 the function (as does normal C).
290 "``fastcc``" - The fast calling convention
291 This calling convention attempts to make calls as fast as possible
292 (e.g. by passing things in registers). This calling convention
293 allows the target to use whatever tricks it wants to produce fast
294 code for the target, without having to conform to an externally
295 specified ABI (Application Binary Interface). `Tail calls can only
296 be optimized when this, the GHC or the HiPE convention is
297 used. <CodeGenerator.html#id80>`_ This calling convention does not
298 support varargs and requires the prototype of all callees to exactly
299 match the prototype of the function definition.
300 "``coldcc``" - The cold calling convention
301 This calling convention attempts to make code in the caller as
302 efficient as possible under the assumption that the call is not
303 commonly executed. As such, these calls often preserve all registers
304 so that the call does not break any live ranges in the caller side.
305 This calling convention does not support varargs and requires the
306 prototype of all callees to exactly match the prototype of the
307 function definition. Furthermore the inliner doesn't consider such function
309 "``cc 10``" - GHC convention
310 This calling convention has been implemented specifically for use by
311 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
312 It passes everything in registers, going to extremes to achieve this
313 by disabling callee save registers. This calling convention should
314 not be used lightly but only for specific situations such as an
315 alternative to the *register pinning* performance technique often
316 used when implementing functional programming languages. At the
317 moment only X86 supports this convention and it has the following
320 - On *X86-32* only supports up to 4 bit type parameters. No
321 floating point types are supported.
322 - On *X86-64* only supports up to 10 bit type parameters and 6
323 floating point parameters.
325 This calling convention supports `tail call
326 optimization <CodeGenerator.html#id80>`_ but requires both the
327 caller and callee are using it.
328 "``cc 11``" - The HiPE calling convention
329 This calling convention has been implemented specifically for use by
330 the `High-Performance Erlang
331 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
332 native code compiler of the `Ericsson's Open Source Erlang/OTP
333 system <http://www.erlang.org/download.shtml>`_. It uses more
334 registers for argument passing than the ordinary C calling
335 convention and defines no callee-saved registers. The calling
336 convention properly supports `tail call
337 optimization <CodeGenerator.html#id80>`_ but requires that both the
338 caller and the callee use it. It uses a *register pinning*
339 mechanism, similar to GHC's convention, for keeping frequently
340 accessed runtime components pinned to specific hardware registers.
341 At the moment only X86 supports this convention (both 32 and 64
343 "``webkit_jscc``" - WebKit's JavaScript calling convention
344 This calling convention has been implemented for `WebKit FTL JIT
345 <https://trac.webkit.org/wiki/FTLJIT>`_. It passes arguments on the
346 stack right to left (as cdecl does), and returns a value in the
347 platform's customary return register.
348 "``anyregcc``" - Dynamic calling convention for code patching
349 This is a special convention that supports patching an arbitrary code
350 sequence in place of a call site. This convention forces the call
351 arguments into registers but allows them to be dynamically
352 allocated. This can currently only be used with calls to
353 llvm.experimental.patchpoint because only this intrinsic records
354 the location of its arguments in a side table. See :doc:`StackMaps`.
355 "``preserve_mostcc``" - The `PreserveMost` calling convention
356 This calling convention attempts to make the code in the caller as
357 unintrusive as possible. This convention behaves identically to the `C`
358 calling convention on how arguments and return values are passed, but it
359 uses a different set of caller/callee-saved registers. This alleviates the
360 burden of saving and recovering a large register set before and after the
361 call in the caller. If the arguments are passed in callee-saved registers,
362 then they will be preserved by the callee across the call. This doesn't
363 apply for values returned in callee-saved registers.
365 - On X86-64 the callee preserves all general purpose registers, except for
366 R11. R11 can be used as a scratch register. Floating-point registers
367 (XMMs/YMMs) are not preserved and need to be saved by the caller.
369 The idea behind this convention is to support calls to runtime functions
370 that have a hot path and a cold path. The hot path is usually a small piece
371 of code that doesn't use many registers. The cold path might need to call out to
372 another function and therefore only needs to preserve the caller-saved
373 registers, which haven't already been saved by the caller. The
374 `PreserveMost` calling convention is very similar to the `cold` calling
375 convention in terms of caller/callee-saved registers, but they are used for
376 different types of function calls. `coldcc` is for function calls that are
377 rarely executed, whereas `preserve_mostcc` function calls are intended to be
378 on the hot path and definitely executed a lot. Furthermore `preserve_mostcc`
379 doesn't prevent the inliner from inlining the function call.
381 This calling convention will be used by a future version of the ObjectiveC
382 runtime and should therefore still be considered experimental at this time.
383 Although this convention was created to optimize certain runtime calls to
384 the ObjectiveC runtime, it is not limited to this runtime and might be used
385 by other runtimes in the future too. The current implementation only
386 supports X86-64, but the intention is to support more architectures in the
388 "``preserve_allcc``" - The `PreserveAll` calling convention
389 This calling convention attempts to make the code in the caller even less
390 intrusive than the `PreserveMost` calling convention. This calling
391 convention also behaves identical to the `C` calling convention on how
392 arguments and return values are passed, but it uses a different set of
393 caller/callee-saved registers. This removes the burden of saving and
394 recovering a large register set before and after the call in the caller. If
395 the arguments are passed in callee-saved registers, then they will be
396 preserved by the callee across the call. This doesn't apply for values
397 returned in callee-saved registers.
399 - On X86-64 the callee preserves all general purpose registers, except for
400 R11. R11 can be used as a scratch register. Furthermore it also preserves
401 all floating-point registers (XMMs/YMMs).
403 The idea behind this convention is to support calls to runtime functions
404 that don't need to call out to any other functions.
406 This calling convention, like the `PreserveMost` calling convention, will be
407 used by a future version of the ObjectiveC runtime and should be considered
408 experimental at this time.
409 "``cc <n>``" - Numbered convention
410 Any calling convention may be specified by number, allowing
411 target-specific calling conventions to be used. Target specific
412 calling conventions start at 64.
414 More calling conventions can be added/defined on an as-needed basis, to
415 support Pascal conventions or any other well-known target-independent
418 .. _visibilitystyles:
423 All Global Variables and Functions have one of the following visibility
426 "``default``" - Default style
427 On targets that use the ELF object file format, default visibility
428 means that the declaration is visible to other modules and, in
429 shared libraries, means that the declared entity may be overridden.
430 On Darwin, default visibility means that the declaration is visible
431 to other modules. Default visibility corresponds to "external
432 linkage" in the language.
433 "``hidden``" - Hidden style
434 Two declarations of an object with hidden visibility refer to the
435 same object if they are in the same shared object. Usually, hidden
436 visibility indicates that the symbol will not be placed into the
437 dynamic symbol table, so no other module (executable or shared
438 library) can reference it directly.
439 "``protected``" - Protected style
440 On ELF, protected visibility indicates that the symbol will be
441 placed in the dynamic symbol table, but that references within the
442 defining module will bind to the local symbol. That is, the symbol
443 cannot be overridden by another module.
445 A symbol with ``internal`` or ``private`` linkage must have ``default``
453 All Global Variables, Functions and Aliases can have one of the following
457 "``dllimport``" causes the compiler to reference a function or variable via
458 a global pointer to a pointer that is set up by the DLL exporting the
459 symbol. On Microsoft Windows targets, the pointer name is formed by
460 combining ``__imp_`` and the function or variable name.
462 "``dllexport``" causes the compiler to provide a global pointer to a pointer
463 in a DLL, so that it can be referenced with the ``dllimport`` attribute. On
464 Microsoft Windows targets, the pointer name is formed by combining
465 ``__imp_`` and the function or variable name. Since this storage class
466 exists for defining a dll interface, the compiler, assembler and linker know
467 it is externally referenced and must refrain from deleting the symbol.
471 Thread Local Storage Models
472 ---------------------------
474 A variable may be defined as ``thread_local``, which means that it will
475 not be shared by threads (each thread will have a separated copy of the
476 variable). Not all targets support thread-local variables. Optionally, a
477 TLS model may be specified:
480 For variables that are only used within the current shared library.
482 For variables in modules that will not be loaded dynamically.
484 For variables defined in the executable and only used within it.
486 If no explicit model is given, the "general dynamic" model is used.
488 The models correspond to the ELF TLS models; see `ELF Handling For
489 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
490 more information on under which circumstances the different models may
491 be used. The target may choose a different TLS model if the specified
492 model is not supported, or if a better choice of model can be made.
494 A model can also be specified in a alias, but then it only governs how
495 the alias is accessed. It will not have any effect in the aliasee.
497 For platforms without linker support of ELF TLS model, the -femulated-tls
498 flag can be used to generate GCC compatible emulated TLS code.
505 LLVM IR allows you to specify both "identified" and "literal" :ref:`structure
506 types <t_struct>`. Literal types are uniqued structurally, but identified types
507 are never uniqued. An :ref:`opaque structural type <t_opaque>` can also be used
508 to forward declare a type that is not yet available.
510 An example of a identified structure specification is:
514 %mytype = type { %mytype*, i32 }
516 Prior to the LLVM 3.0 release, identified types were structurally uniqued. Only
517 literal types are uniqued in recent versions of LLVM.
524 Global variables define regions of memory allocated at compilation time
527 Global variable definitions must be initialized.
529 Global variables in other translation units can also be declared, in which
530 case they don't have an initializer.
532 Either global variable definitions or declarations may have an explicit section
533 to be placed in and may have an optional explicit alignment specified.
535 A variable may be defined as a global ``constant``, which indicates that
536 the contents of the variable will **never** be modified (enabling better
537 optimization, allowing the global data to be placed in the read-only
538 section of an executable, etc). Note that variables that need runtime
539 initialization cannot be marked ``constant`` as there is a store to the
542 LLVM explicitly allows *declarations* of global variables to be marked
543 constant, even if the final definition of the global is not. This
544 capability can be used to enable slightly better optimization of the
545 program, but requires the language definition to guarantee that
546 optimizations based on the 'constantness' are valid for the translation
547 units that do not include the definition.
549 As SSA values, global variables define pointer values that are in scope
550 (i.e. they dominate) all basic blocks in the program. Global variables
551 always define a pointer to their "content" type because they describe a
552 region of memory, and all memory objects in LLVM are accessed through
555 Global variables can be marked with ``unnamed_addr`` which indicates
556 that the address is not significant, only the content. Constants marked
557 like this can be merged with other constants if they have the same
558 initializer. Note that a constant with significant address *can* be
559 merged with a ``unnamed_addr`` constant, the result being a constant
560 whose address is significant.
562 A global variable may be declared to reside in a target-specific
563 numbered address space. For targets that support them, address spaces
564 may affect how optimizations are performed and/or what target
565 instructions are used to access the variable. The default address space
566 is zero. The address space qualifier must precede any other attributes.
568 LLVM allows an explicit section to be specified for globals. If the
569 target supports it, it will emit globals to the section specified.
570 Additionally, the global can placed in a comdat if the target has the necessary
573 By default, global initializers are optimized by assuming that global
574 variables defined within the module are not modified from their
575 initial values before the start of the global initializer. This is
576 true even for variables potentially accessible from outside the
577 module, including those with external linkage or appearing in
578 ``@llvm.used`` or dllexported variables. This assumption may be suppressed
579 by marking the variable with ``externally_initialized``.
581 An explicit alignment may be specified for a global, which must be a
582 power of 2. If not present, or if the alignment is set to zero, the
583 alignment of the global is set by the target to whatever it feels
584 convenient. If an explicit alignment is specified, the global is forced
585 to have exactly that alignment. Targets and optimizers are not allowed
586 to over-align the global if the global has an assigned section. In this
587 case, the extra alignment could be observable: for example, code could
588 assume that the globals are densely packed in their section and try to
589 iterate over them as an array, alignment padding would break this
590 iteration. The maximum alignment is ``1 << 29``.
592 Globals can also have a :ref:`DLL storage class <dllstorageclass>`.
594 Variables and aliases can have a
595 :ref:`Thread Local Storage Model <tls_model>`.
599 [@<GlobalVarName> =] [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal]
600 [unnamed_addr] [AddrSpace] [ExternallyInitialized]
601 <global | constant> <Type> [<InitializerConstant>]
602 [, section "name"] [, comdat [($name)]]
603 [, align <Alignment>]
605 For example, the following defines a global in a numbered address space
606 with an initializer, section, and alignment:
610 @G = addrspace(5) constant float 1.0, section "foo", align 4
612 The following example just declares a global variable
616 @G = external global i32
618 The following example defines a thread-local global with the
619 ``initialexec`` TLS model:
623 @G = thread_local(initialexec) global i32 0, align 4
625 .. _functionstructure:
630 LLVM function definitions consist of the "``define``" keyword, an
631 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
632 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
633 an optional :ref:`calling convention <callingconv>`,
634 an optional ``unnamed_addr`` attribute, a return type, an optional
635 :ref:`parameter attribute <paramattrs>` for the return type, a function
636 name, a (possibly empty) argument list (each with optional :ref:`parameter
637 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
638 an optional section, an optional alignment,
639 an optional :ref:`comdat <langref_comdats>`,
640 an optional :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
641 an optional :ref:`prologue <prologuedata>`,
642 an optional :ref:`personality <personalityfn>`,
643 an opening curly brace, a list of basic blocks, and a closing curly brace.
645 LLVM function declarations consist of the "``declare``" keyword, an
646 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
647 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
648 an optional :ref:`calling convention <callingconv>`,
649 an optional ``unnamed_addr`` attribute, a return type, an optional
650 :ref:`parameter attribute <paramattrs>` for the return type, a function
651 name, a possibly empty list of arguments, an optional alignment, an optional
652 :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`,
653 and an optional :ref:`prologue <prologuedata>`.
655 A function definition contains a list of basic blocks, forming the CFG (Control
656 Flow Graph) for the function. Each basic block may optionally start with a label
657 (giving the basic block a symbol table entry), contains a list of instructions,
658 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
659 function return). If an explicit label is not provided, a block is assigned an
660 implicit numbered label, using the next value from the same counter as used for
661 unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
662 entry block does not have an explicit label, it will be assigned label "%0",
663 then the first unnamed temporary in that block will be "%1", etc.
665 The first basic block in a function is special in two ways: it is
666 immediately executed on entrance to the function, and it is not allowed
667 to have predecessor basic blocks (i.e. there can not be any branches to
668 the entry block of a function). Because the block can have no
669 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
671 LLVM allows an explicit section to be specified for functions. If the
672 target supports it, it will emit functions to the section specified.
673 Additionally, the function can be placed in a COMDAT.
675 An explicit alignment may be specified for a function. If not present,
676 or if the alignment is set to zero, the alignment of the function is set
677 by the target to whatever it feels convenient. If an explicit alignment
678 is specified, the function is forced to have at least that much
679 alignment. All alignments must be a power of 2.
681 If the ``unnamed_addr`` attribute is given, the address is known to not
682 be significant and two identical functions can be merged.
686 define [linkage] [visibility] [DLLStorageClass]
688 <ResultType> @<FunctionName> ([argument list])
689 [unnamed_addr] [fn Attrs] [section "name"] [comdat [($name)]]
690 [align N] [gc] [prefix Constant] [prologue Constant]
691 [personality Constant] { ... }
693 The argument list is a comma seperated sequence of arguments where each
694 argument is of the following form
698 <type> [parameter Attrs] [name]
706 Aliases, unlike function or variables, don't create any new data. They
707 are just a new symbol and metadata for an existing position.
709 Aliases have a name and an aliasee that is either a global value or a
712 Aliases may have an optional :ref:`linkage type <linkage>`, an optional
713 :ref:`visibility style <visibility>`, an optional :ref:`DLL storage class
714 <dllstorageclass>` and an optional :ref:`tls model <tls_model>`.
718 @<Name> = [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal] [unnamed_addr] alias <AliaseeTy> @<Aliasee>
720 The linkage must be one of ``private``, ``internal``, ``linkonce``, ``weak``,
721 ``linkonce_odr``, ``weak_odr``, ``external``. Note that some system linkers
722 might not correctly handle dropping a weak symbol that is aliased.
724 Aliases that are not ``unnamed_addr`` are guaranteed to have the same address as
725 the aliasee expression. ``unnamed_addr`` ones are only guaranteed to point
728 Since aliases are only a second name, some restrictions apply, of which
729 some can only be checked when producing an object file:
731 * The expression defining the aliasee must be computable at assembly
732 time. Since it is just a name, no relocations can be used.
734 * No alias in the expression can be weak as the possibility of the
735 intermediate alias being overridden cannot be represented in an
738 * No global value in the expression can be a declaration, since that
739 would require a relocation, which is not possible.
746 Comdat IR provides access to COFF and ELF object file COMDAT functionality.
748 Comdats have a name which represents the COMDAT key. All global objects that
749 specify this key will only end up in the final object file if the linker chooses
750 that key over some other key. Aliases are placed in the same COMDAT that their
751 aliasee computes to, if any.
753 Comdats have a selection kind to provide input on how the linker should
754 choose between keys in two different object files.
758 $<Name> = comdat SelectionKind
760 The selection kind must be one of the following:
763 The linker may choose any COMDAT key, the choice is arbitrary.
765 The linker may choose any COMDAT key but the sections must contain the
768 The linker will choose the section containing the largest COMDAT key.
770 The linker requires that only section with this COMDAT key exist.
772 The linker may choose any COMDAT key but the sections must contain the
775 Note that the Mach-O platform doesn't support COMDATs and ELF only supports
776 ``any`` as a selection kind.
778 Here is an example of a COMDAT group where a function will only be selected if
779 the COMDAT key's section is the largest:
783 $foo = comdat largest
784 @foo = global i32 2, comdat($foo)
786 define void @bar() comdat($foo) {
790 As a syntactic sugar the ``$name`` can be omitted if the name is the same as
796 @foo = global i32 2, comdat
799 In a COFF object file, this will create a COMDAT section with selection kind
800 ``IMAGE_COMDAT_SELECT_LARGEST`` containing the contents of the ``@foo`` symbol
801 and another COMDAT section with selection kind
802 ``IMAGE_COMDAT_SELECT_ASSOCIATIVE`` which is associated with the first COMDAT
803 section and contains the contents of the ``@bar`` symbol.
805 There are some restrictions on the properties of the global object.
806 It, or an alias to it, must have the same name as the COMDAT group when
808 The contents and size of this object may be used during link-time to determine
809 which COMDAT groups get selected depending on the selection kind.
810 Because the name of the object must match the name of the COMDAT group, the
811 linkage of the global object must not be local; local symbols can get renamed
812 if a collision occurs in the symbol table.
814 The combined use of COMDATS and section attributes may yield surprising results.
821 @g1 = global i32 42, section "sec", comdat($foo)
822 @g2 = global i32 42, section "sec", comdat($bar)
824 From the object file perspective, this requires the creation of two sections
825 with the same name. This is necessary because both globals belong to different
826 COMDAT groups and COMDATs, at the object file level, are represented by
829 Note that certain IR constructs like global variables and functions may
830 create COMDATs in the object file in addition to any which are specified using
831 COMDAT IR. This arises when the code generator is configured to emit globals
832 in individual sections (e.g. when `-data-sections` or `-function-sections`
833 is supplied to `llc`).
835 .. _namedmetadatastructure:
840 Named metadata is a collection of metadata. :ref:`Metadata
841 nodes <metadata>` (but not metadata strings) are the only valid
842 operands for a named metadata.
844 #. Named metadata are represented as a string of characters with the
845 metadata prefix. The rules for metadata names are the same as for
846 identifiers, but quoted names are not allowed. ``"\xx"`` type escapes
847 are still valid, which allows any character to be part of a name.
851 ; Some unnamed metadata nodes, which are referenced by the named metadata.
856 !name = !{!0, !1, !2}
863 The return type and each parameter of a function type may have a set of
864 *parameter attributes* associated with them. Parameter attributes are
865 used to communicate additional information about the result or
866 parameters of a function. Parameter attributes are considered to be part
867 of the function, not of the function type, so functions with different
868 parameter attributes can have the same function type.
870 Parameter attributes are simple keywords that follow the type specified.
871 If multiple parameter attributes are needed, they are space separated.
876 declare i32 @printf(i8* noalias nocapture, ...)
877 declare i32 @atoi(i8 zeroext)
878 declare signext i8 @returns_signed_char()
880 Note that any attributes for the function result (``nounwind``,
881 ``readonly``) come immediately after the argument list.
883 Currently, only the following parameter attributes are defined:
886 This indicates to the code generator that the parameter or return
887 value should be zero-extended to the extent required by the target's
888 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
889 the caller (for a parameter) or the callee (for a return value).
891 This indicates to the code generator that the parameter or return
892 value should be sign-extended to the extent required by the target's
893 ABI (which is usually 32-bits) by the caller (for a parameter) or
894 the callee (for a return value).
896 This indicates that this parameter or return value should be treated
897 in a special target-dependent fashion during while emitting code for
898 a function call or return (usually, by putting it in a register as
899 opposed to memory, though some targets use it to distinguish between
900 two different kinds of registers). Use of this attribute is
903 This indicates that the pointer parameter should really be passed by
904 value to the function. The attribute implies that a hidden copy of
905 the pointee is made between the caller and the callee, so the callee
906 is unable to modify the value in the caller. This attribute is only
907 valid on LLVM pointer arguments. It is generally used to pass
908 structs and arrays by value, but is also valid on pointers to
909 scalars. The copy is considered to belong to the caller not the
910 callee (for example, ``readonly`` functions should not write to
911 ``byval`` parameters). This is not a valid attribute for return
914 The byval attribute also supports specifying an alignment with the
915 align attribute. It indicates the alignment of the stack slot to
916 form and the known alignment of the pointer specified to the call
917 site. If the alignment is not specified, then the code generator
918 makes a target-specific assumption.
924 The ``inalloca`` argument attribute allows the caller to take the
925 address of outgoing stack arguments. An ``inalloca`` argument must
926 be a pointer to stack memory produced by an ``alloca`` instruction.
927 The alloca, or argument allocation, must also be tagged with the
928 inalloca keyword. Only the last argument may have the ``inalloca``
929 attribute, and that argument is guaranteed to be passed in memory.
931 An argument allocation may be used by a call at most once because
932 the call may deallocate it. The ``inalloca`` attribute cannot be
933 used in conjunction with other attributes that affect argument
934 storage, like ``inreg``, ``nest``, ``sret``, or ``byval``. The
935 ``inalloca`` attribute also disables LLVM's implicit lowering of
936 large aggregate return values, which means that frontend authors
937 must lower them with ``sret`` pointers.
939 When the call site is reached, the argument allocation must have
940 been the most recent stack allocation that is still live, or the
941 results are undefined. It is possible to allocate additional stack
942 space after an argument allocation and before its call site, but it
943 must be cleared off with :ref:`llvm.stackrestore
946 See :doc:`InAlloca` for more information on how to use this
950 This indicates that the pointer parameter specifies the address of a
951 structure that is the return value of the function in the source
952 program. This pointer must be guaranteed by the caller to be valid:
953 loads and stores to the structure may be assumed by the callee
954 not to trap and to be properly aligned. This may only be applied to
955 the first parameter. This is not a valid attribute for return
959 This indicates that the pointer value may be assumed by the optimizer to
960 have the specified alignment.
962 Note that this attribute has additional semantics when combined with the
968 This indicates that objects accessed via pointer values
969 :ref:`based <pointeraliasing>` on the argument or return value are not also
970 accessed, during the execution of the function, via pointer values not
971 *based* on the argument or return value. The attribute on a return value
972 also has additional semantics described below. The caller shares the
973 responsibility with the callee for ensuring that these requirements are met.
974 For further details, please see the discussion of the NoAlias response in
975 :ref:`alias analysis <Must, May, or No>`.
977 Note that this definition of ``noalias`` is intentionally similar
978 to the definition of ``restrict`` in C99 for function arguments.
980 For function return values, C99's ``restrict`` is not meaningful,
981 while LLVM's ``noalias`` is. Furthermore, the semantics of the ``noalias``
982 attribute on return values are stronger than the semantics of the attribute
983 when used on function arguments. On function return values, the ``noalias``
984 attribute indicates that the function acts like a system memory allocation
985 function, returning a pointer to allocated storage disjoint from the
986 storage for any other object accessible to the caller.
989 This indicates that the callee does not make any copies of the
990 pointer that outlive the callee itself. This is not a valid
991 attribute for return values.
996 This indicates that the pointer parameter can be excised using the
997 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
998 attribute for return values and can only be applied to one parameter.
1001 This indicates that the function always returns the argument as its return
1002 value. This is an optimization hint to the code generator when generating
1003 the caller, allowing tail call optimization and omission of register saves
1004 and restores in some cases; it is not checked or enforced when generating
1005 the callee. The parameter and the function return type must be valid
1006 operands for the :ref:`bitcast instruction <i_bitcast>`. This is not a
1007 valid attribute for return values and can only be applied to one parameter.
1010 This indicates that the parameter or return pointer is not null. This
1011 attribute may only be applied to pointer typed parameters. This is not
1012 checked or enforced by LLVM, the caller must ensure that the pointer
1013 passed in is non-null, or the callee must ensure that the returned pointer
1016 ``dereferenceable(<n>)``
1017 This indicates that the parameter or return pointer is dereferenceable. This
1018 attribute may only be applied to pointer typed parameters. A pointer that
1019 is dereferenceable can be loaded from speculatively without a risk of
1020 trapping. The number of bytes known to be dereferenceable must be provided
1021 in parentheses. It is legal for the number of bytes to be less than the
1022 size of the pointee type. The ``nonnull`` attribute does not imply
1023 dereferenceability (consider a pointer to one element past the end of an
1024 array), however ``dereferenceable(<n>)`` does imply ``nonnull`` in
1025 ``addrspace(0)`` (which is the default address space).
1027 ``dereferenceable_or_null(<n>)``
1028 This indicates that the parameter or return value isn't both
1029 non-null and non-dereferenceable (up to ``<n>`` bytes) at the same
1030 time. All non-null pointers tagged with
1031 ``dereferenceable_or_null(<n>)`` are ``dereferenceable(<n>)``.
1032 For address space 0 ``dereferenceable_or_null(<n>)`` implies that
1033 a pointer is exactly one of ``dereferenceable(<n>)`` or ``null``,
1034 and in other address spaces ``dereferenceable_or_null(<n>)``
1035 implies that a pointer is at least one of ``dereferenceable(<n>)``
1036 or ``null`` (i.e. it may be both ``null`` and
1037 ``dereferenceable(<n>)``). This attribute may only be applied to
1038 pointer typed parameters.
1042 Garbage Collector Strategy Names
1043 --------------------------------
1045 Each function may specify a garbage collector strategy name, which is simply a
1048 .. code-block:: llvm
1050 define void @f() gc "name" { ... }
1052 The supported values of *name* includes those :ref:`built in to LLVM
1053 <builtin-gc-strategies>` and any provided by loaded plugins. Specifying a GC
1054 strategy will cause the compiler to alter its output in order to support the
1055 named garbage collection algorithm. Note that LLVM itself does not contain a
1056 garbage collector, this functionality is restricted to generating machine code
1057 which can interoperate with a collector provided externally.
1064 Prefix data is data associated with a function which the code
1065 generator will emit immediately before the function's entrypoint.
1066 The purpose of this feature is to allow frontends to associate
1067 language-specific runtime metadata with specific functions and make it
1068 available through the function pointer while still allowing the
1069 function pointer to be called.
1071 To access the data for a given function, a program may bitcast the
1072 function pointer to a pointer to the constant's type and dereference
1073 index -1. This implies that the IR symbol points just past the end of
1074 the prefix data. For instance, take the example of a function annotated
1075 with a single ``i32``,
1077 .. code-block:: llvm
1079 define void @f() prefix i32 123 { ... }
1081 The prefix data can be referenced as,
1083 .. code-block:: llvm
1085 %0 = bitcast void* () @f to i32*
1086 %a = getelementptr inbounds i32, i32* %0, i32 -1
1087 %b = load i32, i32* %a
1089 Prefix data is laid out as if it were an initializer for a global variable
1090 of the prefix data's type. The function will be placed such that the
1091 beginning of the prefix data is aligned. This means that if the size
1092 of the prefix data is not a multiple of the alignment size, the
1093 function's entrypoint will not be aligned. If alignment of the
1094 function's entrypoint is desired, padding must be added to the prefix
1097 A function may have prefix data but no body. This has similar semantics
1098 to the ``available_externally`` linkage in that the data may be used by the
1099 optimizers but will not be emitted in the object file.
1106 The ``prologue`` attribute allows arbitrary code (encoded as bytes) to
1107 be inserted prior to the function body. This can be used for enabling
1108 function hot-patching and instrumentation.
1110 To maintain the semantics of ordinary function calls, the prologue data must
1111 have a particular format. Specifically, it must begin with a sequence of
1112 bytes which decode to a sequence of machine instructions, valid for the
1113 module's target, which transfer control to the point immediately succeeding
1114 the prologue data, without performing any other visible action. This allows
1115 the inliner and other passes to reason about the semantics of the function
1116 definition without needing to reason about the prologue data. Obviously this
1117 makes the format of the prologue data highly target dependent.
1119 A trivial example of valid prologue data for the x86 architecture is ``i8 144``,
1120 which encodes the ``nop`` instruction:
1122 .. code-block:: llvm
1124 define void @f() prologue i8 144 { ... }
1126 Generally prologue data can be formed by encoding a relative branch instruction
1127 which skips the metadata, as in this example of valid prologue data for the
1128 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
1130 .. code-block:: llvm
1132 %0 = type <{ i8, i8, i8* }>
1134 define void @f() prologue %0 <{ i8 235, i8 8, i8* @md}> { ... }
1136 A function may have prologue data but no body. This has similar semantics
1137 to the ``available_externally`` linkage in that the data may be used by the
1138 optimizers but will not be emitted in the object file.
1142 Personality Function
1143 --------------------
1145 The ``personality`` attribute permits functions to specify what function
1146 to use for exception handling.
1153 Attribute groups are groups of attributes that are referenced by objects within
1154 the IR. They are important for keeping ``.ll`` files readable, because a lot of
1155 functions will use the same set of attributes. In the degenerative case of a
1156 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
1157 group will capture the important command line flags used to build that file.
1159 An attribute group is a module-level object. To use an attribute group, an
1160 object references the attribute group's ID (e.g. ``#37``). An object may refer
1161 to more than one attribute group. In that situation, the attributes from the
1162 different groups are merged.
1164 Here is an example of attribute groups for a function that should always be
1165 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
1167 .. code-block:: llvm
1169 ; Target-independent attributes:
1170 attributes #0 = { alwaysinline alignstack=4 }
1172 ; Target-dependent attributes:
1173 attributes #1 = { "no-sse" }
1175 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
1176 define void @f() #0 #1 { ... }
1183 Function attributes are set to communicate additional information about
1184 a function. Function attributes are considered to be part of the
1185 function, not of the function type, so functions with different function
1186 attributes can have the same function type.
1188 Function attributes are simple keywords that follow the type specified.
1189 If multiple attributes are needed, they are space separated. For
1192 .. code-block:: llvm
1194 define void @f() noinline { ... }
1195 define void @f() alwaysinline { ... }
1196 define void @f() alwaysinline optsize { ... }
1197 define void @f() optsize { ... }
1200 This attribute indicates that, when emitting the prologue and
1201 epilogue, the backend should forcibly align the stack pointer.
1202 Specify the desired alignment, which must be a power of two, in
1205 This attribute indicates that the inliner should attempt to inline
1206 this function into callers whenever possible, ignoring any active
1207 inlining size threshold for this caller.
1209 This indicates that the callee function at a call site should be
1210 recognized as a built-in function, even though the function's declaration
1211 uses the ``nobuiltin`` attribute. This is only valid at call sites for
1212 direct calls to functions that are declared with the ``nobuiltin``
1215 This attribute indicates that this function is rarely called. When
1216 computing edge weights, basic blocks post-dominated by a cold
1217 function call are also considered to be cold; and, thus, given low
1220 This attribute indicates that the callee is dependent on a convergent
1221 thread execution pattern under certain parallel execution models.
1222 Transformations that are execution model agnostic may only move or
1223 tranform this call if the final location is control equivalent to its
1224 original position in the program, where control equivalence is defined as
1225 A dominates B and B post-dominates A, or vice versa.
1227 This attribute indicates that the source code contained a hint that
1228 inlining this function is desirable (such as the "inline" keyword in
1229 C/C++). It is just a hint; it imposes no requirements on the
1232 This attribute indicates that the function should be added to a
1233 jump-instruction table at code-generation time, and that all address-taken
1234 references to this function should be replaced with a reference to the
1235 appropriate jump-instruction-table function pointer. Note that this creates
1236 a new pointer for the original function, which means that code that depends
1237 on function-pointer identity can break. So, any function annotated with
1238 ``jumptable`` must also be ``unnamed_addr``.
1240 This attribute suggests that optimization passes and code generator
1241 passes make choices that keep the code size of this function as small
1242 as possible and perform optimizations that may sacrifice runtime
1243 performance in order to minimize the size of the generated code.
1245 This attribute disables prologue / epilogue emission for the
1246 function. This can have very system-specific consequences.
1248 This indicates that the callee function at a call site is not recognized as
1249 a built-in function. LLVM will retain the original call and not replace it
1250 with equivalent code based on the semantics of the built-in function, unless
1251 the call site uses the ``builtin`` attribute. This is valid at call sites
1252 and on function declarations and definitions.
1254 This attribute indicates that calls to the function cannot be
1255 duplicated. A call to a ``noduplicate`` function may be moved
1256 within its parent function, but may not be duplicated within
1257 its parent function.
1259 A function containing a ``noduplicate`` call may still
1260 be an inlining candidate, provided that the call is not
1261 duplicated by inlining. That implies that the function has
1262 internal linkage and only has one call site, so the original
1263 call is dead after inlining.
1265 This attributes disables implicit floating point instructions.
1267 This attribute indicates that the inliner should never inline this
1268 function in any situation. This attribute may not be used together
1269 with the ``alwaysinline`` attribute.
1271 This attribute suppresses lazy symbol binding for the function. This
1272 may make calls to the function faster, at the cost of extra program
1273 startup time if the function is not called during program startup.
1275 This attribute indicates that the code generator should not use a
1276 red zone, even if the target-specific ABI normally permits it.
1278 This function attribute indicates that the function never returns
1279 normally. This produces undefined behavior at runtime if the
1280 function ever does dynamically return.
1282 This function attribute indicates that the function never raises an
1283 exception. If the function does raise an exception, its runtime
1284 behavior is undefined. However, functions marked nounwind may still
1285 trap or generate asynchronous exceptions. Exception handling schemes
1286 that are recognized by LLVM to handle asynchronous exceptions, such
1287 as SEH, will still provide their implementation defined semantics.
1289 This function attribute indicates that the function is not optimized
1290 by any optimization or code generator passes with the
1291 exception of interprocedural optimization passes.
1292 This attribute cannot be used together with the ``alwaysinline``
1293 attribute; this attribute is also incompatible
1294 with the ``minsize`` attribute and the ``optsize`` attribute.
1296 This attribute requires the ``noinline`` attribute to be specified on
1297 the function as well, so the function is never inlined into any caller.
1298 Only functions with the ``alwaysinline`` attribute are valid
1299 candidates for inlining into the body of this function.
1301 This attribute suggests that optimization passes and code generator
1302 passes make choices that keep the code size of this function low,
1303 and otherwise do optimizations specifically to reduce code size as
1304 long as they do not significantly impact runtime performance.
1306 On a function, this attribute indicates that the function computes its
1307 result (or decides to unwind an exception) based strictly on its arguments,
1308 without dereferencing any pointer arguments or otherwise accessing
1309 any mutable state (e.g. memory, control registers, etc) visible to
1310 caller functions. It does not write through any pointer arguments
1311 (including ``byval`` arguments) and never changes any state visible
1312 to callers. This means that it cannot unwind exceptions by calling
1313 the ``C++`` exception throwing methods.
1315 On an argument, this attribute indicates that the function does not
1316 dereference that pointer argument, even though it may read or write the
1317 memory that the pointer points to if accessed through other pointers.
1319 On a function, this attribute indicates that the function does not write
1320 through any pointer arguments (including ``byval`` arguments) or otherwise
1321 modify any state (e.g. memory, control registers, etc) visible to
1322 caller functions. It may dereference pointer arguments and read
1323 state that may be set in the caller. A readonly function always
1324 returns the same value (or unwinds an exception identically) when
1325 called with the same set of arguments and global state. It cannot
1326 unwind an exception by calling the ``C++`` exception throwing
1329 On an argument, this attribute indicates that the function does not write
1330 through this pointer argument, even though it may write to the memory that
1331 the pointer points to.
1333 This attribute indicates that the only memory accesses inside function are
1334 loads and stores from objects pointed to by its pointer-typed arguments,
1335 with arbitrary offsets. Or in other words, all memory operations in the
1336 function can refer to memory only using pointers based on its function
1338 Note that ``argmemonly`` can be used together with ``readonly`` attribute
1339 in order to specify that function reads only from its arguments.
1341 This attribute indicates that this function can return twice. The C
1342 ``setjmp`` is an example of such a function. The compiler disables
1343 some optimizations (like tail calls) in the caller of these
1346 This attribute indicates that
1347 `SafeStack <http://clang.llvm.org/docs/SafeStack.html>`_
1348 protection is enabled for this function.
1350 If a function that has a ``safestack`` attribute is inlined into a
1351 function that doesn't have a ``safestack`` attribute or which has an
1352 ``ssp``, ``sspstrong`` or ``sspreq`` attribute, then the resulting
1353 function will have a ``safestack`` attribute.
1354 ``sanitize_address``
1355 This attribute indicates that AddressSanitizer checks
1356 (dynamic address safety analysis) are enabled for this function.
1358 This attribute indicates that MemorySanitizer checks (dynamic detection
1359 of accesses to uninitialized memory) are enabled for this function.
1361 This attribute indicates that ThreadSanitizer checks
1362 (dynamic thread safety analysis) are enabled for this function.
1364 This attribute indicates that the function should emit a stack
1365 smashing protector. It is in the form of a "canary" --- a random value
1366 placed on the stack before the local variables that's checked upon
1367 return from the function to see if it has been overwritten. A
1368 heuristic is used to determine if a function needs stack protectors
1369 or not. The heuristic used will enable protectors for functions with:
1371 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1372 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1373 - Calls to alloca() with variable sizes or constant sizes greater than
1374 ``ssp-buffer-size``.
1376 Variables that are identified as requiring a protector will be arranged
1377 on the stack such that they are adjacent to the stack protector guard.
1379 If a function that has an ``ssp`` attribute is inlined into a
1380 function that doesn't have an ``ssp`` attribute, then the resulting
1381 function will have an ``ssp`` attribute.
1383 This attribute indicates that the function should *always* emit a
1384 stack smashing protector. This overrides the ``ssp`` function
1387 Variables that are identified as requiring a protector will be arranged
1388 on the stack such that they are adjacent to the stack protector guard.
1389 The specific layout rules are:
1391 #. Large arrays and structures containing large arrays
1392 (``>= ssp-buffer-size``) are closest to the stack protector.
1393 #. Small arrays and structures containing small arrays
1394 (``< ssp-buffer-size``) are 2nd closest to the protector.
1395 #. Variables that have had their address taken are 3rd closest to the
1398 If a function that has an ``sspreq`` attribute is inlined into a
1399 function that doesn't have an ``sspreq`` attribute or which has an
1400 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1401 an ``sspreq`` attribute.
1403 This attribute indicates that the function should emit a stack smashing
1404 protector. This attribute causes a strong heuristic to be used when
1405 determining if a function needs stack protectors. The strong heuristic
1406 will enable protectors for functions with:
1408 - Arrays of any size and type
1409 - Aggregates containing an array of any size and type.
1410 - Calls to alloca().
1411 - Local variables that have had their address taken.
1413 Variables that are identified as requiring a protector will be arranged
1414 on the stack such that they are adjacent to the stack protector guard.
1415 The specific layout rules are:
1417 #. Large arrays and structures containing large arrays
1418 (``>= ssp-buffer-size``) are closest to the stack protector.
1419 #. Small arrays and structures containing small arrays
1420 (``< ssp-buffer-size``) are 2nd closest to the protector.
1421 #. Variables that have had their address taken are 3rd closest to the
1424 This overrides the ``ssp`` function attribute.
1426 If a function that has an ``sspstrong`` attribute is inlined into a
1427 function that doesn't have an ``sspstrong`` attribute, then the
1428 resulting function will have an ``sspstrong`` attribute.
1430 This attribute indicates that the function will delegate to some other
1431 function with a tail call. The prototype of a thunk should not be used for
1432 optimization purposes. The caller is expected to cast the thunk prototype to
1433 match the thunk target prototype.
1435 This attribute indicates that the ABI being targeted requires that
1436 an unwind table entry be produce for this function even if we can
1437 show that no exceptions passes by it. This is normally the case for
1438 the ELF x86-64 abi, but it can be disabled for some compilation
1443 Module-Level Inline Assembly
1444 ----------------------------
1446 Modules may contain "module-level inline asm" blocks, which corresponds
1447 to the GCC "file scope inline asm" blocks. These blocks are internally
1448 concatenated by LLVM and treated as a single unit, but may be separated
1449 in the ``.ll`` file if desired. The syntax is very simple:
1451 .. code-block:: llvm
1453 module asm "inline asm code goes here"
1454 module asm "more can go here"
1456 The strings can contain any character by escaping non-printable
1457 characters. The escape sequence used is simply "\\xx" where "xx" is the
1458 two digit hex code for the number.
1460 Note that the assembly string *must* be parseable by LLVM's integrated assembler
1461 (unless it is disabled), even when emitting a ``.s`` file.
1463 .. _langref_datalayout:
1468 A module may specify a target specific data layout string that specifies
1469 how data is to be laid out in memory. The syntax for the data layout is
1472 .. code-block:: llvm
1474 target datalayout = "layout specification"
1476 The *layout specification* consists of a list of specifications
1477 separated by the minus sign character ('-'). Each specification starts
1478 with a letter and may include other information after the letter to
1479 define some aspect of the data layout. The specifications accepted are
1483 Specifies that the target lays out data in big-endian form. That is,
1484 the bits with the most significance have the lowest address
1487 Specifies that the target lays out data in little-endian form. That
1488 is, the bits with the least significance have the lowest address
1491 Specifies the natural alignment of the stack in bits. Alignment
1492 promotion of stack variables is limited to the natural stack
1493 alignment to avoid dynamic stack realignment. The stack alignment
1494 must be a multiple of 8-bits. If omitted, the natural stack
1495 alignment defaults to "unspecified", which does not prevent any
1496 alignment promotions.
1497 ``p[n]:<size>:<abi>:<pref>``
1498 This specifies the *size* of a pointer and its ``<abi>`` and
1499 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1500 bits. The address space, ``n`` is optional, and if not specified,
1501 denotes the default address space 0. The value of ``n`` must be
1502 in the range [1,2^23).
1503 ``i<size>:<abi>:<pref>``
1504 This specifies the alignment for an integer type of a given bit
1505 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1506 ``v<size>:<abi>:<pref>``
1507 This specifies the alignment for a vector type of a given bit
1509 ``f<size>:<abi>:<pref>``
1510 This specifies the alignment for a floating point type of a given bit
1511 ``<size>``. Only values of ``<size>`` that are supported by the target
1512 will work. 32 (float) and 64 (double) are supported on all targets; 80
1513 or 128 (different flavors of long double) are also supported on some
1516 This specifies the alignment for an object of aggregate type.
1518 If present, specifies that llvm names are mangled in the output. The
1521 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1522 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1523 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1524 symbols get a ``_`` prefix.
1525 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1526 functions also get a suffix based on the frame size.
1527 ``n<size1>:<size2>:<size3>...``
1528 This specifies a set of native integer widths for the target CPU in
1529 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1530 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1531 this set are considered to support most general arithmetic operations
1534 On every specification that takes a ``<abi>:<pref>``, specifying the
1535 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1536 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1538 When constructing the data layout for a given target, LLVM starts with a
1539 default set of specifications which are then (possibly) overridden by
1540 the specifications in the ``datalayout`` keyword. The default
1541 specifications are given in this list:
1543 - ``E`` - big endian
1544 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1545 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1546 same as the default address space.
1547 - ``S0`` - natural stack alignment is unspecified
1548 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1549 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1550 - ``i16:16:16`` - i16 is 16-bit aligned
1551 - ``i32:32:32`` - i32 is 32-bit aligned
1552 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1553 alignment of 64-bits
1554 - ``f16:16:16`` - half is 16-bit aligned
1555 - ``f32:32:32`` - float is 32-bit aligned
1556 - ``f64:64:64`` - double is 64-bit aligned
1557 - ``f128:128:128`` - quad is 128-bit aligned
1558 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1559 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1560 - ``a:0:64`` - aggregates are 64-bit aligned
1562 When LLVM is determining the alignment for a given type, it uses the
1565 #. If the type sought is an exact match for one of the specifications,
1566 that specification is used.
1567 #. If no match is found, and the type sought is an integer type, then
1568 the smallest integer type that is larger than the bitwidth of the
1569 sought type is used. If none of the specifications are larger than
1570 the bitwidth then the largest integer type is used. For example,
1571 given the default specifications above, the i7 type will use the
1572 alignment of i8 (next largest) while both i65 and i256 will use the
1573 alignment of i64 (largest specified).
1574 #. If no match is found, and the type sought is a vector type, then the
1575 largest vector type that is smaller than the sought vector type will
1576 be used as a fall back. This happens because <128 x double> can be
1577 implemented in terms of 64 <2 x double>, for example.
1579 The function of the data layout string may not be what you expect.
1580 Notably, this is not a specification from the frontend of what alignment
1581 the code generator should use.
1583 Instead, if specified, the target data layout is required to match what
1584 the ultimate *code generator* expects. This string is used by the
1585 mid-level optimizers to improve code, and this only works if it matches
1586 what the ultimate code generator uses. There is no way to generate IR
1587 that does not embed this target-specific detail into the IR. If you
1588 don't specify the string, the default specifications will be used to
1589 generate a Data Layout and the optimization phases will operate
1590 accordingly and introduce target specificity into the IR with respect to
1591 these default specifications.
1598 A module may specify a target triple string that describes the target
1599 host. The syntax for the target triple is simply:
1601 .. code-block:: llvm
1603 target triple = "x86_64-apple-macosx10.7.0"
1605 The *target triple* string consists of a series of identifiers delimited
1606 by the minus sign character ('-'). The canonical forms are:
1610 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1611 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1613 This information is passed along to the backend so that it generates
1614 code for the proper architecture. It's possible to override this on the
1615 command line with the ``-mtriple`` command line option.
1617 .. _pointeraliasing:
1619 Pointer Aliasing Rules
1620 ----------------------
1622 Any memory access must be done through a pointer value associated with
1623 an address range of the memory access, otherwise the behavior is
1624 undefined. Pointer values are associated with address ranges according
1625 to the following rules:
1627 - A pointer value is associated with the addresses associated with any
1628 value it is *based* on.
1629 - An address of a global variable is associated with the address range
1630 of the variable's storage.
1631 - The result value of an allocation instruction is associated with the
1632 address range of the allocated storage.
1633 - A null pointer in the default address-space is associated with no
1635 - An integer constant other than zero or a pointer value returned from
1636 a function not defined within LLVM may be associated with address
1637 ranges allocated through mechanisms other than those provided by
1638 LLVM. Such ranges shall not overlap with any ranges of addresses
1639 allocated by mechanisms provided by LLVM.
1641 A pointer value is *based* on another pointer value according to the
1644 - A pointer value formed from a ``getelementptr`` operation is *based*
1645 on the first value operand of the ``getelementptr``.
1646 - The result value of a ``bitcast`` is *based* on the operand of the
1648 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1649 values that contribute (directly or indirectly) to the computation of
1650 the pointer's value.
1651 - The "*based* on" relationship is transitive.
1653 Note that this definition of *"based"* is intentionally similar to the
1654 definition of *"based"* in C99, though it is slightly weaker.
1656 LLVM IR does not associate types with memory. The result type of a
1657 ``load`` merely indicates the size and alignment of the memory from
1658 which to load, as well as the interpretation of the value. The first
1659 operand type of a ``store`` similarly only indicates the size and
1660 alignment of the store.
1662 Consequently, type-based alias analysis, aka TBAA, aka
1663 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1664 :ref:`Metadata <metadata>` may be used to encode additional information
1665 which specialized optimization passes may use to implement type-based
1670 Volatile Memory Accesses
1671 ------------------------
1673 Certain memory accesses, such as :ref:`load <i_load>`'s,
1674 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1675 marked ``volatile``. The optimizers must not change the number of
1676 volatile operations or change their order of execution relative to other
1677 volatile operations. The optimizers *may* change the order of volatile
1678 operations relative to non-volatile operations. This is not Java's
1679 "volatile" and has no cross-thread synchronization behavior.
1681 IR-level volatile loads and stores cannot safely be optimized into
1682 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1683 flagged volatile. Likewise, the backend should never split or merge
1684 target-legal volatile load/store instructions.
1686 .. admonition:: Rationale
1688 Platforms may rely on volatile loads and stores of natively supported
1689 data width to be executed as single instruction. For example, in C
1690 this holds for an l-value of volatile primitive type with native
1691 hardware support, but not necessarily for aggregate types. The
1692 frontend upholds these expectations, which are intentionally
1693 unspecified in the IR. The rules above ensure that IR transformation
1694 do not violate the frontend's contract with the language.
1698 Memory Model for Concurrent Operations
1699 --------------------------------------
1701 The LLVM IR does not define any way to start parallel threads of
1702 execution or to register signal handlers. Nonetheless, there are
1703 platform-specific ways to create them, and we define LLVM IR's behavior
1704 in their presence. This model is inspired by the C++0x memory model.
1706 For a more informal introduction to this model, see the :doc:`Atomics`.
1708 We define a *happens-before* partial order as the least partial order
1711 - Is a superset of single-thread program order, and
1712 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1713 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1714 techniques, like pthread locks, thread creation, thread joining,
1715 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1716 Constraints <ordering>`).
1718 Note that program order does not introduce *happens-before* edges
1719 between a thread and signals executing inside that thread.
1721 Every (defined) read operation (load instructions, memcpy, atomic
1722 loads/read-modify-writes, etc.) R reads a series of bytes written by
1723 (defined) write operations (store instructions, atomic
1724 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1725 section, initialized globals are considered to have a write of the
1726 initializer which is atomic and happens before any other read or write
1727 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1728 may see any write to the same byte, except:
1730 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1731 write\ :sub:`2` happens before R\ :sub:`byte`, then
1732 R\ :sub:`byte` does not see write\ :sub:`1`.
1733 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1734 R\ :sub:`byte` does not see write\ :sub:`3`.
1736 Given that definition, R\ :sub:`byte` is defined as follows:
1738 - If R is volatile, the result is target-dependent. (Volatile is
1739 supposed to give guarantees which can support ``sig_atomic_t`` in
1740 C/C++, and may be used for accesses to addresses that do not behave
1741 like normal memory. It does not generally provide cross-thread
1743 - Otherwise, if there is no write to the same byte that happens before
1744 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1745 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1746 R\ :sub:`byte` returns the value written by that write.
1747 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1748 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1749 Memory Ordering Constraints <ordering>` section for additional
1750 constraints on how the choice is made.
1751 - Otherwise R\ :sub:`byte` returns ``undef``.
1753 R returns the value composed of the series of bytes it read. This
1754 implies that some bytes within the value may be ``undef`` **without**
1755 the entire value being ``undef``. Note that this only defines the
1756 semantics of the operation; it doesn't mean that targets will emit more
1757 than one instruction to read the series of bytes.
1759 Note that in cases where none of the atomic intrinsics are used, this
1760 model places only one restriction on IR transformations on top of what
1761 is required for single-threaded execution: introducing a store to a byte
1762 which might not otherwise be stored is not allowed in general.
1763 (Specifically, in the case where another thread might write to and read
1764 from an address, introducing a store can change a load that may see
1765 exactly one write into a load that may see multiple writes.)
1769 Atomic Memory Ordering Constraints
1770 ----------------------------------
1772 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1773 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1774 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1775 ordering parameters that determine which other atomic instructions on
1776 the same address they *synchronize with*. These semantics are borrowed
1777 from Java and C++0x, but are somewhat more colloquial. If these
1778 descriptions aren't precise enough, check those specs (see spec
1779 references in the :doc:`atomics guide <Atomics>`).
1780 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1781 differently since they don't take an address. See that instruction's
1782 documentation for details.
1784 For a simpler introduction to the ordering constraints, see the
1788 The set of values that can be read is governed by the happens-before
1789 partial order. A value cannot be read unless some operation wrote
1790 it. This is intended to provide a guarantee strong enough to model
1791 Java's non-volatile shared variables. This ordering cannot be
1792 specified for read-modify-write operations; it is not strong enough
1793 to make them atomic in any interesting way.
1795 In addition to the guarantees of ``unordered``, there is a single
1796 total order for modifications by ``monotonic`` operations on each
1797 address. All modification orders must be compatible with the
1798 happens-before order. There is no guarantee that the modification
1799 orders can be combined to a global total order for the whole program
1800 (and this often will not be possible). The read in an atomic
1801 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1802 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1803 order immediately before the value it writes. If one atomic read
1804 happens before another atomic read of the same address, the later
1805 read must see the same value or a later value in the address's
1806 modification order. This disallows reordering of ``monotonic`` (or
1807 stronger) operations on the same address. If an address is written
1808 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1809 read that address repeatedly, the other threads must eventually see
1810 the write. This corresponds to the C++0x/C1x
1811 ``memory_order_relaxed``.
1813 In addition to the guarantees of ``monotonic``, a
1814 *synchronizes-with* edge may be formed with a ``release`` operation.
1815 This is intended to model C++'s ``memory_order_acquire``.
1817 In addition to the guarantees of ``monotonic``, if this operation
1818 writes a value which is subsequently read by an ``acquire``
1819 operation, it *synchronizes-with* that operation. (This isn't a
1820 complete description; see the C++0x definition of a release
1821 sequence.) This corresponds to the C++0x/C1x
1822 ``memory_order_release``.
1823 ``acq_rel`` (acquire+release)
1824 Acts as both an ``acquire`` and ``release`` operation on its
1825 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1826 ``seq_cst`` (sequentially consistent)
1827 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1828 operation that only reads, ``release`` for an operation that only
1829 writes), there is a global total order on all
1830 sequentially-consistent operations on all addresses, which is
1831 consistent with the *happens-before* partial order and with the
1832 modification orders of all the affected addresses. Each
1833 sequentially-consistent read sees the last preceding write to the
1834 same address in this global order. This corresponds to the C++0x/C1x
1835 ``memory_order_seq_cst`` and Java volatile.
1839 If an atomic operation is marked ``singlethread``, it only *synchronizes
1840 with* or participates in modification and seq\_cst total orderings with
1841 other operations running in the same thread (for example, in signal
1849 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1850 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1851 :ref:`frem <i_frem>`, :ref:`fcmp <i_fcmp>`) have the following flags that can
1852 be set to enable otherwise unsafe floating point operations
1855 No NaNs - Allow optimizations to assume the arguments and result are not
1856 NaN. Such optimizations are required to retain defined behavior over
1857 NaNs, but the value of the result is undefined.
1860 No Infs - Allow optimizations to assume the arguments and result are not
1861 +/-Inf. Such optimizations are required to retain defined behavior over
1862 +/-Inf, but the value of the result is undefined.
1865 No Signed Zeros - Allow optimizations to treat the sign of a zero
1866 argument or result as insignificant.
1869 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1870 argument rather than perform division.
1873 Fast - Allow algebraically equivalent transformations that may
1874 dramatically change results in floating point (e.g. reassociate). This
1875 flag implies all the others.
1879 Use-list Order Directives
1880 -------------------------
1882 Use-list directives encode the in-memory order of each use-list, allowing the
1883 order to be recreated. ``<order-indexes>`` is a comma-separated list of
1884 indexes that are assigned to the referenced value's uses. The referenced
1885 value's use-list is immediately sorted by these indexes.
1887 Use-list directives may appear at function scope or global scope. They are not
1888 instructions, and have no effect on the semantics of the IR. When they're at
1889 function scope, they must appear after the terminator of the final basic block.
1891 If basic blocks have their address taken via ``blockaddress()`` expressions,
1892 ``uselistorder_bb`` can be used to reorder their use-lists from outside their
1899 uselistorder <ty> <value>, { <order-indexes> }
1900 uselistorder_bb @function, %block { <order-indexes> }
1906 define void @foo(i32 %arg1, i32 %arg2) {
1908 ; ... instructions ...
1910 ; ... instructions ...
1912 ; At function scope.
1913 uselistorder i32 %arg1, { 1, 0, 2 }
1914 uselistorder label %bb, { 1, 0 }
1918 uselistorder i32* @global, { 1, 2, 0 }
1919 uselistorder i32 7, { 1, 0 }
1920 uselistorder i32 (i32) @bar, { 1, 0 }
1921 uselistorder_bb @foo, %bb, { 5, 1, 3, 2, 0, 4 }
1928 The LLVM type system is one of the most important features of the
1929 intermediate representation. Being typed enables a number of
1930 optimizations to be performed on the intermediate representation
1931 directly, without having to do extra analyses on the side before the
1932 transformation. A strong type system makes it easier to read the
1933 generated code and enables novel analyses and transformations that are
1934 not feasible to perform on normal three address code representations.
1944 The void type does not represent any value and has no size.
1962 The function type can be thought of as a function signature. It consists of a
1963 return type and a list of formal parameter types. The return type of a function
1964 type is a void type or first class type --- except for :ref:`label <t_label>`
1965 and :ref:`metadata <t_metadata>` types.
1971 <returntype> (<parameter list>)
1973 ...where '``<parameter list>``' is a comma-separated list of type
1974 specifiers. Optionally, the parameter list may include a type ``...``, which
1975 indicates that the function takes a variable number of arguments. Variable
1976 argument functions can access their arguments with the :ref:`variable argument
1977 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
1978 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
1982 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1983 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1984 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1985 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1986 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1987 | ``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. |
1988 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1989 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1990 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1997 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1998 Values of these types are the only ones which can be produced by
2006 These are the types that are valid in registers from CodeGen's perspective.
2015 The integer type is a very simple type that simply specifies an
2016 arbitrary bit width for the integer type desired. Any bit width from 1
2017 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
2025 The number of bits the integer will occupy is specified by the ``N``
2031 +----------------+------------------------------------------------+
2032 | ``i1`` | a single-bit integer. |
2033 +----------------+------------------------------------------------+
2034 | ``i32`` | a 32-bit integer. |
2035 +----------------+------------------------------------------------+
2036 | ``i1942652`` | a really big integer of over 1 million bits. |
2037 +----------------+------------------------------------------------+
2041 Floating Point Types
2042 """"""""""""""""""""
2051 - 16-bit floating point value
2054 - 32-bit floating point value
2057 - 64-bit floating point value
2060 - 128-bit floating point value (112-bit mantissa)
2063 - 80-bit floating point value (X87)
2066 - 128-bit floating point value (two 64-bits)
2073 The x86_mmx type represents a value held in an MMX register on an x86
2074 machine. The operations allowed on it are quite limited: parameters and
2075 return values, load and store, and bitcast. User-specified MMX
2076 instructions are represented as intrinsic or asm calls with arguments
2077 and/or results of this type. There are no arrays, vectors or constants
2094 The pointer type is used to specify memory locations. Pointers are
2095 commonly used to reference objects in memory.
2097 Pointer types may have an optional address space attribute defining the
2098 numbered address space where the pointed-to object resides. The default
2099 address space is number zero. The semantics of non-zero address spaces
2100 are target-specific.
2102 Note that LLVM does not permit pointers to void (``void*``) nor does it
2103 permit pointers to labels (``label*``). Use ``i8*`` instead.
2113 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2114 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
2115 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2116 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
2117 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2118 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
2119 +-------------------------+--------------------------------------------------------------------------------------------------------------+
2128 A vector type is a simple derived type that represents a vector of
2129 elements. Vector types are used when multiple primitive data are
2130 operated in parallel using a single instruction (SIMD). A vector type
2131 requires a size (number of elements) and an underlying primitive data
2132 type. Vector types are considered :ref:`first class <t_firstclass>`.
2138 < <# elements> x <elementtype> >
2140 The number of elements is a constant integer value larger than 0;
2141 elementtype may be any integer, floating point or pointer type. Vectors
2142 of size zero are not allowed.
2146 +-------------------+--------------------------------------------------+
2147 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
2148 +-------------------+--------------------------------------------------+
2149 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
2150 +-------------------+--------------------------------------------------+
2151 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
2152 +-------------------+--------------------------------------------------+
2153 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
2154 +-------------------+--------------------------------------------------+
2163 The label type represents code labels.
2178 The metadata type represents embedded metadata. No derived types may be
2179 created from metadata except for :ref:`function <t_function>` arguments.
2192 Aggregate Types are a subset of derived types that can contain multiple
2193 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
2194 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
2204 The array type is a very simple derived type that arranges elements
2205 sequentially in memory. The array type requires a size (number of
2206 elements) and an underlying data type.
2212 [<# elements> x <elementtype>]
2214 The number of elements is a constant integer value; ``elementtype`` may
2215 be any type with a size.
2219 +------------------+--------------------------------------+
2220 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
2221 +------------------+--------------------------------------+
2222 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
2223 +------------------+--------------------------------------+
2224 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
2225 +------------------+--------------------------------------+
2227 Here are some examples of multidimensional arrays:
2229 +-----------------------------+----------------------------------------------------------+
2230 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
2231 +-----------------------------+----------------------------------------------------------+
2232 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
2233 +-----------------------------+----------------------------------------------------------+
2234 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
2235 +-----------------------------+----------------------------------------------------------+
2237 There is no restriction on indexing beyond the end of the array implied
2238 by a static type (though there are restrictions on indexing beyond the
2239 bounds of an allocated object in some cases). This means that
2240 single-dimension 'variable sized array' addressing can be implemented in
2241 LLVM with a zero length array type. An implementation of 'pascal style
2242 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
2252 The structure type is used to represent a collection of data members
2253 together in memory. The elements of a structure may be any type that has
2256 Structures in memory are accessed using '``load``' and '``store``' by
2257 getting a pointer to a field with the '``getelementptr``' instruction.
2258 Structures in registers are accessed using the '``extractvalue``' and
2259 '``insertvalue``' instructions.
2261 Structures may optionally be "packed" structures, which indicate that
2262 the alignment of the struct is one byte, and that there is no padding
2263 between the elements. In non-packed structs, padding between field types
2264 is inserted as defined by the DataLayout string in the module, which is
2265 required to match what the underlying code generator expects.
2267 Structures can either be "literal" or "identified". A literal structure
2268 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
2269 identified types are always defined at the top level with a name.
2270 Literal types are uniqued by their contents and can never be recursive
2271 or opaque since there is no way to write one. Identified types can be
2272 recursive, can be opaqued, and are never uniqued.
2278 %T1 = type { <type list> } ; Identified normal struct type
2279 %T2 = type <{ <type list> }> ; Identified packed struct type
2283 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2284 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
2285 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2286 | ``{ 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``. |
2287 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2288 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
2289 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2293 Opaque Structure Types
2294 """"""""""""""""""""""
2298 Opaque structure types are used to represent named structure types that
2299 do not have a body specified. This corresponds (for example) to the C
2300 notion of a forward declared structure.
2311 +--------------+-------------------+
2312 | ``opaque`` | An opaque type. |
2313 +--------------+-------------------+
2320 LLVM has several different basic types of constants. This section
2321 describes them all and their syntax.
2326 **Boolean constants**
2327 The two strings '``true``' and '``false``' are both valid constants
2329 **Integer constants**
2330 Standard integers (such as '4') are constants of the
2331 :ref:`integer <t_integer>` type. Negative numbers may be used with
2333 **Floating point constants**
2334 Floating point constants use standard decimal notation (e.g.
2335 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2336 hexadecimal notation (see below). The assembler requires the exact
2337 decimal value of a floating-point constant. For example, the
2338 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2339 decimal in binary. Floating point constants must have a :ref:`floating
2340 point <t_floating>` type.
2341 **Null pointer constants**
2342 The identifier '``null``' is recognized as a null pointer constant
2343 and must be of :ref:`pointer type <t_pointer>`.
2345 The one non-intuitive notation for constants is the hexadecimal form of
2346 floating point constants. For example, the form
2347 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2348 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2349 constants are required (and the only time that they are generated by the
2350 disassembler) is when a floating point constant must be emitted but it
2351 cannot be represented as a decimal floating point number in a reasonable
2352 number of digits. For example, NaN's, infinities, and other special
2353 values are represented in their IEEE hexadecimal format so that assembly
2354 and disassembly do not cause any bits to change in the constants.
2356 When using the hexadecimal form, constants of types half, float, and
2357 double are represented using the 16-digit form shown above (which
2358 matches the IEEE754 representation for double); half and float values
2359 must, however, be exactly representable as IEEE 754 half and single
2360 precision, respectively. Hexadecimal format is always used for long
2361 double, and there are three forms of long double. The 80-bit format used
2362 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2363 128-bit format used by PowerPC (two adjacent doubles) is represented by
2364 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2365 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2366 will only work if they match the long double format on your target.
2367 The IEEE 16-bit format (half precision) is represented by ``0xH``
2368 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2369 (sign bit at the left).
2371 There are no constants of type x86_mmx.
2373 .. _complexconstants:
2378 Complex constants are a (potentially recursive) combination of simple
2379 constants and smaller complex constants.
2381 **Structure constants**
2382 Structure constants are represented with notation similar to
2383 structure type definitions (a comma separated list of elements,
2384 surrounded by braces (``{}``)). For example:
2385 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2386 "``@G = external global i32``". Structure constants must have
2387 :ref:`structure type <t_struct>`, and the number and types of elements
2388 must match those specified by the type.
2390 Array constants are represented with notation similar to array type
2391 definitions (a comma separated list of elements, surrounded by
2392 square brackets (``[]``)). For example:
2393 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2394 :ref:`array type <t_array>`, and the number and types of elements must
2395 match those specified by the type. As a special case, character array
2396 constants may also be represented as a double-quoted string using the ``c``
2397 prefix. For example: "``c"Hello World\0A\00"``".
2398 **Vector constants**
2399 Vector constants are represented with notation similar to vector
2400 type definitions (a comma separated list of elements, surrounded by
2401 less-than/greater-than's (``<>``)). For example:
2402 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2403 must have :ref:`vector type <t_vector>`, and the number and types of
2404 elements must match those specified by the type.
2405 **Zero initialization**
2406 The string '``zeroinitializer``' can be used to zero initialize a
2407 value to zero of *any* type, including scalar and
2408 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2409 having to print large zero initializers (e.g. for large arrays) and
2410 is always exactly equivalent to using explicit zero initializers.
2412 A metadata node is a constant tuple without types. For example:
2413 "``!{!0, !{!2, !0}, !"test"}``". Metadata can reference constant values,
2414 for example: "``!{!0, i32 0, i8* @global, i64 (i64)* @function, !"str"}``".
2415 Unlike other typed constants that are meant to be interpreted as part of
2416 the instruction stream, metadata is a place to attach additional
2417 information such as debug info.
2419 Global Variable and Function Addresses
2420 --------------------------------------
2422 The addresses of :ref:`global variables <globalvars>` and
2423 :ref:`functions <functionstructure>` are always implicitly valid
2424 (link-time) constants. These constants are explicitly referenced when
2425 the :ref:`identifier for the global <identifiers>` is used and always have
2426 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2429 .. code-block:: llvm
2433 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2440 The string '``undef``' can be used anywhere a constant is expected, and
2441 indicates that the user of the value may receive an unspecified
2442 bit-pattern. Undefined values may be of any type (other than '``label``'
2443 or '``void``') and be used anywhere a constant is permitted.
2445 Undefined values are useful because they indicate to the compiler that
2446 the program is well defined no matter what value is used. This gives the
2447 compiler more freedom to optimize. Here are some examples of
2448 (potentially surprising) transformations that are valid (in pseudo IR):
2450 .. code-block:: llvm
2460 This is safe because all of the output bits are affected by the undef
2461 bits. Any output bit can have a zero or one depending on the input bits.
2463 .. code-block:: llvm
2474 These logical operations have bits that are not always affected by the
2475 input. For example, if ``%X`` has a zero bit, then the output of the
2476 '``and``' operation will always be a zero for that bit, no matter what
2477 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2478 optimize or assume that the result of the '``and``' is '``undef``'.
2479 However, it is safe to assume that all bits of the '``undef``' could be
2480 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2481 all the bits of the '``undef``' operand to the '``or``' could be set,
2482 allowing the '``or``' to be folded to -1.
2484 .. code-block:: llvm
2486 %A = select undef, %X, %Y
2487 %B = select undef, 42, %Y
2488 %C = select %X, %Y, undef
2498 This set of examples shows that undefined '``select``' (and conditional
2499 branch) conditions can go *either way*, but they have to come from one
2500 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2501 both known to have a clear low bit, then ``%A`` would have to have a
2502 cleared low bit. However, in the ``%C`` example, the optimizer is
2503 allowed to assume that the '``undef``' operand could be the same as
2504 ``%Y``, allowing the whole '``select``' to be eliminated.
2506 .. code-block:: llvm
2508 %A = xor undef, undef
2525 This example points out that two '``undef``' operands are not
2526 necessarily the same. This can be surprising to people (and also matches
2527 C semantics) where they assume that "``X^X``" is always zero, even if
2528 ``X`` is undefined. This isn't true for a number of reasons, but the
2529 short answer is that an '``undef``' "variable" can arbitrarily change
2530 its value over its "live range". This is true because the variable
2531 doesn't actually *have a live range*. Instead, the value is logically
2532 read from arbitrary registers that happen to be around when needed, so
2533 the value is not necessarily consistent over time. In fact, ``%A`` and
2534 ``%C`` need to have the same semantics or the core LLVM "replace all
2535 uses with" concept would not hold.
2537 .. code-block:: llvm
2545 These examples show the crucial difference between an *undefined value*
2546 and *undefined behavior*. An undefined value (like '``undef``') is
2547 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2548 operation can be constant folded to '``undef``', because the '``undef``'
2549 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2550 However, in the second example, we can make a more aggressive
2551 assumption: because the ``undef`` is allowed to be an arbitrary value,
2552 we are allowed to assume that it could be zero. Since a divide by zero
2553 has *undefined behavior*, we are allowed to assume that the operation
2554 does not execute at all. This allows us to delete the divide and all
2555 code after it. Because the undefined operation "can't happen", the
2556 optimizer can assume that it occurs in dead code.
2558 .. code-block:: llvm
2560 a: store undef -> %X
2561 b: store %X -> undef
2566 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2567 value can be assumed to not have any effect; we can assume that the
2568 value is overwritten with bits that happen to match what was already
2569 there. However, a store *to* an undefined location could clobber
2570 arbitrary memory, therefore, it has undefined behavior.
2577 Poison values are similar to :ref:`undef values <undefvalues>`, however
2578 they also represent the fact that an instruction or constant expression
2579 that cannot evoke side effects has nevertheless detected a condition
2580 that results in undefined behavior.
2582 There is currently no way of representing a poison value in the IR; they
2583 only exist when produced by operations such as :ref:`add <i_add>` with
2586 Poison value behavior is defined in terms of value *dependence*:
2588 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2589 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2590 their dynamic predecessor basic block.
2591 - Function arguments depend on the corresponding actual argument values
2592 in the dynamic callers of their functions.
2593 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2594 instructions that dynamically transfer control back to them.
2595 - :ref:`Invoke <i_invoke>` instructions depend on the
2596 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2597 call instructions that dynamically transfer control back to them.
2598 - Non-volatile loads and stores depend on the most recent stores to all
2599 of the referenced memory addresses, following the order in the IR
2600 (including loads and stores implied by intrinsics such as
2601 :ref:`@llvm.memcpy <int_memcpy>`.)
2602 - An instruction with externally visible side effects depends on the
2603 most recent preceding instruction with externally visible side
2604 effects, following the order in the IR. (This includes :ref:`volatile
2605 operations <volatile>`.)
2606 - An instruction *control-depends* on a :ref:`terminator
2607 instruction <terminators>` if the terminator instruction has
2608 multiple successors and the instruction is always executed when
2609 control transfers to one of the successors, and may not be executed
2610 when control is transferred to another.
2611 - Additionally, an instruction also *control-depends* on a terminator
2612 instruction if the set of instructions it otherwise depends on would
2613 be different if the terminator had transferred control to a different
2615 - Dependence is transitive.
2617 Poison values have the same behavior as :ref:`undef values <undefvalues>`,
2618 with the additional effect that any instruction that has a *dependence*
2619 on a poison value has undefined behavior.
2621 Here are some examples:
2623 .. code-block:: llvm
2626 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2627 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2628 %poison_yet_again = getelementptr i32, i32* @h, i32 %still_poison
2629 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2631 store i32 %poison, i32* @g ; Poison value stored to memory.
2632 %poison2 = load i32, i32* @g ; Poison value loaded back from memory.
2634 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2636 %narrowaddr = bitcast i32* @g to i16*
2637 %wideaddr = bitcast i32* @g to i64*
2638 %poison3 = load i16, i16* %narrowaddr ; Returns a poison value.
2639 %poison4 = load i64, i64* %wideaddr ; Returns a poison value.
2641 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2642 br i1 %cmp, label %true, label %end ; Branch to either destination.
2645 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2646 ; it has undefined behavior.
2650 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2651 ; Both edges into this PHI are
2652 ; control-dependent on %cmp, so this
2653 ; always results in a poison value.
2655 store volatile i32 0, i32* @g ; This would depend on the store in %true
2656 ; if %cmp is true, or the store in %entry
2657 ; otherwise, so this is undefined behavior.
2659 br i1 %cmp, label %second_true, label %second_end
2660 ; The same branch again, but this time the
2661 ; true block doesn't have side effects.
2668 store volatile i32 0, i32* @g ; This time, the instruction always depends
2669 ; on the store in %end. Also, it is
2670 ; control-equivalent to %end, so this is
2671 ; well-defined (ignoring earlier undefined
2672 ; behavior in this example).
2676 Addresses of Basic Blocks
2677 -------------------------
2679 ``blockaddress(@function, %block)``
2681 The '``blockaddress``' constant computes the address of the specified
2682 basic block in the specified function, and always has an ``i8*`` type.
2683 Taking the address of the entry block is illegal.
2685 This value only has defined behavior when used as an operand to the
2686 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2687 against null. Pointer equality tests between labels addresses results in
2688 undefined behavior --- though, again, comparison against null is ok, and
2689 no label is equal to the null pointer. This may be passed around as an
2690 opaque pointer sized value as long as the bits are not inspected. This
2691 allows ``ptrtoint`` and arithmetic to be performed on these values so
2692 long as the original value is reconstituted before the ``indirectbr``
2695 Finally, some targets may provide defined semantics when using the value
2696 as the operand to an inline assembly, but that is target specific.
2700 Constant Expressions
2701 --------------------
2703 Constant expressions are used to allow expressions involving other
2704 constants to be used as constants. Constant expressions may be of any
2705 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2706 that does not have side effects (e.g. load and call are not supported).
2707 The following is the syntax for constant expressions:
2709 ``trunc (CST to TYPE)``
2710 Truncate a constant to another type. The bit size of CST must be
2711 larger than the bit size of TYPE. Both types must be integers.
2712 ``zext (CST to TYPE)``
2713 Zero extend a constant to another type. The bit size of CST must be
2714 smaller than the bit size of TYPE. Both types must be integers.
2715 ``sext (CST to TYPE)``
2716 Sign extend a constant to another type. The bit size of CST must be
2717 smaller than the bit size of TYPE. Both types must be integers.
2718 ``fptrunc (CST to TYPE)``
2719 Truncate a floating point constant to another floating point type.
2720 The size of CST must be larger than the size of TYPE. Both types
2721 must be floating point.
2722 ``fpext (CST to TYPE)``
2723 Floating point extend a constant to another type. The size of CST
2724 must be smaller or equal to the size of TYPE. Both types must be
2726 ``fptoui (CST to TYPE)``
2727 Convert a floating point constant to the corresponding unsigned
2728 integer constant. TYPE must be a scalar or vector integer type. CST
2729 must be of scalar or vector floating point type. Both CST and TYPE
2730 must be scalars, or vectors of the same number of elements. If the
2731 value won't fit in the integer type, the results are undefined.
2732 ``fptosi (CST to TYPE)``
2733 Convert a floating point constant to the corresponding signed
2734 integer constant. TYPE must be a scalar or vector integer type. CST
2735 must be of scalar or vector floating point type. Both CST and TYPE
2736 must be scalars, or vectors of the same number of elements. If the
2737 value won't fit in the integer type, the results are undefined.
2738 ``uitofp (CST to TYPE)``
2739 Convert an unsigned integer constant to the corresponding floating
2740 point constant. TYPE must be a scalar or vector floating point type.
2741 CST must be of scalar or vector integer type. Both CST and TYPE must
2742 be scalars, or vectors of the same number of elements. If the value
2743 won't fit in the floating point type, the results are undefined.
2744 ``sitofp (CST to TYPE)``
2745 Convert a signed integer constant to the corresponding floating
2746 point constant. TYPE must be a scalar or vector floating point type.
2747 CST must be of scalar or vector integer type. Both CST and TYPE must
2748 be scalars, or vectors of the same number of elements. If the value
2749 won't fit in the floating point type, the results are undefined.
2750 ``ptrtoint (CST to TYPE)``
2751 Convert a pointer typed constant to the corresponding integer
2752 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2753 pointer type. The ``CST`` value is zero extended, truncated, or
2754 unchanged to make it fit in ``TYPE``.
2755 ``inttoptr (CST to TYPE)``
2756 Convert an integer constant to a pointer constant. TYPE must be a
2757 pointer type. CST must be of integer type. The CST value is zero
2758 extended, truncated, or unchanged to make it fit in a pointer size.
2759 This one is *really* dangerous!
2760 ``bitcast (CST to TYPE)``
2761 Convert a constant, CST, to another TYPE. The constraints of the
2762 operands are the same as those for the :ref:`bitcast
2763 instruction <i_bitcast>`.
2764 ``addrspacecast (CST to TYPE)``
2765 Convert a constant pointer or constant vector of pointer, CST, to another
2766 TYPE in a different address space. The constraints of the operands are the
2767 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2768 ``getelementptr (TY, CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (TY, CSTPTR, IDX0, IDX1, ...)``
2769 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2770 constants. As with the :ref:`getelementptr <i_getelementptr>`
2771 instruction, the index list may have zero or more indexes, which are
2772 required to make sense for the type of "pointer to TY".
2773 ``select (COND, VAL1, VAL2)``
2774 Perform the :ref:`select operation <i_select>` on constants.
2775 ``icmp COND (VAL1, VAL2)``
2776 Performs the :ref:`icmp operation <i_icmp>` on constants.
2777 ``fcmp COND (VAL1, VAL2)``
2778 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2779 ``extractelement (VAL, IDX)``
2780 Perform the :ref:`extractelement operation <i_extractelement>` on
2782 ``insertelement (VAL, ELT, IDX)``
2783 Perform the :ref:`insertelement operation <i_insertelement>` on
2785 ``shufflevector (VEC1, VEC2, IDXMASK)``
2786 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2788 ``extractvalue (VAL, IDX0, IDX1, ...)``
2789 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2790 constants. The index list is interpreted in a similar manner as
2791 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2792 least one index value must be specified.
2793 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2794 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2795 The index list is interpreted in a similar manner as indices in a
2796 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2797 value must be specified.
2798 ``OPCODE (LHS, RHS)``
2799 Perform the specified operation of the LHS and RHS constants. OPCODE
2800 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2801 binary <bitwiseops>` operations. The constraints on operands are
2802 the same as those for the corresponding instruction (e.g. no bitwise
2803 operations on floating point values are allowed).
2810 Inline Assembler Expressions
2811 ----------------------------
2813 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2814 Inline Assembly <moduleasm>`) through the use of a special value. This value
2815 represents the inline assembler as a template string (containing the
2816 instructions to emit), a list of operand constraints (stored as a string), a
2817 flag that indicates whether or not the inline asm expression has side effects,
2818 and a flag indicating whether the function containing the asm needs to align its
2819 stack conservatively.
2821 The template string supports argument substitution of the operands using "``$``"
2822 followed by a number, to indicate substitution of the given register/memory
2823 location, as specified by the constraint string. "``${NUM:MODIFIER}``" may also
2824 be used, where ``MODIFIER`` is a target-specific annotation for how to print the
2825 operand (See :ref:`inline-asm-modifiers`).
2827 A literal "``$``" may be included by using "``$$``" in the template. To include
2828 other special characters into the output, the usual "``\XX``" escapes may be
2829 used, just as in other strings. Note that after template substitution, the
2830 resulting assembly string is parsed by LLVM's integrated assembler unless it is
2831 disabled -- even when emitting a ``.s`` file -- and thus must contain assembly
2832 syntax known to LLVM.
2834 LLVM's support for inline asm is modeled closely on the requirements of Clang's
2835 GCC-compatible inline-asm support. Thus, the feature-set and the constraint and
2836 modifier codes listed here are similar or identical to those in GCC's inline asm
2837 support. However, to be clear, the syntax of the template and constraint strings
2838 described here is *not* the same as the syntax accepted by GCC and Clang, and,
2839 while most constraint letters are passed through as-is by Clang, some get
2840 translated to other codes when converting from the C source to the LLVM
2843 An example inline assembler expression is:
2845 .. code-block:: llvm
2847 i32 (i32) asm "bswap $0", "=r,r"
2849 Inline assembler expressions may **only** be used as the callee operand
2850 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2851 Thus, typically we have:
2853 .. code-block:: llvm
2855 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2857 Inline asms with side effects not visible in the constraint list must be
2858 marked as having side effects. This is done through the use of the
2859 '``sideeffect``' keyword, like so:
2861 .. code-block:: llvm
2863 call void asm sideeffect "eieio", ""()
2865 In some cases inline asms will contain code that will not work unless
2866 the stack is aligned in some way, such as calls or SSE instructions on
2867 x86, yet will not contain code that does that alignment within the asm.
2868 The compiler should make conservative assumptions about what the asm
2869 might contain and should generate its usual stack alignment code in the
2870 prologue if the '``alignstack``' keyword is present:
2872 .. code-block:: llvm
2874 call void asm alignstack "eieio", ""()
2876 Inline asms also support using non-standard assembly dialects. The
2877 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2878 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2879 the only supported dialects. An example is:
2881 .. code-block:: llvm
2883 call void asm inteldialect "eieio", ""()
2885 If multiple keywords appear the '``sideeffect``' keyword must come
2886 first, the '``alignstack``' keyword second and the '``inteldialect``'
2889 Inline Asm Constraint String
2890 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2892 The constraint list is a comma-separated string, each element containing one or
2893 more constraint codes.
2895 For each element in the constraint list an appropriate register or memory
2896 operand will be chosen, and it will be made available to assembly template
2897 string expansion as ``$0`` for the first constraint in the list, ``$1`` for the
2900 There are three different types of constraints, which are distinguished by a
2901 prefix symbol in front of the constraint code: Output, Input, and Clobber. The
2902 constraints must always be given in that order: outputs first, then inputs, then
2903 clobbers. They cannot be intermingled.
2905 There are also three different categories of constraint codes:
2907 - Register constraint. This is either a register class, or a fixed physical
2908 register. This kind of constraint will allocate a register, and if necessary,
2909 bitcast the argument or result to the appropriate type.
2910 - Memory constraint. This kind of constraint is for use with an instruction
2911 taking a memory operand. Different constraints allow for different addressing
2912 modes used by the target.
2913 - Immediate value constraint. This kind of constraint is for an integer or other
2914 immediate value which can be rendered directly into an instruction. The
2915 various target-specific constraints allow the selection of a value in the
2916 proper range for the instruction you wish to use it with.
2921 Output constraints are specified by an "``=``" prefix (e.g. "``=r``"). This
2922 indicates that the assembly will write to this operand, and the operand will
2923 then be made available as a return value of the ``asm`` expression. Output
2924 constraints do not consume an argument from the call instruction. (Except, see
2925 below about indirect outputs).
2927 Normally, it is expected that no output locations are written to by the assembly
2928 expression until *all* of the inputs have been read. As such, LLVM may assign
2929 the same register to an output and an input. If this is not safe (e.g. if the
2930 assembly contains two instructions, where the first writes to one output, and
2931 the second reads an input and writes to a second output), then the "``&``"
2932 modifier must be used (e.g. "``=&r``") to specify that the output is an
2933 "early-clobber" output. Marking an ouput as "early-clobber" ensures that LLVM
2934 will not use the same register for any inputs (other than an input tied to this
2940 Input constraints do not have a prefix -- just the constraint codes. Each input
2941 constraint will consume one argument from the call instruction. It is not
2942 permitted for the asm to write to any input register or memory location (unless
2943 that input is tied to an output). Note also that multiple inputs may all be
2944 assigned to the same register, if LLVM can determine that they necessarily all
2945 contain the same value.
2947 Instead of providing a Constraint Code, input constraints may also "tie"
2948 themselves to an output constraint, by providing an integer as the constraint
2949 string. Tied inputs still consume an argument from the call instruction, and
2950 take up a position in the asm template numbering as is usual -- they will simply
2951 be constrained to always use the same register as the output they've been tied
2952 to. For example, a constraint string of "``=r,0``" says to assign a register for
2953 output, and use that register as an input as well (it being the 0'th
2956 It is permitted to tie an input to an "early-clobber" output. In that case, no
2957 *other* input may share the same register as the input tied to the early-clobber
2958 (even when the other input has the same value).
2960 You may only tie an input to an output which has a register constraint, not a
2961 memory constraint. Only a single input may be tied to an output.
2963 There is also an "interesting" feature which deserves a bit of explanation: if a
2964 register class constraint allocates a register which is too small for the value
2965 type operand provided as input, the input value will be split into multiple
2966 registers, and all of them passed to the inline asm.
2968 However, this feature is often not as useful as you might think.
2970 Firstly, the registers are *not* guaranteed to be consecutive. So, on those
2971 architectures that have instructions which operate on multiple consecutive
2972 instructions, this is not an appropriate way to support them. (e.g. the 32-bit
2973 SparcV8 has a 64-bit load, which instruction takes a single 32-bit register. The
2974 hardware then loads into both the named register, and the next register. This
2975 feature of inline asm would not be useful to support that.)
2977 A few of the targets provide a template string modifier allowing explicit access
2978 to the second register of a two-register operand (e.g. MIPS ``L``, ``M``, and
2979 ``D``). On such an architecture, you can actually access the second allocated
2980 register (yet, still, not any subsequent ones). But, in that case, you're still
2981 probably better off simply splitting the value into two separate operands, for
2982 clarity. (e.g. see the description of the ``A`` constraint on X86, which,
2983 despite existing only for use with this feature, is not really a good idea to
2986 Indirect inputs and outputs
2987 """""""""""""""""""""""""""
2989 Indirect output or input constraints can be specified by the "``*``" modifier
2990 (which goes after the "``=``" in case of an output). This indicates that the asm
2991 will write to or read from the contents of an *address* provided as an input
2992 argument. (Note that in this way, indirect outputs act more like an *input* than
2993 an output: just like an input, they consume an argument of the call expression,
2994 rather than producing a return value. An indirect output constraint is an
2995 "output" only in that the asm is expected to write to the contents of the input
2996 memory location, instead of just read from it).
2998 This is most typically used for memory constraint, e.g. "``=*m``", to pass the
2999 address of a variable as a value.
3001 It is also possible to use an indirect *register* constraint, but only on output
3002 (e.g. "``=*r``"). This will cause LLVM to allocate a register for an output
3003 value normally, and then, separately emit a store to the address provided as
3004 input, after the provided inline asm. (It's not clear what value this
3005 functionality provides, compared to writing the store explicitly after the asm
3006 statement, and it can only produce worse code, since it bypasses many
3007 optimization passes. I would recommend not using it.)
3013 A clobber constraint is indicated by a "``~``" prefix. A clobber does not
3014 consume an input operand, nor generate an output. Clobbers cannot use any of the
3015 general constraint code letters -- they may use only explicit register
3016 constraints, e.g. "``~{eax}``". The one exception is that a clobber string of
3017 "``~{memory}``" indicates that the assembly writes to arbitrary undeclared
3018 memory locations -- not only the memory pointed to by a declared indirect
3024 After a potential prefix comes constraint code, or codes.
3026 A Constraint Code is either a single letter (e.g. "``r``"), a "``^``" character
3027 followed by two letters (e.g. "``^wc``"), or "``{``" register-name "``}``"
3030 The one and two letter constraint codes are typically chosen to be the same as
3031 GCC's constraint codes.
3033 A single constraint may include one or more than constraint code in it, leaving
3034 it up to LLVM to choose which one to use. This is included mainly for
3035 compatibility with the translation of GCC inline asm coming from clang.
3037 There are two ways to specify alternatives, and either or both may be used in an
3038 inline asm constraint list:
3040 1) Append the codes to each other, making a constraint code set. E.g. "``im``"
3041 or "``{eax}m``". This means "choose any of the options in the set". The
3042 choice of constraint is made independently for each constraint in the
3045 2) Use "``|``" between constraint code sets, creating alternatives. Every
3046 constraint in the constraint list must have the same number of alternative
3047 sets. With this syntax, the same alternative in *all* of the items in the
3048 constraint list will be chosen together.
3050 Putting those together, you might have a two operand constraint string like
3051 ``"rm|r,ri|rm"``. This indicates that if operand 0 is ``r`` or ``m``, then
3052 operand 1 may be one of ``r`` or ``i``. If operand 0 is ``r``, then operand 1
3053 may be one of ``r`` or ``m``. But, operand 0 and 1 cannot both be of type m.
3055 However, the use of either of the alternatives features is *NOT* recommended, as
3056 LLVM is not able to make an intelligent choice about which one to use. (At the
3057 point it currently needs to choose, not enough information is available to do so
3058 in a smart way.) Thus, it simply tries to make a choice that's most likely to
3059 compile, not one that will be optimal performance. (e.g., given "``rm``", it'll
3060 always choose to use memory, not registers). And, if given multiple registers,
3061 or multiple register classes, it will simply choose the first one. (In fact, it
3062 doesn't currently even ensure explicitly specified physical registers are
3063 unique, so specifying multiple physical registers as alternatives, like
3064 ``{r11}{r12},{r11}{r12}``, will assign r11 to both operands, not at all what was
3067 Supported Constraint Code List
3068 """"""""""""""""""""""""""""""
3070 The constraint codes are, in general, expected to behave the same way they do in
3071 GCC. LLVM's support is often implemented on an 'as-needed' basis, to support C
3072 inline asm code which was supported by GCC. A mismatch in behavior between LLVM
3073 and GCC likely indicates a bug in LLVM.
3075 Some constraint codes are typically supported by all targets:
3077 - ``r``: A register in the target's general purpose register class.
3078 - ``m``: A memory address operand. It is target-specific what addressing modes
3079 are supported, typical examples are register, or register + register offset,
3080 or register + immediate offset (of some target-specific size).
3081 - ``i``: An integer constant (of target-specific width). Allows either a simple
3082 immediate, or a relocatable value.
3083 - ``n``: An integer constant -- *not* including relocatable values.
3084 - ``s``: An integer constant, but allowing *only* relocatable values.
3085 - ``X``: Allows an operand of any kind, no constraint whatsoever. Typically
3086 useful to pass a label for an asm branch or call.
3088 .. FIXME: but that surely isn't actually okay to jump out of an asm
3089 block without telling llvm about the control transfer???)
3091 - ``{register-name}``: Requires exactly the named physical register.
3093 Other constraints are target-specific:
3097 - ``z``: An immediate integer 0. Outputs ``WZR`` or ``XZR``, as appropriate.
3098 - ``I``: An immediate integer valid for an ``ADD`` or ``SUB`` instruction,
3099 i.e. 0 to 4095 with optional shift by 12.
3100 - ``J``: An immediate integer that, when negated, is valid for an ``ADD`` or
3101 ``SUB`` instruction, i.e. -1 to -4095 with optional left shift by 12.
3102 - ``K``: An immediate integer that is valid for the 'bitmask immediate 32' of a
3103 logical instruction like ``AND``, ``EOR``, or ``ORR`` with a 32-bit register.
3104 - ``L``: An immediate integer that is valid for the 'bitmask immediate 64' of a
3105 logical instruction like ``AND``, ``EOR``, or ``ORR`` with a 64-bit register.
3106 - ``M``: An immediate integer for use with the ``MOV`` assembly alias on a
3107 32-bit register. This is a superset of ``K``: in addition to the bitmask
3108 immediate, also allows immediate integers which can be loaded with a single
3109 ``MOVZ`` or ``MOVL`` instruction.
3110 - ``N``: An immediate integer for use with the ``MOV`` assembly alias on a
3111 64-bit register. This is a superset of ``L``.
3112 - ``Q``: Memory address operand must be in a single register (no
3113 offsets). (However, LLVM currently does this for the ``m`` constraint as
3115 - ``r``: A 32 or 64-bit integer register (W* or X*).
3116 - ``w``: A 32, 64, or 128-bit floating-point/SIMD register.
3117 - ``x``: A lower 128-bit floating-point/SIMD register (``V0`` to ``V15``).
3121 - ``r``: A 32 or 64-bit integer register.
3122 - ``[0-9]v``: The 32-bit VGPR register, number 0-9.
3123 - ``[0-9]s``: The 32-bit SGPR register, number 0-9.
3128 - ``Q``, ``Um``, ``Un``, ``Uq``, ``Us``, ``Ut``, ``Uv``, ``Uy``: Memory address
3129 operand. Treated the same as operand ``m``, at the moment.
3131 ARM and ARM's Thumb2 mode:
3133 - ``j``: An immediate integer between 0 and 65535 (valid for ``MOVW``)
3134 - ``I``: An immediate integer valid for a data-processing instruction.
3135 - ``J``: An immediate integer between -4095 and 4095.
3136 - ``K``: An immediate integer whose bitwise inverse is valid for a
3137 data-processing instruction. (Can be used with template modifier "``B``" to
3138 print the inverted value).
3139 - ``L``: An immediate integer whose negation is valid for a data-processing
3140 instruction. (Can be used with template modifier "``n``" to print the negated
3142 - ``M``: A power of two or a integer between 0 and 32.
3143 - ``N``: Invalid immediate constraint.
3144 - ``O``: Invalid immediate constraint.
3145 - ``r``: A general-purpose 32-bit integer register (``r0-r15``).
3146 - ``l``: In Thumb2 mode, low 32-bit GPR registers (``r0-r7``). In ARM mode, same
3148 - ``h``: In Thumb2 mode, a high 32-bit GPR register (``r8-r15``). In ARM mode,
3150 - ``w``: A 32, 64, or 128-bit floating-point/SIMD register: ``s0-s31``,
3151 ``d0-d31``, or ``q0-q15``.
3152 - ``x``: A 32, 64, or 128-bit floating-point/SIMD register: ``s0-s15``,
3153 ``d0-d7``, or ``q0-q3``.
3154 - ``t``: A floating-point/SIMD register, only supports 32-bit values:
3159 - ``I``: An immediate integer between 0 and 255.
3160 - ``J``: An immediate integer between -255 and -1.
3161 - ``K``: An immediate integer between 0 and 255, with optional left-shift by
3163 - ``L``: An immediate integer between -7 and 7.
3164 - ``M``: An immediate integer which is a multiple of 4 between 0 and 1020.
3165 - ``N``: An immediate integer between 0 and 31.
3166 - ``O``: An immediate integer which is a multiple of 4 between -508 and 508.
3167 - ``r``: A low 32-bit GPR register (``r0-r7``).
3168 - ``l``: A low 32-bit GPR register (``r0-r7``).
3169 - ``h``: A high GPR register (``r0-r7``).
3170 - ``w``: A 32, 64, or 128-bit floating-point/SIMD register: ``s0-s31``,
3171 ``d0-d31``, or ``q0-q15``.
3172 - ``x``: A 32, 64, or 128-bit floating-point/SIMD register: ``s0-s15``,
3173 ``d0-d7``, or ``q0-q3``.
3174 - ``t``: A floating-point/SIMD register, only supports 32-bit values:
3180 - ``o``, ``v``: A memory address operand, treated the same as constraint ``m``,
3182 - ``r``: A 32 or 64-bit register.
3186 - ``r``: An 8 or 16-bit register.
3190 - ``I``: An immediate signed 16-bit integer.
3191 - ``J``: An immediate integer zero.
3192 - ``K``: An immediate unsigned 16-bit integer.
3193 - ``L``: An immediate 32-bit integer, where the lower 16 bits are 0.
3194 - ``N``: An immediate integer between -65535 and -1.
3195 - ``O``: An immediate signed 15-bit integer.
3196 - ``P``: An immediate integer between 1 and 65535.
3197 - ``m``: A memory address operand. In MIPS-SE mode, allows a base address
3198 register plus 16-bit immediate offset. In MIPS mode, just a base register.
3199 - ``R``: A memory address operand. In MIPS-SE mode, allows a base address
3200 register plus a 9-bit signed offset. In MIPS mode, the same as constraint
3202 - ``ZC``: A memory address operand, suitable for use in a ``pref``, ``ll``, or
3203 ``sc`` instruction on the given subtarget (details vary).
3204 - ``r``, ``d``, ``y``: A 32 or 64-bit GPR register.
3205 - ``f``: A 32 or 64-bit FPU register (``F0-F31``), or a 128-bit MSA register
3206 (``W0-W31``). In the case of MSA registers, it is recommended to use the ``w``
3207 argument modifier for compatibility with GCC.
3208 - ``c``: A 32-bit or 64-bit GPR register suitable for indirect jump (always
3210 - ``l``: The ``lo`` register, 32 or 64-bit.
3215 - ``b``: A 1-bit integer register.
3216 - ``c`` or ``h``: A 16-bit integer register.
3217 - ``r``: A 32-bit integer register.
3218 - ``l`` or ``N``: A 64-bit integer register.
3219 - ``f``: A 32-bit float register.
3220 - ``d``: A 64-bit float register.
3225 - ``I``: An immediate signed 16-bit integer.
3226 - ``J``: An immediate unsigned 16-bit integer, shifted left 16 bits.
3227 - ``K``: An immediate unsigned 16-bit integer.
3228 - ``L``: An immediate signed 16-bit integer, shifted left 16 bits.
3229 - ``M``: An immediate integer greater than 31.
3230 - ``N``: An immediate integer that is an exact power of 2.
3231 - ``O``: The immediate integer constant 0.
3232 - ``P``: An immediate integer constant whose negation is a signed 16-bit
3234 - ``es``, ``o``, ``Q``, ``Z``, ``Zy``: A memory address operand, currently
3235 treated the same as ``m``.
3236 - ``r``: A 32 or 64-bit integer register.
3237 - ``b``: A 32 or 64-bit integer register, excluding ``R0`` (that is:
3239 - ``f``: A 32 or 64-bit float register (``F0-F31``), or when QPX is enabled, a
3240 128 or 256-bit QPX register (``Q0-Q31``; aliases the ``F`` registers).
3241 - ``v``: For ``4 x f32`` or ``4 x f64`` types, when QPX is enabled, a
3242 128 or 256-bit QPX register (``Q0-Q31``), otherwise a 128-bit
3243 altivec vector register (``V0-V31``).
3245 .. FIXME: is this a bug that v accepts QPX registers? I think this
3246 is supposed to only use the altivec vector registers?
3248 - ``y``: Condition register (``CR0-CR7``).
3249 - ``wc``: An individual CR bit in a CR register.
3250 - ``wa``, ``wd``, ``wf``: Any 128-bit VSX vector register, from the full VSX
3251 register set (overlapping both the floating-point and vector register files).
3252 - ``ws``: A 32 or 64-bit floating point register, from the full VSX register
3257 - ``I``: An immediate 13-bit signed integer.
3258 - ``r``: A 32-bit integer register.
3262 - ``I``: An immediate unsigned 8-bit integer.
3263 - ``J``: An immediate unsigned 12-bit integer.
3264 - ``K``: An immediate signed 16-bit integer.
3265 - ``L``: An immediate signed 20-bit integer.
3266 - ``M``: An immediate integer 0x7fffffff.
3267 - ``Q``, ``R``, ``S``, ``T``: A memory address operand, treated the same as
3268 ``m``, at the moment.
3269 - ``r`` or ``d``: A 32, 64, or 128-bit integer register.
3270 - ``a``: A 32, 64, or 128-bit integer address register (excludes R0, which in an
3271 address context evaluates as zero).
3272 - ``h``: A 32-bit value in the high part of a 64bit data register
3274 - ``f``: A 32, 64, or 128-bit floating point register.
3278 - ``I``: An immediate integer between 0 and 31.
3279 - ``J``: An immediate integer between 0 and 64.
3280 - ``K``: An immediate signed 8-bit integer.
3281 - ``L``: An immediate integer, 0xff or 0xffff or (in 64-bit mode only)
3283 - ``M``: An immediate integer between 0 and 3.
3284 - ``N``: An immediate unsigned 8-bit integer.
3285 - ``O``: An immediate integer between 0 and 127.
3286 - ``e``: An immediate 32-bit signed integer.
3287 - ``Z``: An immediate 32-bit unsigned integer.
3288 - ``o``, ``v``: Treated the same as ``m``, at the moment.
3289 - ``q``: An 8, 16, 32, or 64-bit register which can be accessed as an 8-bit
3290 ``l`` integer register. On X86-32, this is the ``a``, ``b``, ``c``, and ``d``
3291 registers, and on X86-64, it is all of the integer registers.
3292 - ``Q``: An 8, 16, 32, or 64-bit register which can be accessed as an 8-bit
3293 ``h`` integer register. This is the ``a``, ``b``, ``c``, and ``d`` registers.
3294 - ``r`` or ``l``: An 8, 16, 32, or 64-bit integer register.
3295 - ``R``: An 8, 16, 32, or 64-bit "legacy" integer register -- one which has
3296 existed since i386, and can be accessed without the REX prefix.
3297 - ``f``: A 32, 64, or 80-bit '387 FPU stack pseudo-register.
3298 - ``y``: A 64-bit MMX register, if MMX is enabled.
3299 - ``x``: If SSE is enabled: a 32 or 64-bit scalar operand, or 128-bit vector
3300 operand in a SSE register. If AVX is also enabled, can also be a 256-bit
3301 vector operand in an AVX register. If AVX-512 is also enabled, can also be a
3302 512-bit vector operand in an AVX512 register, Otherwise, an error.
3303 - ``Y``: The same as ``x``, if *SSE2* is enabled, otherwise an error.
3304 - ``A``: Special case: allocates EAX first, then EDX, for a single operand (in
3305 32-bit mode, a 64-bit integer operand will get split into two registers). It
3306 is not recommended to use this constraint, as in 64-bit mode, the 64-bit
3307 operand will get allocated only to RAX -- if two 32-bit operands are needed,
3308 you're better off splitting it yourself, before passing it to the asm
3313 - ``r``: A 32-bit integer register.
3316 .. _inline-asm-modifiers:
3318 Asm template argument modifiers
3319 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3321 In the asm template string, modifiers can be used on the operand reference, like
3324 The modifiers are, in general, expected to behave the same way they do in
3325 GCC. LLVM's support is often implemented on an 'as-needed' basis, to support C
3326 inline asm code which was supported by GCC. A mismatch in behavior between LLVM
3327 and GCC likely indicates a bug in LLVM.
3331 - ``c``: Print an immediate integer constant unadorned, without
3332 the target-specific immediate punctuation (e.g. no ``$`` prefix).
3333 - ``n``: Negate and print immediate integer constant unadorned, without the
3334 target-specific immediate punctuation (e.g. no ``$`` prefix).
3335 - ``l``: Print as an unadorned label, without the target-specific label
3336 punctuation (e.g. no ``$`` prefix).
3340 - ``w``: Print a GPR register with a ``w*`` name instead of ``x*`` name. E.g.,
3341 instead of ``x30``, print ``w30``.
3342 - ``x``: Print a GPR register with a ``x*`` name. (this is the default, anyhow).
3343 - ``b``, ``h``, ``s``, ``d``, ``q``: Print a floating-point/SIMD register with a
3344 ``b*``, ``h*``, ``s*``, ``d*``, or ``q*`` name, rather than the default of
3353 - ``a``: Print an operand as an address (with ``[`` and ``]`` surrounding a
3357 - ``y``: Print a VFP single-precision register as an indexed double (e.g. print
3358 as ``d4[1]`` instead of ``s9``)
3359 - ``B``: Bitwise invert and print an immediate integer constant without ``#``
3361 - ``L``: Print the low 16-bits of an immediate integer constant.
3362 - ``M``: Print as a register set suitable for ldm/stm. Also prints *all*
3363 register operands subsequent to the specified one (!), so use carefully.
3364 - ``Q``: Print the low-order register of a register-pair, or the low-order
3365 register of a two-register operand.
3366 - ``R``: Print the high-order register of a register-pair, or the high-order
3367 register of a two-register operand.
3368 - ``H``: Print the second register of a register-pair. (On a big-endian system,
3369 ``H`` is equivalent to ``Q``, and on little-endian system, ``H`` is equivalent
3372 .. FIXME: H doesn't currently support printing the second register
3373 of a two-register operand.
3375 - ``e``: Print the low doubleword register of a NEON quad register.
3376 - ``f``: Print the high doubleword register of a NEON quad register.
3377 - ``m``: Print the base register of a memory operand without the ``[`` and ``]``
3382 - ``L``: Print the second register of a two-register operand. Requires that it
3383 has been allocated consecutively to the first.
3385 .. FIXME: why is it restricted to consecutive ones? And there's
3386 nothing that ensures that happens, is there?
3388 - ``I``: Print the letter 'i' if the operand is an integer constant, otherwise
3389 nothing. Used to print 'addi' vs 'add' instructions.
3393 No additional modifiers.
3397 - ``X``: Print an immediate integer as hexadecimal
3398 - ``x``: Print the low 16 bits of an immediate integer as hexadecimal.
3399 - ``d``: Print an immediate integer as decimal.
3400 - ``m``: Subtract one and print an immediate integer as decimal.
3401 - ``z``: Print $0 if an immediate zero, otherwise print normally.
3402 - ``L``: Print the low-order register of a two-register operand, or prints the
3403 address of the low-order word of a double-word memory operand.
3405 .. FIXME: L seems to be missing memory operand support.
3407 - ``M``: Print the high-order register of a two-register operand, or prints the
3408 address of the high-order word of a double-word memory operand.
3410 .. FIXME: M seems to be missing memory operand support.
3412 - ``D``: Print the second register of a two-register operand, or prints the
3413 second word of a double-word memory operand. (On a big-endian system, ``D`` is
3414 equivalent to ``L``, and on little-endian system, ``D`` is equivalent to
3416 - ``w``: No effect. Provided for compatibility with GCC which requires this
3417 modifier in order to print MSA registers (``W0-W31``) with the ``f``
3426 - ``L``: Print the second register of a two-register operand. Requires that it
3427 has been allocated consecutively to the first.
3429 .. FIXME: why is it restricted to consecutive ones? And there's
3430 nothing that ensures that happens, is there?
3432 - ``I``: Print the letter 'i' if the operand is an integer constant, otherwise
3433 nothing. Used to print 'addi' vs 'add' instructions.
3434 - ``y``: For a memory operand, prints formatter for a two-register X-form
3435 instruction. (Currently always prints ``r0,OPERAND``).
3436 - ``U``: Prints 'u' if the memory operand is an update form, and nothing
3437 otherwise. (NOTE: LLVM does not support update form, so this will currently
3438 always print nothing)
3439 - ``X``: Prints 'x' if the memory operand is an indexed form. (NOTE: LLVM does
3440 not support indexed form, so this will currently always print nothing)
3448 SystemZ implements only ``n``, and does *not* support any of the other
3449 target-independent modifiers.
3453 - ``c``: Print an unadorned integer or symbol name. (The latter is
3454 target-specific behavior for this typically target-independent modifier).
3455 - ``A``: Print a register name with a '``*``' before it.
3456 - ``b``: Print an 8-bit register name (e.g. ``al``); do nothing on a memory
3458 - ``h``: Print the upper 8-bit register name (e.g. ``ah``); do nothing on a
3460 - ``w``: Print the 16-bit register name (e.g. ``ax``); do nothing on a memory
3462 - ``k``: Print the 32-bit register name (e.g. ``eax``); do nothing on a memory
3464 - ``q``: Print the 64-bit register name (e.g. ``rax``), if 64-bit registers are
3465 available, otherwise the 32-bit register name; do nothing on a memory operand.
3466 - ``n``: Negate and print an unadorned integer, or, for operands other than an
3467 immediate integer (e.g. a relocatable symbol expression), print a '-' before
3468 the operand. (The behavior for relocatable symbol expressions is a
3469 target-specific behavior for this typically target-independent modifier)
3470 - ``H``: Print a memory reference with additional offset +8.
3471 - ``P``: Print a memory reference or operand for use as the argument of a call
3472 instruction. (E.g. omit ``(rip)``, even though it's PC-relative.)
3476 No additional modifiers.
3482 The call instructions that wrap inline asm nodes may have a
3483 "``!srcloc``" MDNode attached to it that contains a list of constant
3484 integers. If present, the code generator will use the integer as the
3485 location cookie value when report errors through the ``LLVMContext``
3486 error reporting mechanisms. This allows a front-end to correlate backend
3487 errors that occur with inline asm back to the source code that produced
3490 .. code-block:: llvm
3492 call void asm sideeffect "something bad", ""(), !srcloc !42
3494 !42 = !{ i32 1234567 }
3496 It is up to the front-end to make sense of the magic numbers it places
3497 in the IR. If the MDNode contains multiple constants, the code generator
3498 will use the one that corresponds to the line of the asm that the error
3506 LLVM IR allows metadata to be attached to instructions in the program
3507 that can convey extra information about the code to the optimizers and
3508 code generator. One example application of metadata is source-level
3509 debug information. There are two metadata primitives: strings and nodes.
3511 Metadata does not have a type, and is not a value. If referenced from a
3512 ``call`` instruction, it uses the ``metadata`` type.
3514 All metadata are identified in syntax by a exclamation point ('``!``').
3516 .. _metadata-string:
3518 Metadata Nodes and Metadata Strings
3519 -----------------------------------
3521 A metadata string is a string surrounded by double quotes. It can
3522 contain any character by escaping non-printable characters with
3523 "``\xx``" where "``xx``" is the two digit hex code. For example:
3526 Metadata nodes are represented with notation similar to structure
3527 constants (a comma separated list of elements, surrounded by braces and
3528 preceded by an exclamation point). Metadata nodes can have any values as
3529 their operand. For example:
3531 .. code-block:: llvm
3533 !{ !"test\00", i32 10}
3535 Metadata nodes that aren't uniqued use the ``distinct`` keyword. For example:
3537 .. code-block:: llvm
3539 !0 = distinct !{!"test\00", i32 10}
3541 ``distinct`` nodes are useful when nodes shouldn't be merged based on their
3542 content. They can also occur when transformations cause uniquing collisions
3543 when metadata operands change.
3545 A :ref:`named metadata <namedmetadatastructure>` is a collection of
3546 metadata nodes, which can be looked up in the module symbol table. For
3549 .. code-block:: llvm
3553 Metadata can be used as function arguments. Here ``llvm.dbg.value``
3554 function is using two metadata arguments:
3556 .. code-block:: llvm
3558 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
3560 Metadata can be attached with an instruction. Here metadata ``!21`` is
3561 attached to the ``add`` instruction using the ``!dbg`` identifier:
3563 .. code-block:: llvm
3565 %indvar.next = add i64 %indvar, 1, !dbg !21
3567 More information about specific metadata nodes recognized by the
3568 optimizers and code generator is found below.
3570 .. _specialized-metadata:
3572 Specialized Metadata Nodes
3573 ^^^^^^^^^^^^^^^^^^^^^^^^^^
3575 Specialized metadata nodes are custom data structures in metadata (as opposed
3576 to generic tuples). Their fields are labelled, and can be specified in any
3579 These aren't inherently debug info centric, but currently all the specialized
3580 metadata nodes are related to debug info.
3587 ``DICompileUnit`` nodes represent a compile unit. The ``enums:``,
3588 ``retainedTypes:``, ``subprograms:``, ``globals:`` and ``imports:`` fields are
3589 tuples containing the debug info to be emitted along with the compile unit,
3590 regardless of code optimizations (some nodes are only emitted if there are
3591 references to them from instructions).
3593 .. code-block:: llvm
3595 !0 = !DICompileUnit(language: DW_LANG_C99, file: !1, producer: "clang",
3596 isOptimized: true, flags: "-O2", runtimeVersion: 2,
3597 splitDebugFilename: "abc.debug", emissionKind: 1,
3598 enums: !2, retainedTypes: !3, subprograms: !4,
3599 globals: !5, imports: !6)
3601 Compile unit descriptors provide the root scope for objects declared in a
3602 specific compilation unit. File descriptors are defined using this scope.
3603 These descriptors are collected by a named metadata ``!llvm.dbg.cu``. They
3604 keep track of subprograms, global variables, type information, and imported
3605 entities (declarations and namespaces).
3612 ``DIFile`` nodes represent files. The ``filename:`` can include slashes.
3614 .. code-block:: llvm
3616 !0 = !DIFile(filename: "path/to/file", directory: "/path/to/dir")
3618 Files are sometimes used in ``scope:`` fields, and are the only valid target
3619 for ``file:`` fields.
3626 ``DIBasicType`` nodes represent primitive types, such as ``int``, ``bool`` and
3627 ``float``. ``tag:`` defaults to ``DW_TAG_base_type``.
3629 .. code-block:: llvm
3631 !0 = !DIBasicType(name: "unsigned char", size: 8, align: 8,
3632 encoding: DW_ATE_unsigned_char)
3633 !1 = !DIBasicType(tag: DW_TAG_unspecified_type, name: "decltype(nullptr)")
3635 The ``encoding:`` describes the details of the type. Usually it's one of the
3638 .. code-block:: llvm
3644 DW_ATE_signed_char = 6
3646 DW_ATE_unsigned_char = 8
3648 .. _DISubroutineType:
3653 ``DISubroutineType`` nodes represent subroutine types. Their ``types:`` field
3654 refers to a tuple; the first operand is the return type, while the rest are the
3655 types of the formal arguments in order. If the first operand is ``null``, that
3656 represents a function with no return value (such as ``void foo() {}`` in C++).
3658 .. code-block:: llvm
3660 !0 = !BasicType(name: "int", size: 32, align: 32, DW_ATE_signed)
3661 !1 = !BasicType(name: "char", size: 8, align: 8, DW_ATE_signed_char)
3662 !2 = !DISubroutineType(types: !{null, !0, !1}) ; void (int, char)
3669 ``DIDerivedType`` nodes represent types derived from other types, such as
3672 .. code-block:: llvm
3674 !0 = !DIBasicType(name: "unsigned char", size: 8, align: 8,
3675 encoding: DW_ATE_unsigned_char)
3676 !1 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !0, size: 32,
3679 The following ``tag:`` values are valid:
3681 .. code-block:: llvm
3683 DW_TAG_formal_parameter = 5
3685 DW_TAG_pointer_type = 15
3686 DW_TAG_reference_type = 16
3688 DW_TAG_ptr_to_member_type = 31
3689 DW_TAG_const_type = 38
3690 DW_TAG_volatile_type = 53
3691 DW_TAG_restrict_type = 55
3693 ``DW_TAG_member`` is used to define a member of a :ref:`composite type
3694 <DICompositeType>` or :ref:`subprogram <DISubprogram>`. The type of the member
3695 is the ``baseType:``. The ``offset:`` is the member's bit offset.
3696 ``DW_TAG_formal_parameter`` is used to define a member which is a formal
3697 argument of a subprogram.
3699 ``DW_TAG_typedef`` is used to provide a name for the ``baseType:``.
3701 ``DW_TAG_pointer_type``, ``DW_TAG_reference_type``, ``DW_TAG_const_type``,
3702 ``DW_TAG_volatile_type`` and ``DW_TAG_restrict_type`` are used to qualify the
3705 Note that the ``void *`` type is expressed as a type derived from NULL.
3707 .. _DICompositeType:
3712 ``DICompositeType`` nodes represent types composed of other types, like
3713 structures and unions. ``elements:`` points to a tuple of the composed types.
3715 If the source language supports ODR, the ``identifier:`` field gives the unique
3716 identifier used for type merging between modules. When specified, other types
3717 can refer to composite types indirectly via a :ref:`metadata string
3718 <metadata-string>` that matches their identifier.
3720 .. code-block:: llvm
3722 !0 = !DIEnumerator(name: "SixKind", value: 7)
3723 !1 = !DIEnumerator(name: "SevenKind", value: 7)
3724 !2 = !DIEnumerator(name: "NegEightKind", value: -8)
3725 !3 = !DICompositeType(tag: DW_TAG_enumeration_type, name: "Enum", file: !12,
3726 line: 2, size: 32, align: 32, identifier: "_M4Enum",
3727 elements: !{!0, !1, !2})
3729 The following ``tag:`` values are valid:
3731 .. code-block:: llvm
3733 DW_TAG_array_type = 1
3734 DW_TAG_class_type = 2
3735 DW_TAG_enumeration_type = 4
3736 DW_TAG_structure_type = 19
3737 DW_TAG_union_type = 23
3738 DW_TAG_subroutine_type = 21
3739 DW_TAG_inheritance = 28
3742 For ``DW_TAG_array_type``, the ``elements:`` should be :ref:`subrange
3743 descriptors <DISubrange>`, each representing the range of subscripts at that
3744 level of indexing. The ``DIFlagVector`` flag to ``flags:`` indicates that an
3745 array type is a native packed vector.
3747 For ``DW_TAG_enumeration_type``, the ``elements:`` should be :ref:`enumerator
3748 descriptors <DIEnumerator>`, each representing the definition of an enumeration
3749 value for the set. All enumeration type descriptors are collected in the
3750 ``enums:`` field of the :ref:`compile unit <DICompileUnit>`.
3752 For ``DW_TAG_structure_type``, ``DW_TAG_class_type``, and
3753 ``DW_TAG_union_type``, the ``elements:`` should be :ref:`derived types
3754 <DIDerivedType>` with ``tag: DW_TAG_member`` or ``tag: DW_TAG_inheritance``.
3761 ``DISubrange`` nodes are the elements for ``DW_TAG_array_type`` variants of
3762 :ref:`DICompositeType`. ``count: -1`` indicates an empty array.
3764 .. code-block:: llvm
3766 !0 = !DISubrange(count: 5, lowerBound: 0) ; array counting from 0
3767 !1 = !DISubrange(count: 5, lowerBound: 1) ; array counting from 1
3768 !2 = !DISubrange(count: -1) ; empty array.
3775 ``DIEnumerator`` nodes are the elements for ``DW_TAG_enumeration_type``
3776 variants of :ref:`DICompositeType`.
3778 .. code-block:: llvm
3780 !0 = !DIEnumerator(name: "SixKind", value: 7)
3781 !1 = !DIEnumerator(name: "SevenKind", value: 7)
3782 !2 = !DIEnumerator(name: "NegEightKind", value: -8)
3784 DITemplateTypeParameter
3785 """""""""""""""""""""""
3787 ``DITemplateTypeParameter`` nodes represent type parameters to generic source
3788 language constructs. They are used (optionally) in :ref:`DICompositeType` and
3789 :ref:`DISubprogram` ``templateParams:`` fields.
3791 .. code-block:: llvm
3793 !0 = !DITemplateTypeParameter(name: "Ty", type: !1)
3795 DITemplateValueParameter
3796 """"""""""""""""""""""""
3798 ``DITemplateValueParameter`` nodes represent value parameters to generic source
3799 language constructs. ``tag:`` defaults to ``DW_TAG_template_value_parameter``,
3800 but if specified can also be set to ``DW_TAG_GNU_template_template_param`` or
3801 ``DW_TAG_GNU_template_param_pack``. They are used (optionally) in
3802 :ref:`DICompositeType` and :ref:`DISubprogram` ``templateParams:`` fields.
3804 .. code-block:: llvm
3806 !0 = !DITemplateValueParameter(name: "Ty", type: !1, value: i32 7)
3811 ``DINamespace`` nodes represent namespaces in the source language.
3813 .. code-block:: llvm
3815 !0 = !DINamespace(name: "myawesomeproject", scope: !1, file: !2, line: 7)
3820 ``DIGlobalVariable`` nodes represent global variables in the source language.
3822 .. code-block:: llvm
3824 !0 = !DIGlobalVariable(name: "foo", linkageName: "foo", scope: !1,
3825 file: !2, line: 7, type: !3, isLocal: true,
3826 isDefinition: false, variable: i32* @foo,
3829 All global variables should be referenced by the `globals:` field of a
3830 :ref:`compile unit <DICompileUnit>`.
3837 ``DISubprogram`` nodes represent functions from the source language. The
3838 ``variables:`` field points at :ref:`variables <DILocalVariable>` that must be
3839 retained, even if their IR counterparts are optimized out of the IR. The
3840 ``type:`` field must point at an :ref:`DISubroutineType`.
3842 .. code-block:: llvm
3844 !0 = !DISubprogram(name: "foo", linkageName: "_Zfoov", scope: !1,
3845 file: !2, line: 7, type: !3, isLocal: true,
3846 isDefinition: false, scopeLine: 8, containingType: !4,
3847 virtuality: DW_VIRTUALITY_pure_virtual, virtualIndex: 10,
3848 flags: DIFlagPrototyped, isOptimized: true,
3849 function: void ()* @_Z3foov,
3850 templateParams: !5, declaration: !6, variables: !7)
3857 ``DILexicalBlock`` nodes describe nested blocks within a :ref:`subprogram
3858 <DISubprogram>`. The line number and column numbers are used to dinstinguish
3859 two lexical blocks at same depth. They are valid targets for ``scope:``
3862 .. code-block:: llvm
3864 !0 = distinct !DILexicalBlock(scope: !1, file: !2, line: 7, column: 35)
3866 Usually lexical blocks are ``distinct`` to prevent node merging based on
3869 .. _DILexicalBlockFile:
3874 ``DILexicalBlockFile`` nodes are used to discriminate between sections of a
3875 :ref:`lexical block <DILexicalBlock>`. The ``file:`` field can be changed to
3876 indicate textual inclusion, or the ``discriminator:`` field can be used to
3877 discriminate between control flow within a single block in the source language.
3879 .. code-block:: llvm
3881 !0 = !DILexicalBlock(scope: !3, file: !4, line: 7, column: 35)
3882 !1 = !DILexicalBlockFile(scope: !0, file: !4, discriminator: 0)
3883 !2 = !DILexicalBlockFile(scope: !0, file: !4, discriminator: 1)
3890 ``DILocation`` nodes represent source debug locations. The ``scope:`` field is
3891 mandatory, and points at an :ref:`DILexicalBlockFile`, an
3892 :ref:`DILexicalBlock`, or an :ref:`DISubprogram`.
3894 .. code-block:: llvm
3896 !0 = !DILocation(line: 2900, column: 42, scope: !1, inlinedAt: !2)
3898 .. _DILocalVariable:
3903 ``DILocalVariable`` nodes represent local variables in the source language. If
3904 the ``arg:`` field is set to non-zero, then this variable is a subprogram
3905 parameter, and it will be included in the ``variables:`` field of its
3906 :ref:`DISubprogram`.
3908 .. code-block:: llvm
3910 !0 = !DILocalVariable(name: "this", arg: 1, scope: !3, file: !2, line: 7,
3911 type: !3, flags: DIFlagArtificial)
3912 !1 = !DILocalVariable(name: "x", arg: 2, scope: !4, file: !2, line: 7,
3914 !2 = !DILocalVariable(name: "y", scope: !5, file: !2, line: 7, type: !3)
3919 ``DIExpression`` nodes represent DWARF expression sequences. They are used in
3920 :ref:`debug intrinsics<dbg_intrinsics>` (such as ``llvm.dbg.declare``) to
3921 describe how the referenced LLVM variable relates to the source language
3924 The current supported vocabulary is limited:
3926 - ``DW_OP_deref`` dereferences the working expression.
3927 - ``DW_OP_plus, 93`` adds ``93`` to the working expression.
3928 - ``DW_OP_bit_piece, 16, 8`` specifies the offset and size (``16`` and ``8``
3929 here, respectively) of the variable piece from the working expression.
3931 .. code-block:: llvm
3933 !0 = !DIExpression(DW_OP_deref)
3934 !1 = !DIExpression(DW_OP_plus, 3)
3935 !2 = !DIExpression(DW_OP_bit_piece, 3, 7)
3936 !3 = !DIExpression(DW_OP_deref, DW_OP_plus, 3, DW_OP_bit_piece, 3, 7)
3941 ``DIObjCProperty`` nodes represent Objective-C property nodes.
3943 .. code-block:: llvm
3945 !3 = !DIObjCProperty(name: "foo", file: !1, line: 7, setter: "setFoo",
3946 getter: "getFoo", attributes: 7, type: !2)
3951 ``DIImportedEntity`` nodes represent entities (such as modules) imported into a
3954 .. code-block:: llvm
3956 !2 = !DIImportedEntity(tag: DW_TAG_imported_module, name: "foo", scope: !0,
3957 entity: !1, line: 7)
3962 In LLVM IR, memory does not have types, so LLVM's own type system is not
3963 suitable for doing TBAA. Instead, metadata is added to the IR to
3964 describe a type system of a higher level language. This can be used to
3965 implement typical C/C++ TBAA, but it can also be used to implement
3966 custom alias analysis behavior for other languages.
3968 The current metadata format is very simple. TBAA metadata nodes have up
3969 to three fields, e.g.:
3971 .. code-block:: llvm
3973 !0 = !{ !"an example type tree" }
3974 !1 = !{ !"int", !0 }
3975 !2 = !{ !"float", !0 }
3976 !3 = !{ !"const float", !2, i64 1 }
3978 The first field is an identity field. It can be any value, usually a
3979 metadata string, which uniquely identifies the type. The most important
3980 name in the tree is the name of the root node. Two trees with different
3981 root node names are entirely disjoint, even if they have leaves with
3984 The second field identifies the type's parent node in the tree, or is
3985 null or omitted for a root node. A type is considered to alias all of
3986 its descendants and all of its ancestors in the tree. Also, a type is
3987 considered to alias all types in other trees, so that bitcode produced
3988 from multiple front-ends is handled conservatively.
3990 If the third field is present, it's an integer which if equal to 1
3991 indicates that the type is "constant" (meaning
3992 ``pointsToConstantMemory`` should return true; see `other useful
3993 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
3995 '``tbaa.struct``' Metadata
3996 ^^^^^^^^^^^^^^^^^^^^^^^^^^
3998 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
3999 aggregate assignment operations in C and similar languages, however it
4000 is defined to copy a contiguous region of memory, which is more than
4001 strictly necessary for aggregate types which contain holes due to
4002 padding. Also, it doesn't contain any TBAA information about the fields
4005 ``!tbaa.struct`` metadata can describe which memory subregions in a
4006 memcpy are padding and what the TBAA tags of the struct are.
4008 The current metadata format is very simple. ``!tbaa.struct`` metadata
4009 nodes are a list of operands which are in conceptual groups of three.
4010 For each group of three, the first operand gives the byte offset of a
4011 field in bytes, the second gives its size in bytes, and the third gives
4014 .. code-block:: llvm
4016 !4 = !{ i64 0, i64 4, !1, i64 8, i64 4, !2 }
4018 This describes a struct with two fields. The first is at offset 0 bytes
4019 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
4020 and has size 4 bytes and has tbaa tag !2.
4022 Note that the fields need not be contiguous. In this example, there is a
4023 4 byte gap between the two fields. This gap represents padding which
4024 does not carry useful data and need not be preserved.
4026 '``noalias``' and '``alias.scope``' Metadata
4027 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4029 ``noalias`` and ``alias.scope`` metadata provide the ability to specify generic
4030 noalias memory-access sets. This means that some collection of memory access
4031 instructions (loads, stores, memory-accessing calls, etc.) that carry
4032 ``noalias`` metadata can specifically be specified not to alias with some other
4033 collection of memory access instructions that carry ``alias.scope`` metadata.
4034 Each type of metadata specifies a list of scopes where each scope has an id and
4035 a domain. When evaluating an aliasing query, if for some domain, the set
4036 of scopes with that domain in one instruction's ``alias.scope`` list is a
4037 subset of (or equal to) the set of scopes for that domain in another
4038 instruction's ``noalias`` list, then the two memory accesses are assumed not to
4041 The metadata identifying each domain is itself a list containing one or two
4042 entries. The first entry is the name of the domain. Note that if the name is a
4043 string then it can be combined accross functions and translation units. A
4044 self-reference can be used to create globally unique domain names. A
4045 descriptive string may optionally be provided as a second list entry.
4047 The metadata identifying each scope is also itself a list containing two or
4048 three entries. The first entry is the name of the scope. Note that if the name
4049 is a string then it can be combined accross functions and translation units. A
4050 self-reference can be used to create globally unique scope names. A metadata
4051 reference to the scope's domain is the second entry. A descriptive string may
4052 optionally be provided as a third list entry.
4056 .. code-block:: llvm
4058 ; Two scope domains:
4062 ; Some scopes in these domains:
4068 !5 = !{!4} ; A list containing only scope !4
4072 ; These two instructions don't alias:
4073 %0 = load float, float* %c, align 4, !alias.scope !5
4074 store float %0, float* %arrayidx.i, align 4, !noalias !5
4076 ; These two instructions also don't alias (for domain !1, the set of scopes
4077 ; in the !alias.scope equals that in the !noalias list):
4078 %2 = load float, float* %c, align 4, !alias.scope !5
4079 store float %2, float* %arrayidx.i2, align 4, !noalias !6
4081 ; These two instructions may alias (for domain !0, the set of scopes in
4082 ; the !noalias list is not a superset of, or equal to, the scopes in the
4083 ; !alias.scope list):
4084 %2 = load float, float* %c, align 4, !alias.scope !6
4085 store float %0, float* %arrayidx.i, align 4, !noalias !7
4087 '``fpmath``' Metadata
4088 ^^^^^^^^^^^^^^^^^^^^^
4090 ``fpmath`` metadata may be attached to any instruction of floating point
4091 type. It can be used to express the maximum acceptable error in the
4092 result of that instruction, in ULPs, thus potentially allowing the
4093 compiler to use a more efficient but less accurate method of computing
4094 it. ULP is defined as follows:
4096 If ``x`` is a real number that lies between two finite consecutive
4097 floating-point numbers ``a`` and ``b``, without being equal to one
4098 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
4099 distance between the two non-equal finite floating-point numbers
4100 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
4102 The metadata node shall consist of a single positive floating point
4103 number representing the maximum relative error, for example:
4105 .. code-block:: llvm
4107 !0 = !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
4111 '``range``' Metadata
4112 ^^^^^^^^^^^^^^^^^^^^
4114 ``range`` metadata may be attached only to ``load``, ``call`` and ``invoke`` of
4115 integer types. It expresses the possible ranges the loaded value or the value
4116 returned by the called function at this call site is in. The ranges are
4117 represented with a flattened list of integers. The loaded value or the value
4118 returned is known to be in the union of the ranges defined by each consecutive
4119 pair. Each pair has the following properties:
4121 - The type must match the type loaded by the instruction.
4122 - The pair ``a,b`` represents the range ``[a,b)``.
4123 - Both ``a`` and ``b`` are constants.
4124 - The range is allowed to wrap.
4125 - The range should not represent the full or empty set. That is,
4128 In addition, the pairs must be in signed order of the lower bound and
4129 they must be non-contiguous.
4133 .. code-block:: llvm
4135 %a = load i8, i8* %x, align 1, !range !0 ; Can only be 0 or 1
4136 %b = load i8, i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
4137 %c = call i8 @foo(), !range !2 ; Can only be 0, 1, 3, 4 or 5
4138 %d = invoke i8 @bar() to label %cont
4139 unwind label %lpad, !range !3 ; Can only be -2, -1, 3, 4 or 5
4141 !0 = !{ i8 0, i8 2 }
4142 !1 = !{ i8 255, i8 2 }
4143 !2 = !{ i8 0, i8 2, i8 3, i8 6 }
4144 !3 = !{ i8 -2, i8 0, i8 3, i8 6 }
4149 It is sometimes useful to attach information to loop constructs. Currently,
4150 loop metadata is implemented as metadata attached to the branch instruction
4151 in the loop latch block. This type of metadata refer to a metadata node that is
4152 guaranteed to be separate for each loop. The loop identifier metadata is
4153 specified with the name ``llvm.loop``.
4155 The loop identifier metadata is implemented using a metadata that refers to
4156 itself to avoid merging it with any other identifier metadata, e.g.,
4157 during module linkage or function inlining. That is, each loop should refer
4158 to their own identification metadata even if they reside in separate functions.
4159 The following example contains loop identifier metadata for two separate loop
4162 .. code-block:: llvm
4167 The loop identifier metadata can be used to specify additional
4168 per-loop metadata. Any operands after the first operand can be treated
4169 as user-defined metadata. For example the ``llvm.loop.unroll.count``
4170 suggests an unroll factor to the loop unroller:
4172 .. code-block:: llvm
4174 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
4177 !1 = !{!"llvm.loop.unroll.count", i32 4}
4179 '``llvm.loop.vectorize``' and '``llvm.loop.interleave``'
4180 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4182 Metadata prefixed with ``llvm.loop.vectorize`` or ``llvm.loop.interleave`` are
4183 used to control per-loop vectorization and interleaving parameters such as
4184 vectorization width and interleave count. These metadata should be used in
4185 conjunction with ``llvm.loop`` loop identification metadata. The
4186 ``llvm.loop.vectorize`` and ``llvm.loop.interleave`` metadata are only
4187 optimization hints and the optimizer will only interleave and vectorize loops if
4188 it believes it is safe to do so. The ``llvm.mem.parallel_loop_access`` metadata
4189 which contains information about loop-carried memory dependencies can be helpful
4190 in determining the safety of these transformations.
4192 '``llvm.loop.interleave.count``' Metadata
4193 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4195 This metadata suggests an interleave count to the loop interleaver.
4196 The first operand is the string ``llvm.loop.interleave.count`` and the
4197 second operand is an integer specifying the interleave count. For
4200 .. code-block:: llvm
4202 !0 = !{!"llvm.loop.interleave.count", i32 4}
4204 Note that setting ``llvm.loop.interleave.count`` to 1 disables interleaving
4205 multiple iterations of the loop. If ``llvm.loop.interleave.count`` is set to 0
4206 then the interleave count will be determined automatically.
4208 '``llvm.loop.vectorize.enable``' Metadata
4209 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4211 This metadata selectively enables or disables vectorization for the loop. The
4212 first operand is the string ``llvm.loop.vectorize.enable`` and the second operand
4213 is a bit. If the bit operand value is 1 vectorization is enabled. A value of
4214 0 disables vectorization:
4216 .. code-block:: llvm
4218 !0 = !{!"llvm.loop.vectorize.enable", i1 0}
4219 !1 = !{!"llvm.loop.vectorize.enable", i1 1}
4221 '``llvm.loop.vectorize.width``' Metadata
4222 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4224 This metadata sets the target width of the vectorizer. The first
4225 operand is the string ``llvm.loop.vectorize.width`` and the second
4226 operand is an integer specifying the width. For example:
4228 .. code-block:: llvm
4230 !0 = !{!"llvm.loop.vectorize.width", i32 4}
4232 Note that setting ``llvm.loop.vectorize.width`` to 1 disables
4233 vectorization of the loop. If ``llvm.loop.vectorize.width`` is set to
4234 0 or if the loop does not have this metadata the width will be
4235 determined automatically.
4237 '``llvm.loop.unroll``'
4238 ^^^^^^^^^^^^^^^^^^^^^^
4240 Metadata prefixed with ``llvm.loop.unroll`` are loop unrolling
4241 optimization hints such as the unroll factor. ``llvm.loop.unroll``
4242 metadata should be used in conjunction with ``llvm.loop`` loop
4243 identification metadata. The ``llvm.loop.unroll`` metadata are only
4244 optimization hints and the unrolling will only be performed if the
4245 optimizer believes it is safe to do so.
4247 '``llvm.loop.unroll.count``' Metadata
4248 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4250 This metadata suggests an unroll factor to the loop unroller. The
4251 first operand is the string ``llvm.loop.unroll.count`` and the second
4252 operand is a positive integer specifying the unroll factor. For
4255 .. code-block:: llvm
4257 !0 = !{!"llvm.loop.unroll.count", i32 4}
4259 If the trip count of the loop is less than the unroll count the loop
4260 will be partially unrolled.
4262 '``llvm.loop.unroll.disable``' Metadata
4263 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4265 This metadata disables loop unrolling. The metadata has a single operand
4266 which is the string ``llvm.loop.unroll.disable``. For example:
4268 .. code-block:: llvm
4270 !0 = !{!"llvm.loop.unroll.disable"}
4272 '``llvm.loop.unroll.runtime.disable``' Metadata
4273 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4275 This metadata disables runtime loop unrolling. The metadata has a single
4276 operand which is the string ``llvm.loop.unroll.runtime.disable``. For example:
4278 .. code-block:: llvm
4280 !0 = !{!"llvm.loop.unroll.runtime.disable"}
4282 '``llvm.loop.unroll.full``' Metadata
4283 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4285 This metadata suggests that the loop should be unrolled fully. The
4286 metadata has a single operand which is the string ``llvm.loop.unroll.full``.
4289 .. code-block:: llvm
4291 !0 = !{!"llvm.loop.unroll.full"}
4296 Metadata types used to annotate memory accesses with information helpful
4297 for optimizations are prefixed with ``llvm.mem``.
4299 '``llvm.mem.parallel_loop_access``' Metadata
4300 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4302 The ``llvm.mem.parallel_loop_access`` metadata refers to a loop identifier,
4303 or metadata containing a list of loop identifiers for nested loops.
4304 The metadata is attached to memory accessing instructions and denotes that
4305 no loop carried memory dependence exist between it and other instructions denoted
4306 with the same loop identifier.
4308 Precisely, given two instructions ``m1`` and ``m2`` that both have the
4309 ``llvm.mem.parallel_loop_access`` metadata, with ``L1`` and ``L2`` being the
4310 set of loops associated with that metadata, respectively, then there is no loop
4311 carried dependence between ``m1`` and ``m2`` for loops in both ``L1`` and
4314 As a special case, if all memory accessing instructions in a loop have
4315 ``llvm.mem.parallel_loop_access`` metadata that refers to that loop, then the
4316 loop has no loop carried memory dependences and is considered to be a parallel
4319 Note that if not all memory access instructions have such metadata referring to
4320 the loop, then the loop is considered not being trivially parallel. Additional
4321 memory dependence analysis is required to make that determination. As a fail
4322 safe mechanism, this causes loops that were originally parallel to be considered
4323 sequential (if optimization passes that are unaware of the parallel semantics
4324 insert new memory instructions into the loop body).
4326 Example of a loop that is considered parallel due to its correct use of
4327 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
4328 metadata types that refer to the same loop identifier metadata.
4330 .. code-block:: llvm
4334 %val0 = load i32, i32* %arrayidx, !llvm.mem.parallel_loop_access !0
4336 store i32 %val0, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
4338 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
4344 It is also possible to have nested parallel loops. In that case the
4345 memory accesses refer to a list of loop identifier metadata nodes instead of
4346 the loop identifier metadata node directly:
4348 .. code-block:: llvm
4352 %val1 = load i32, i32* %arrayidx3, !llvm.mem.parallel_loop_access !2
4354 br label %inner.for.body
4358 %val0 = load i32, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
4360 store i32 %val0, i32* %arrayidx2, !llvm.mem.parallel_loop_access !0
4362 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
4366 store i32 %val1, i32* %arrayidx4, !llvm.mem.parallel_loop_access !2
4368 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
4370 outer.for.end: ; preds = %for.body
4372 !0 = !{!1, !2} ; a list of loop identifiers
4373 !1 = !{!1} ; an identifier for the inner loop
4374 !2 = !{!2} ; an identifier for the outer loop
4379 The ``llvm.bitsets`` global metadata is used to implement
4380 :doc:`bitsets <BitSets>`.
4382 Module Flags Metadata
4383 =====================
4385 Information about the module as a whole is difficult to convey to LLVM's
4386 subsystems. The LLVM IR isn't sufficient to transmit this information.
4387 The ``llvm.module.flags`` named metadata exists in order to facilitate
4388 this. These flags are in the form of key / value pairs --- much like a
4389 dictionary --- making it easy for any subsystem who cares about a flag to
4392 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
4393 Each triplet has the following form:
4395 - The first element is a *behavior* flag, which specifies the behavior
4396 when two (or more) modules are merged together, and it encounters two
4397 (or more) metadata with the same ID. The supported behaviors are
4399 - The second element is a metadata string that is a unique ID for the
4400 metadata. Each module may only have one flag entry for each unique ID (not
4401 including entries with the **Require** behavior).
4402 - The third element is the value of the flag.
4404 When two (or more) modules are merged together, the resulting
4405 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
4406 each unique metadata ID string, there will be exactly one entry in the merged
4407 modules ``llvm.module.flags`` metadata table, and the value for that entry will
4408 be determined by the merge behavior flag, as described below. The only exception
4409 is that entries with the *Require* behavior are always preserved.
4411 The following behaviors are supported:
4422 Emits an error if two values disagree, otherwise the resulting value
4423 is that of the operands.
4427 Emits a warning if two values disagree. The result value will be the
4428 operand for the flag from the first module being linked.
4432 Adds a requirement that another module flag be present and have a
4433 specified value after linking is performed. The value must be a
4434 metadata pair, where the first element of the pair is the ID of the
4435 module flag to be restricted, and the second element of the pair is
4436 the value the module flag should be restricted to. This behavior can
4437 be used to restrict the allowable results (via triggering of an
4438 error) of linking IDs with the **Override** behavior.
4442 Uses the specified value, regardless of the behavior or value of the
4443 other module. If both modules specify **Override**, but the values
4444 differ, an error will be emitted.
4448 Appends the two values, which are required to be metadata nodes.
4452 Appends the two values, which are required to be metadata
4453 nodes. However, duplicate entries in the second list are dropped
4454 during the append operation.
4456 It is an error for a particular unique flag ID to have multiple behaviors,
4457 except in the case of **Require** (which adds restrictions on another metadata
4458 value) or **Override**.
4460 An example of module flags:
4462 .. code-block:: llvm
4464 !0 = !{ i32 1, !"foo", i32 1 }
4465 !1 = !{ i32 4, !"bar", i32 37 }
4466 !2 = !{ i32 2, !"qux", i32 42 }
4467 !3 = !{ i32 3, !"qux",
4472 !llvm.module.flags = !{ !0, !1, !2, !3 }
4474 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
4475 if two or more ``!"foo"`` flags are seen is to emit an error if their
4476 values are not equal.
4478 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
4479 behavior if two or more ``!"bar"`` flags are seen is to use the value
4482 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
4483 behavior if two or more ``!"qux"`` flags are seen is to emit a
4484 warning if their values are not equal.
4486 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
4492 The behavior is to emit an error if the ``llvm.module.flags`` does not
4493 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
4496 Objective-C Garbage Collection Module Flags Metadata
4497 ----------------------------------------------------
4499 On the Mach-O platform, Objective-C stores metadata about garbage
4500 collection in a special section called "image info". The metadata
4501 consists of a version number and a bitmask specifying what types of
4502 garbage collection are supported (if any) by the file. If two or more
4503 modules are linked together their garbage collection metadata needs to
4504 be merged rather than appended together.
4506 The Objective-C garbage collection module flags metadata consists of the
4507 following key-value pairs:
4516 * - ``Objective-C Version``
4517 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
4519 * - ``Objective-C Image Info Version``
4520 - **[Required]** --- The version of the image info section. Currently
4523 * - ``Objective-C Image Info Section``
4524 - **[Required]** --- The section to place the metadata. Valid values are
4525 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
4526 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
4527 Objective-C ABI version 2.
4529 * - ``Objective-C Garbage Collection``
4530 - **[Required]** --- Specifies whether garbage collection is supported or
4531 not. Valid values are 0, for no garbage collection, and 2, for garbage
4532 collection supported.
4534 * - ``Objective-C GC Only``
4535 - **[Optional]** --- Specifies that only garbage collection is supported.
4536 If present, its value must be 6. This flag requires that the
4537 ``Objective-C Garbage Collection`` flag have the value 2.
4539 Some important flag interactions:
4541 - If a module with ``Objective-C Garbage Collection`` set to 0 is
4542 merged with a module with ``Objective-C Garbage Collection`` set to
4543 2, then the resulting module has the
4544 ``Objective-C Garbage Collection`` flag set to 0.
4545 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
4546 merged with a module with ``Objective-C GC Only`` set to 6.
4548 Automatic Linker Flags Module Flags Metadata
4549 --------------------------------------------
4551 Some targets support embedding flags to the linker inside individual object
4552 files. Typically this is used in conjunction with language extensions which
4553 allow source files to explicitly declare the libraries they depend on, and have
4554 these automatically be transmitted to the linker via object files.
4556 These flags are encoded in the IR using metadata in the module flags section,
4557 using the ``Linker Options`` key. The merge behavior for this flag is required
4558 to be ``AppendUnique``, and the value for the key is expected to be a metadata
4559 node which should be a list of other metadata nodes, each of which should be a
4560 list of metadata strings defining linker options.
4562 For example, the following metadata section specifies two separate sets of
4563 linker options, presumably to link against ``libz`` and the ``Cocoa``
4566 !0 = !{ i32 6, !"Linker Options",
4569 !{ !"-framework", !"Cocoa" } } }
4570 !llvm.module.flags = !{ !0 }
4572 The metadata encoding as lists of lists of options, as opposed to a collapsed
4573 list of options, is chosen so that the IR encoding can use multiple option
4574 strings to specify e.g., a single library, while still having that specifier be
4575 preserved as an atomic element that can be recognized by a target specific
4576 assembly writer or object file emitter.
4578 Each individual option is required to be either a valid option for the target's
4579 linker, or an option that is reserved by the target specific assembly writer or
4580 object file emitter. No other aspect of these options is defined by the IR.
4582 C type width Module Flags Metadata
4583 ----------------------------------
4585 The ARM backend emits a section into each generated object file describing the
4586 options that it was compiled with (in a compiler-independent way) to prevent
4587 linking incompatible objects, and to allow automatic library selection. Some
4588 of these options are not visible at the IR level, namely wchar_t width and enum
4591 To pass this information to the backend, these options are encoded in module
4592 flags metadata, using the following key-value pairs:
4602 - * 0 --- sizeof(wchar_t) == 4
4603 * 1 --- sizeof(wchar_t) == 2
4606 - * 0 --- Enums are at least as large as an ``int``.
4607 * 1 --- Enums are stored in the smallest integer type which can
4608 represent all of its values.
4610 For example, the following metadata section specifies that the module was
4611 compiled with a ``wchar_t`` width of 4 bytes, and the underlying type of an
4612 enum is the smallest type which can represent all of its values::
4614 !llvm.module.flags = !{!0, !1}
4615 !0 = !{i32 1, !"short_wchar", i32 1}
4616 !1 = !{i32 1, !"short_enum", i32 0}
4618 .. _intrinsicglobalvariables:
4620 Intrinsic Global Variables
4621 ==========================
4623 LLVM has a number of "magic" global variables that contain data that
4624 affect code generation or other IR semantics. These are documented here.
4625 All globals of this sort should have a section specified as
4626 "``llvm.metadata``". This section and all globals that start with
4627 "``llvm.``" are reserved for use by LLVM.
4631 The '``llvm.used``' Global Variable
4632 -----------------------------------
4634 The ``@llvm.used`` global is an array which has
4635 :ref:`appending linkage <linkage_appending>`. This array contains a list of
4636 pointers to named global variables, functions and aliases which may optionally
4637 have a pointer cast formed of bitcast or getelementptr. For example, a legal
4640 .. code-block:: llvm
4645 @llvm.used = appending global [2 x i8*] [
4647 i8* bitcast (i32* @Y to i8*)
4648 ], section "llvm.metadata"
4650 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
4651 and linker are required to treat the symbol as if there is a reference to the
4652 symbol that it cannot see (which is why they have to be named). For example, if
4653 a variable has internal linkage and no references other than that from the
4654 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
4655 references from inline asms and other things the compiler cannot "see", and
4656 corresponds to "``attribute((used))``" in GNU C.
4658 On some targets, the code generator must emit a directive to the
4659 assembler or object file to prevent the assembler and linker from
4660 molesting the symbol.
4662 .. _gv_llvmcompilerused:
4664 The '``llvm.compiler.used``' Global Variable
4665 --------------------------------------------
4667 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
4668 directive, except that it only prevents the compiler from touching the
4669 symbol. On targets that support it, this allows an intelligent linker to
4670 optimize references to the symbol without being impeded as it would be
4673 This is a rare construct that should only be used in rare circumstances,
4674 and should not be exposed to source languages.
4676 .. _gv_llvmglobalctors:
4678 The '``llvm.global_ctors``' Global Variable
4679 -------------------------------------------
4681 .. code-block:: llvm
4683 %0 = type { i32, void ()*, i8* }
4684 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor, i8* @data }]
4686 The ``@llvm.global_ctors`` array contains a list of constructor
4687 functions, priorities, and an optional associated global or function.
4688 The functions referenced by this array will be called in ascending order
4689 of priority (i.e. lowest first) when the module is loaded. The order of
4690 functions with the same priority is not defined.
4692 If the third field is present, non-null, and points to a global variable
4693 or function, the initializer function will only run if the associated
4694 data from the current module is not discarded.
4696 .. _llvmglobaldtors:
4698 The '``llvm.global_dtors``' Global Variable
4699 -------------------------------------------
4701 .. code-block:: llvm
4703 %0 = type { i32, void ()*, i8* }
4704 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor, i8* @data }]
4706 The ``@llvm.global_dtors`` array contains a list of destructor
4707 functions, priorities, and an optional associated global or function.
4708 The functions referenced by this array will be called in descending
4709 order of priority (i.e. highest first) when the module is unloaded. The
4710 order of functions with the same priority is not defined.
4712 If the third field is present, non-null, and points to a global variable
4713 or function, the destructor function will only run if the associated
4714 data from the current module is not discarded.
4716 Instruction Reference
4717 =====================
4719 The LLVM instruction set consists of several different classifications
4720 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
4721 instructions <binaryops>`, :ref:`bitwise binary
4722 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
4723 :ref:`other instructions <otherops>`.
4727 Terminator Instructions
4728 -----------------------
4730 As mentioned :ref:`previously <functionstructure>`, every basic block in a
4731 program ends with a "Terminator" instruction, which indicates which
4732 block should be executed after the current block is finished. These
4733 terminator instructions typically yield a '``void``' value: they produce
4734 control flow, not values (the one exception being the
4735 ':ref:`invoke <i_invoke>`' instruction).
4737 The terminator instructions are: ':ref:`ret <i_ret>`',
4738 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
4739 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
4740 ':ref:`resume <i_resume>`', ':ref:`catchpad <i_catchpad>`',
4741 ':ref:`catchendpad <i_catchendpad>`',
4742 ':ref:`catchret <i_catchret>`',
4743 ':ref:`cleanupret <i_cleanupret>`',
4744 ':ref:`terminatepad <i_terminatepad>`',
4745 and ':ref:`unreachable <i_unreachable>`'.
4749 '``ret``' Instruction
4750 ^^^^^^^^^^^^^^^^^^^^^
4757 ret <type> <value> ; Return a value from a non-void function
4758 ret void ; Return from void function
4763 The '``ret``' instruction is used to return control flow (and optionally
4764 a value) from a function back to the caller.
4766 There are two forms of the '``ret``' instruction: one that returns a
4767 value and then causes control flow, and one that just causes control
4773 The '``ret``' instruction optionally accepts a single argument, the
4774 return value. The type of the return value must be a ':ref:`first
4775 class <t_firstclass>`' type.
4777 A function is not :ref:`well formed <wellformed>` if it it has a non-void
4778 return type and contains a '``ret``' instruction with no return value or
4779 a return value with a type that does not match its type, or if it has a
4780 void return type and contains a '``ret``' instruction with a return
4786 When the '``ret``' instruction is executed, control flow returns back to
4787 the calling function's context. If the caller is a
4788 ":ref:`call <i_call>`" instruction, execution continues at the
4789 instruction after the call. If the caller was an
4790 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
4791 beginning of the "normal" destination block. If the instruction returns
4792 a value, that value shall set the call or invoke instruction's return
4798 .. code-block:: llvm
4800 ret i32 5 ; Return an integer value of 5
4801 ret void ; Return from a void function
4802 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
4806 '``br``' Instruction
4807 ^^^^^^^^^^^^^^^^^^^^
4814 br i1 <cond>, label <iftrue>, label <iffalse>
4815 br label <dest> ; Unconditional branch
4820 The '``br``' instruction is used to cause control flow to transfer to a
4821 different basic block in the current function. There are two forms of
4822 this instruction, corresponding to a conditional branch and an
4823 unconditional branch.
4828 The conditional branch form of the '``br``' instruction takes a single
4829 '``i1``' value and two '``label``' values. The unconditional form of the
4830 '``br``' instruction takes a single '``label``' value as a target.
4835 Upon execution of a conditional '``br``' instruction, the '``i1``'
4836 argument is evaluated. If the value is ``true``, control flows to the
4837 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
4838 to the '``iffalse``' ``label`` argument.
4843 .. code-block:: llvm
4846 %cond = icmp eq i32 %a, %b
4847 br i1 %cond, label %IfEqual, label %IfUnequal
4855 '``switch``' Instruction
4856 ^^^^^^^^^^^^^^^^^^^^^^^^
4863 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
4868 The '``switch``' instruction is used to transfer control flow to one of
4869 several different places. It is a generalization of the '``br``'
4870 instruction, allowing a branch to occur to one of many possible
4876 The '``switch``' instruction uses three parameters: an integer
4877 comparison value '``value``', a default '``label``' destination, and an
4878 array of pairs of comparison value constants and '``label``'s. The table
4879 is not allowed to contain duplicate constant entries.
4884 The ``switch`` instruction specifies a table of values and destinations.
4885 When the '``switch``' instruction is executed, this table is searched
4886 for the given value. If the value is found, control flow is transferred
4887 to the corresponding destination; otherwise, control flow is transferred
4888 to the default destination.
4893 Depending on properties of the target machine and the particular
4894 ``switch`` instruction, this instruction may be code generated in
4895 different ways. For example, it could be generated as a series of
4896 chained conditional branches or with a lookup table.
4901 .. code-block:: llvm
4903 ; Emulate a conditional br instruction
4904 %Val = zext i1 %value to i32
4905 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
4907 ; Emulate an unconditional br instruction
4908 switch i32 0, label %dest [ ]
4910 ; Implement a jump table:
4911 switch i32 %val, label %otherwise [ i32 0, label %onzero
4913 i32 2, label %ontwo ]
4917 '``indirectbr``' Instruction
4918 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4925 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
4930 The '``indirectbr``' instruction implements an indirect branch to a
4931 label within the current function, whose address is specified by
4932 "``address``". Address must be derived from a
4933 :ref:`blockaddress <blockaddress>` constant.
4938 The '``address``' argument is the address of the label to jump to. The
4939 rest of the arguments indicate the full set of possible destinations
4940 that the address may point to. Blocks are allowed to occur multiple
4941 times in the destination list, though this isn't particularly useful.
4943 This destination list is required so that dataflow analysis has an
4944 accurate understanding of the CFG.
4949 Control transfers to the block specified in the address argument. All
4950 possible destination blocks must be listed in the label list, otherwise
4951 this instruction has undefined behavior. This implies that jumps to
4952 labels defined in other functions have undefined behavior as well.
4957 This is typically implemented with a jump through a register.
4962 .. code-block:: llvm
4964 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
4968 '``invoke``' Instruction
4969 ^^^^^^^^^^^^^^^^^^^^^^^^
4976 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
4977 to label <normal label> unwind label <exception label>
4982 The '``invoke``' instruction causes control to transfer to a specified
4983 function, with the possibility of control flow transfer to either the
4984 '``normal``' label or the '``exception``' label. If the callee function
4985 returns with the "``ret``" instruction, control flow will return to the
4986 "normal" label. If the callee (or any indirect callees) returns via the
4987 ":ref:`resume <i_resume>`" instruction or other exception handling
4988 mechanism, control is interrupted and continued at the dynamically
4989 nearest "exception" label.
4991 The '``exception``' label is a `landing
4992 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
4993 '``exception``' label is required to have the
4994 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
4995 information about the behavior of the program after unwinding happens,
4996 as its first non-PHI instruction. The restrictions on the
4997 "``landingpad``" instruction's tightly couples it to the "``invoke``"
4998 instruction, so that the important information contained within the
4999 "``landingpad``" instruction can't be lost through normal code motion.
5004 This instruction requires several arguments:
5006 #. The optional "cconv" marker indicates which :ref:`calling
5007 convention <callingconv>` the call should use. If none is
5008 specified, the call defaults to using C calling conventions.
5009 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
5010 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
5012 #. '``ptr to function ty``': shall be the signature of the pointer to
5013 function value being invoked. In most cases, this is a direct
5014 function invocation, but indirect ``invoke``'s are just as possible,
5015 branching off an arbitrary pointer to function value.
5016 #. '``function ptr val``': An LLVM value containing a pointer to a
5017 function to be invoked.
5018 #. '``function args``': argument list whose types match the function
5019 signature argument types and parameter attributes. All arguments must
5020 be of :ref:`first class <t_firstclass>` type. If the function signature
5021 indicates the function accepts a variable number of arguments, the
5022 extra arguments can be specified.
5023 #. '``normal label``': the label reached when the called function
5024 executes a '``ret``' instruction.
5025 #. '``exception label``': the label reached when a callee returns via
5026 the :ref:`resume <i_resume>` instruction or other exception handling
5028 #. The optional :ref:`function attributes <fnattrs>` list. Only
5029 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
5030 attributes are valid here.
5035 This instruction is designed to operate as a standard '``call``'
5036 instruction in most regards. The primary difference is that it
5037 establishes an association with a label, which is used by the runtime
5038 library to unwind the stack.
5040 This instruction is used in languages with destructors to ensure that
5041 proper cleanup is performed in the case of either a ``longjmp`` or a
5042 thrown exception. Additionally, this is important for implementation of
5043 '``catch``' clauses in high-level languages that support them.
5045 For the purposes of the SSA form, the definition of the value returned
5046 by the '``invoke``' instruction is deemed to occur on the edge from the
5047 current block to the "normal" label. If the callee unwinds then no
5048 return value is available.
5053 .. code-block:: llvm
5055 %retval = invoke i32 @Test(i32 15) to label %Continue
5056 unwind label %TestCleanup ; i32:retval set
5057 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
5058 unwind label %TestCleanup ; i32:retval set
5062 '``resume``' Instruction
5063 ^^^^^^^^^^^^^^^^^^^^^^^^
5070 resume <type> <value>
5075 The '``resume``' instruction is a terminator instruction that has no
5081 The '``resume``' instruction requires one argument, which must have the
5082 same type as the result of any '``landingpad``' instruction in the same
5088 The '``resume``' instruction resumes propagation of an existing
5089 (in-flight) exception whose unwinding was interrupted with a
5090 :ref:`landingpad <i_landingpad>` instruction.
5095 .. code-block:: llvm
5097 resume { i8*, i32 } %exn
5101 '``catchpad``' Instruction
5102 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5109 <resultval> = catchpad <resultty> [<args>*]
5110 to label <normal label> unwind label <exception label>
5115 The '``catchpad``' instruction is used by `LLVM's exception handling
5116 system <ExceptionHandling.html#overview>`_ to specify that a basic block
5117 is a catch block --- one where a personality routine attempts to transfer
5118 control to catch an exception.
5119 The ``args`` correspond to whatever information the personality
5120 routine requires to know if this is an appropriate place to catch the
5121 exception. Control is tranfered to the ``exception`` label if the
5122 ``catchpad`` is not an appropriate handler for the in-flight exception.
5123 The ``normal`` label should contain the code found in the ``catch``
5124 portion of a ``try``/``catch`` sequence. It defines values supplied by
5125 the :ref:`personality function <personalityfn>` upon re-entry to the
5126 function. The ``resultval`` has the type ``resultty``.
5131 The instruction takes a list of arbitrary values which are interpreted
5132 by the :ref:`personality function <personalityfn>`.
5134 The ``catchpad`` must be provided a ``normal`` label to transfer control
5135 to if the ``catchpad`` matches the exception and an ``exception``
5136 label to transfer control to if it doesn't.
5141 The '``catchpad``' instruction defines the values which are set by the
5142 :ref:`personality function <personalityfn>` upon re-entry to the function, and
5143 therefore the "result type" of the ``catchpad`` instruction. As with
5144 calling conventions, how the personality function results are
5145 represented in LLVM IR is target specific.
5147 When the call stack is being unwound due to an exception being thrown,
5148 the exception is compared against the ``args``. If it doesn't match,
5149 then control is transfered to the ``exception`` basic block.
5151 The ``catchpad`` instruction has several restrictions:
5153 - A catch block is a basic block which is the unwind destination of
5154 an exceptional instruction.
5155 - A catch block must have a '``catchpad``' instruction as its
5156 first non-PHI instruction.
5157 - A catch block's ``exception`` edge must refer to a catch block or a
5159 - There can be only one '``catchpad``' instruction within the
5161 - A basic block that is not a catch block may not include a
5162 '``catchpad``' instruction.
5163 - It is undefined behavior for control to transfer from a ``catchpad`` to a
5164 ``cleanupret`` without first executing a ``catchret`` and a subsequent
5166 - It is undefined behavior for control to transfer from a ``catchpad`` to a
5167 ``ret`` without first executing a ``catchret``.
5172 .. code-block:: llvm
5174 ;; A catch block which can catch an integer.
5175 %res = catchpad { i8*, i32 } [i8** @_ZTIi]
5176 to label %int.handler unwind label %terminate
5180 '``catchendpad``' Instruction
5181 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5188 catchendpad unwind label <nextaction>
5189 catchendpad unwind to caller
5194 The '``catchendpad``' instruction is used by `LLVM's exception handling
5195 system <ExceptionHandling.html#overview>`_ to communicate to the
5196 :ref:`personality function <personalityfn>` which invokes are associated
5197 with a chain of :ref:`catchpad <i_catchpad>` instructions.
5199 The ``nextaction`` label indicates where control should transfer to if
5200 none of the ``catchpad`` instructions are suitable for catching the
5201 in-flight exception.
5203 If a ``nextaction`` label is not present, the instruction unwinds out of
5204 its parent function. The
5205 :ref:`personality function <personalityfn>` will continue processing
5206 exception handling actions in the caller.
5211 The instruction optionally takes a label, ``nextaction``, indicating
5212 where control should transfer to if none of the preceding
5213 ``catchpad`` instructions are suitable for the in-flight exception.
5218 When the call stack is being unwound due to an exception being thrown
5219 and none of the constituent ``catchpad`` instructions match, then
5220 control is transfered to ``nextaction`` if it is present. If it is not
5221 present, control is transfered to the caller.
5223 The ``catchendpad`` instruction has several restrictions:
5225 - A catch-end block is a basic block which is the unwind destination of
5226 an exceptional instruction.
5227 - A catch-end block must have a '``catchendpad``' instruction as its
5228 first non-PHI instruction.
5229 - There can be only one '``catchendpad``' instruction within the
5231 - A basic block that is not a catch-end block may not include a
5232 '``catchendpad``' instruction.
5233 - Exactly one catch block may unwind to a ``catchendpad``.
5234 - The unwind target of invokes between a ``catchpad`` and a
5235 corresponding ``catchret`` must be its ``catchendpad``.
5240 .. code-block:: llvm
5242 catchendpad unwind label %terminate
5243 catchendpad unwind to caller
5247 '``catchret``' Instruction
5248 ^^^^^^^^^^^^^^^^^^^^^^^^^^
5255 catchret label <normal>
5260 The '``catchret``' instruction is a terminator instruction that has a
5267 The '``catchret``' instruction requires one argument which specifies
5268 where control will transfer to next.
5273 The '``catchret``' instruction ends the existing (in-flight) exception
5274 whose unwinding was interrupted with a
5275 :ref:`catchpad <i_catchpad>` instruction.
5276 The :ref:`personality function <personalityfn>` gets a chance to execute
5277 arbitrary code to, for example, run a C++ destructor.
5278 Control then transfers to ``normal``.
5283 .. code-block:: llvm
5285 catchret label %continue
5289 '``cleanupret``' Instruction
5290 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5297 cleanupret <type> <value> unwind label <continue>
5298 cleanupret <type> <value> unwind to caller
5303 The '``cleanupret``' instruction is a terminator instruction that has
5304 an optional successor.
5310 The '``cleanupret``' instruction requires one argument, which must have the
5311 same type as the result of any '``cleanuppad``' instruction in the same
5312 function. It also has an optional successor, ``continue``.
5317 The '``cleanupret``' instruction indicates to the
5318 :ref:`personality function <personalityfn>` that one
5319 :ref:`cleanuppad <i_cleanuppad>` it transferred control to has ended.
5320 It transfers control to ``continue`` or unwinds out of the function.
5325 .. code-block:: llvm
5327 cleanupret void unwind to caller
5328 cleanupret { i8*, i32 } %exn unwind label %continue
5332 '``terminatepad``' Instruction
5333 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5340 terminatepad [<args>*] unwind label <exception label>
5341 terminatepad [<args>*] unwind to caller
5346 The '``terminatepad``' instruction is used by `LLVM's exception handling
5347 system <ExceptionHandling.html#overview>`_ to specify that a basic block
5348 is a terminate block --- one where a personality routine may decide to
5349 terminate the program.
5350 The ``args`` correspond to whatever information the personality
5351 routine requires to know if this is an appropriate place to terminate the
5352 program. Control is transferred to the ``exception`` label if the
5353 personality routine decides not to terminate the program for the
5354 in-flight exception.
5359 The instruction takes a list of arbitrary values which are interpreted
5360 by the :ref:`personality function <personalityfn>`.
5362 The ``terminatepad`` may be given an ``exception`` label to
5363 transfer control to if the in-flight exception matches the ``args``.
5368 When the call stack is being unwound due to an exception being thrown,
5369 the exception is compared against the ``args``. If it matches,
5370 then control is transfered to the ``exception`` basic block. Otherwise,
5371 the program is terminated via personality-specific means. Typically,
5372 the first argument to ``terminatepad`` specifies what function the
5373 personality should defer to in order to terminate the program.
5375 The ``terminatepad`` instruction has several restrictions:
5377 - A terminate block is a basic block which is the unwind destination of
5378 an exceptional instruction.
5379 - A terminate block must have a '``terminatepad``' instruction as its
5380 first non-PHI instruction.
5381 - There can be only one '``terminatepad``' instruction within the
5383 - A basic block that is not a terminate block may not include a
5384 '``terminatepad``' instruction.
5389 .. code-block:: llvm
5391 ;; A terminate block which only permits integers.
5392 terminatepad [i8** @_ZTIi] unwind label %continue
5396 '``unreachable``' Instruction
5397 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5409 The '``unreachable``' instruction has no defined semantics. This
5410 instruction is used to inform the optimizer that a particular portion of
5411 the code is not reachable. This can be used to indicate that the code
5412 after a no-return function cannot be reached, and other facts.
5417 The '``unreachable``' instruction has no defined semantics.
5424 Binary operators are used to do most of the computation in a program.
5425 They require two operands of the same type, execute an operation on
5426 them, and produce a single value. The operands might represent multiple
5427 data, as is the case with the :ref:`vector <t_vector>` data type. The
5428 result value has the same type as its operands.
5430 There are several different binary operators:
5434 '``add``' Instruction
5435 ^^^^^^^^^^^^^^^^^^^^^
5442 <result> = add <ty> <op1>, <op2> ; yields ty:result
5443 <result> = add nuw <ty> <op1>, <op2> ; yields ty:result
5444 <result> = add nsw <ty> <op1>, <op2> ; yields ty:result
5445 <result> = add nuw nsw <ty> <op1>, <op2> ; yields ty:result
5450 The '``add``' instruction returns the sum of its two operands.
5455 The two arguments to the '``add``' instruction must be
5456 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5457 arguments must have identical types.
5462 The value produced is the integer sum of the two operands.
5464 If the sum has unsigned overflow, the result returned is the
5465 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
5468 Because LLVM integers use a two's complement representation, this
5469 instruction is appropriate for both signed and unsigned integers.
5471 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
5472 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
5473 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
5474 unsigned and/or signed overflow, respectively, occurs.
5479 .. code-block:: llvm
5481 <result> = add i32 4, %var ; yields i32:result = 4 + %var
5485 '``fadd``' Instruction
5486 ^^^^^^^^^^^^^^^^^^^^^^
5493 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
5498 The '``fadd``' instruction returns the sum of its two operands.
5503 The two arguments to the '``fadd``' instruction must be :ref:`floating
5504 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
5505 Both arguments must have identical types.
5510 The value produced is the floating point sum of the two operands. This
5511 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
5512 which are optimization hints to enable otherwise unsafe floating point
5518 .. code-block:: llvm
5520 <result> = fadd float 4.0, %var ; yields float:result = 4.0 + %var
5522 '``sub``' Instruction
5523 ^^^^^^^^^^^^^^^^^^^^^
5530 <result> = sub <ty> <op1>, <op2> ; yields ty:result
5531 <result> = sub nuw <ty> <op1>, <op2> ; yields ty:result
5532 <result> = sub nsw <ty> <op1>, <op2> ; yields ty:result
5533 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields ty:result
5538 The '``sub``' instruction returns the difference of its two operands.
5540 Note that the '``sub``' instruction is used to represent the '``neg``'
5541 instruction present in most other intermediate representations.
5546 The two arguments to the '``sub``' instruction must be
5547 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5548 arguments must have identical types.
5553 The value produced is the integer difference of the two operands.
5555 If the difference has unsigned overflow, the result returned is the
5556 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
5559 Because LLVM integers use a two's complement representation, this
5560 instruction is appropriate for both signed and unsigned integers.
5562 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
5563 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
5564 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
5565 unsigned and/or signed overflow, respectively, occurs.
5570 .. code-block:: llvm
5572 <result> = sub i32 4, %var ; yields i32:result = 4 - %var
5573 <result> = sub i32 0, %val ; yields i32:result = -%var
5577 '``fsub``' Instruction
5578 ^^^^^^^^^^^^^^^^^^^^^^
5585 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
5590 The '``fsub``' instruction returns the difference of its two operands.
5592 Note that the '``fsub``' instruction is used to represent the '``fneg``'
5593 instruction present in most other intermediate representations.
5598 The two arguments to the '``fsub``' instruction must be :ref:`floating
5599 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
5600 Both arguments must have identical types.
5605 The value produced is the floating point difference of the two operands.
5606 This instruction can also take any number of :ref:`fast-math
5607 flags <fastmath>`, which are optimization hints to enable otherwise
5608 unsafe floating point optimizations:
5613 .. code-block:: llvm
5615 <result> = fsub float 4.0, %var ; yields float:result = 4.0 - %var
5616 <result> = fsub float -0.0, %val ; yields float:result = -%var
5618 '``mul``' Instruction
5619 ^^^^^^^^^^^^^^^^^^^^^
5626 <result> = mul <ty> <op1>, <op2> ; yields ty:result
5627 <result> = mul nuw <ty> <op1>, <op2> ; yields ty:result
5628 <result> = mul nsw <ty> <op1>, <op2> ; yields ty:result
5629 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields ty:result
5634 The '``mul``' instruction returns the product of its two operands.
5639 The two arguments to the '``mul``' instruction must be
5640 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5641 arguments must have identical types.
5646 The value produced is the integer product of the two operands.
5648 If the result of the multiplication has unsigned overflow, the result
5649 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
5650 bit width of the result.
5652 Because LLVM integers use a two's complement representation, and the
5653 result is the same width as the operands, this instruction returns the
5654 correct result for both signed and unsigned integers. If a full product
5655 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
5656 sign-extended or zero-extended as appropriate to the width of the full
5659 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
5660 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
5661 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
5662 unsigned and/or signed overflow, respectively, occurs.
5667 .. code-block:: llvm
5669 <result> = mul i32 4, %var ; yields i32:result = 4 * %var
5673 '``fmul``' Instruction
5674 ^^^^^^^^^^^^^^^^^^^^^^
5681 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
5686 The '``fmul``' instruction returns the product of its two operands.
5691 The two arguments to the '``fmul``' instruction must be :ref:`floating
5692 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
5693 Both arguments must have identical types.
5698 The value produced is the floating point product of the two operands.
5699 This instruction can also take any number of :ref:`fast-math
5700 flags <fastmath>`, which are optimization hints to enable otherwise
5701 unsafe floating point optimizations:
5706 .. code-block:: llvm
5708 <result> = fmul float 4.0, %var ; yields float:result = 4.0 * %var
5710 '``udiv``' Instruction
5711 ^^^^^^^^^^^^^^^^^^^^^^
5718 <result> = udiv <ty> <op1>, <op2> ; yields ty:result
5719 <result> = udiv exact <ty> <op1>, <op2> ; yields ty:result
5724 The '``udiv``' instruction returns the quotient of its two operands.
5729 The two arguments to the '``udiv``' instruction must be
5730 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5731 arguments must have identical types.
5736 The value produced is the unsigned integer quotient of the two operands.
5738 Note that unsigned integer division and signed integer division are
5739 distinct operations; for signed integer division, use '``sdiv``'.
5741 Division by zero leads to undefined behavior.
5743 If the ``exact`` keyword is present, the result value of the ``udiv`` is
5744 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
5745 such, "((a udiv exact b) mul b) == a").
5750 .. code-block:: llvm
5752 <result> = udiv i32 4, %var ; yields i32:result = 4 / %var
5754 '``sdiv``' Instruction
5755 ^^^^^^^^^^^^^^^^^^^^^^
5762 <result> = sdiv <ty> <op1>, <op2> ; yields ty:result
5763 <result> = sdiv exact <ty> <op1>, <op2> ; yields ty:result
5768 The '``sdiv``' instruction returns the quotient of its two operands.
5773 The two arguments to the '``sdiv``' instruction must be
5774 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5775 arguments must have identical types.
5780 The value produced is the signed integer quotient of the two operands
5781 rounded towards zero.
5783 Note that signed integer division and unsigned integer division are
5784 distinct operations; for unsigned integer division, use '``udiv``'.
5786 Division by zero leads to undefined behavior. Overflow also leads to
5787 undefined behavior; this is a rare case, but can occur, for example, by
5788 doing a 32-bit division of -2147483648 by -1.
5790 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
5791 a :ref:`poison value <poisonvalues>` if the result would be rounded.
5796 .. code-block:: llvm
5798 <result> = sdiv i32 4, %var ; yields i32:result = 4 / %var
5802 '``fdiv``' Instruction
5803 ^^^^^^^^^^^^^^^^^^^^^^
5810 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
5815 The '``fdiv``' instruction returns the quotient of its two operands.
5820 The two arguments to the '``fdiv``' instruction must be :ref:`floating
5821 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
5822 Both arguments must have identical types.
5827 The value produced is the floating point quotient of the two operands.
5828 This instruction can also take any number of :ref:`fast-math
5829 flags <fastmath>`, which are optimization hints to enable otherwise
5830 unsafe floating point optimizations:
5835 .. code-block:: llvm
5837 <result> = fdiv float 4.0, %var ; yields float:result = 4.0 / %var
5839 '``urem``' Instruction
5840 ^^^^^^^^^^^^^^^^^^^^^^
5847 <result> = urem <ty> <op1>, <op2> ; yields ty:result
5852 The '``urem``' instruction returns the remainder from the unsigned
5853 division of its two arguments.
5858 The two arguments to the '``urem``' instruction must be
5859 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5860 arguments must have identical types.
5865 This instruction returns the unsigned integer *remainder* of a division.
5866 This instruction always performs an unsigned division to get the
5869 Note that unsigned integer remainder and signed integer remainder are
5870 distinct operations; for signed integer remainder, use '``srem``'.
5872 Taking the remainder of a division by zero leads to undefined behavior.
5877 .. code-block:: llvm
5879 <result> = urem i32 4, %var ; yields i32:result = 4 % %var
5881 '``srem``' Instruction
5882 ^^^^^^^^^^^^^^^^^^^^^^
5889 <result> = srem <ty> <op1>, <op2> ; yields ty:result
5894 The '``srem``' instruction returns the remainder from the signed
5895 division of its two operands. This instruction can also take
5896 :ref:`vector <t_vector>` versions of the values in which case the elements
5902 The two arguments to the '``srem``' instruction must be
5903 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
5904 arguments must have identical types.
5909 This instruction returns the *remainder* of a division (where the result
5910 is either zero or has the same sign as the dividend, ``op1``), not the
5911 *modulo* operator (where the result is either zero or has the same sign
5912 as the divisor, ``op2``) of a value. For more information about the
5913 difference, see `The Math
5914 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
5915 table of how this is implemented in various languages, please see
5917 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
5919 Note that signed integer remainder and unsigned integer remainder are
5920 distinct operations; for unsigned integer remainder, use '``urem``'.
5922 Taking the remainder of a division by zero leads to undefined behavior.
5923 Overflow also leads to undefined behavior; this is a rare case, but can
5924 occur, for example, by taking the remainder of a 32-bit division of
5925 -2147483648 by -1. (The remainder doesn't actually overflow, but this
5926 rule lets srem be implemented using instructions that return both the
5927 result of the division and the remainder.)
5932 .. code-block:: llvm
5934 <result> = srem i32 4, %var ; yields i32:result = 4 % %var
5938 '``frem``' Instruction
5939 ^^^^^^^^^^^^^^^^^^^^^^
5946 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
5951 The '``frem``' instruction returns the remainder from the division of
5957 The two arguments to the '``frem``' instruction must be :ref:`floating
5958 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
5959 Both arguments must have identical types.
5964 This instruction returns the *remainder* of a division. The remainder
5965 has the same sign as the dividend. This instruction can also take any
5966 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
5967 to enable otherwise unsafe floating point optimizations:
5972 .. code-block:: llvm
5974 <result> = frem float 4.0, %var ; yields float:result = 4.0 % %var
5978 Bitwise Binary Operations
5979 -------------------------
5981 Bitwise binary operators are used to do various forms of bit-twiddling
5982 in a program. They are generally very efficient instructions and can
5983 commonly be strength reduced from other instructions. They require two
5984 operands of the same type, execute an operation on them, and produce a
5985 single value. The resulting value is the same type as its operands.
5987 '``shl``' Instruction
5988 ^^^^^^^^^^^^^^^^^^^^^
5995 <result> = shl <ty> <op1>, <op2> ; yields ty:result
5996 <result> = shl nuw <ty> <op1>, <op2> ; yields ty:result
5997 <result> = shl nsw <ty> <op1>, <op2> ; yields ty:result
5998 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields ty:result
6003 The '``shl``' instruction returns the first operand shifted to the left
6004 a specified number of bits.
6009 Both arguments to the '``shl``' instruction must be the same
6010 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
6011 '``op2``' is treated as an unsigned value.
6016 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
6017 where ``n`` is the width of the result. If ``op2`` is (statically or
6018 dynamically) equal to or larger than the number of bits in
6019 ``op1``, the result is undefined. If the arguments are vectors, each
6020 vector element of ``op1`` is shifted by the corresponding shift amount
6023 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
6024 value <poisonvalues>` if it shifts out any non-zero bits. If the
6025 ``nsw`` keyword is present, then the shift produces a :ref:`poison
6026 value <poisonvalues>` if it shifts out any bits that disagree with the
6027 resultant sign bit. As such, NUW/NSW have the same semantics as they
6028 would if the shift were expressed as a mul instruction with the same
6029 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
6034 .. code-block:: llvm
6036 <result> = shl i32 4, %var ; yields i32: 4 << %var
6037 <result> = shl i32 4, 2 ; yields i32: 16
6038 <result> = shl i32 1, 10 ; yields i32: 1024
6039 <result> = shl i32 1, 32 ; undefined
6040 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
6042 '``lshr``' Instruction
6043 ^^^^^^^^^^^^^^^^^^^^^^
6050 <result> = lshr <ty> <op1>, <op2> ; yields ty:result
6051 <result> = lshr exact <ty> <op1>, <op2> ; yields ty:result
6056 The '``lshr``' instruction (logical shift right) returns the first
6057 operand shifted to the right a specified number of bits with zero fill.
6062 Both arguments to the '``lshr``' instruction must be the same
6063 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
6064 '``op2``' is treated as an unsigned value.
6069 This instruction always performs a logical shift right operation. The
6070 most significant bits of the result will be filled with zero bits after
6071 the shift. If ``op2`` is (statically or dynamically) equal to or larger
6072 than the number of bits in ``op1``, the result is undefined. If the
6073 arguments are vectors, each vector element of ``op1`` is shifted by the
6074 corresponding shift amount in ``op2``.
6076 If the ``exact`` keyword is present, the result value of the ``lshr`` is
6077 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
6083 .. code-block:: llvm
6085 <result> = lshr i32 4, 1 ; yields i32:result = 2
6086 <result> = lshr i32 4, 2 ; yields i32:result = 1
6087 <result> = lshr i8 4, 3 ; yields i8:result = 0
6088 <result> = lshr i8 -2, 1 ; yields i8:result = 0x7F
6089 <result> = lshr i32 1, 32 ; undefined
6090 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
6092 '``ashr``' Instruction
6093 ^^^^^^^^^^^^^^^^^^^^^^
6100 <result> = ashr <ty> <op1>, <op2> ; yields ty:result
6101 <result> = ashr exact <ty> <op1>, <op2> ; yields ty:result
6106 The '``ashr``' instruction (arithmetic shift right) returns the first
6107 operand shifted to the right a specified number of bits with sign
6113 Both arguments to the '``ashr``' instruction must be the same
6114 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
6115 '``op2``' is treated as an unsigned value.
6120 This instruction always performs an arithmetic shift right operation,
6121 The most significant bits of the result will be filled with the sign bit
6122 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
6123 than the number of bits in ``op1``, the result is undefined. If the
6124 arguments are vectors, each vector element of ``op1`` is shifted by the
6125 corresponding shift amount in ``op2``.
6127 If the ``exact`` keyword is present, the result value of the ``ashr`` is
6128 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
6134 .. code-block:: llvm
6136 <result> = ashr i32 4, 1 ; yields i32:result = 2
6137 <result> = ashr i32 4, 2 ; yields i32:result = 1
6138 <result> = ashr i8 4, 3 ; yields i8:result = 0
6139 <result> = ashr i8 -2, 1 ; yields i8:result = -1
6140 <result> = ashr i32 1, 32 ; undefined
6141 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
6143 '``and``' Instruction
6144 ^^^^^^^^^^^^^^^^^^^^^
6151 <result> = and <ty> <op1>, <op2> ; yields ty:result
6156 The '``and``' instruction returns the bitwise logical and of its two
6162 The two arguments to the '``and``' instruction must be
6163 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6164 arguments must have identical types.
6169 The truth table used for the '``and``' instruction is:
6186 .. code-block:: llvm
6188 <result> = and i32 4, %var ; yields i32:result = 4 & %var
6189 <result> = and i32 15, 40 ; yields i32:result = 8
6190 <result> = and i32 4, 8 ; yields i32:result = 0
6192 '``or``' Instruction
6193 ^^^^^^^^^^^^^^^^^^^^
6200 <result> = or <ty> <op1>, <op2> ; yields ty:result
6205 The '``or``' instruction returns the bitwise logical inclusive or of its
6211 The two arguments to the '``or``' instruction must be
6212 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6213 arguments must have identical types.
6218 The truth table used for the '``or``' instruction is:
6237 <result> = or i32 4, %var ; yields i32:result = 4 | %var
6238 <result> = or i32 15, 40 ; yields i32:result = 47
6239 <result> = or i32 4, 8 ; yields i32:result = 12
6241 '``xor``' Instruction
6242 ^^^^^^^^^^^^^^^^^^^^^
6249 <result> = xor <ty> <op1>, <op2> ; yields ty:result
6254 The '``xor``' instruction returns the bitwise logical exclusive or of
6255 its two operands. The ``xor`` is used to implement the "one's
6256 complement" operation, which is the "~" operator in C.
6261 The two arguments to the '``xor``' instruction must be
6262 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
6263 arguments must have identical types.
6268 The truth table used for the '``xor``' instruction is:
6285 .. code-block:: llvm
6287 <result> = xor i32 4, %var ; yields i32:result = 4 ^ %var
6288 <result> = xor i32 15, 40 ; yields i32:result = 39
6289 <result> = xor i32 4, 8 ; yields i32:result = 12
6290 <result> = xor i32 %V, -1 ; yields i32:result = ~%V
6295 LLVM supports several instructions to represent vector operations in a
6296 target-independent manner. These instructions cover the element-access
6297 and vector-specific operations needed to process vectors effectively.
6298 While LLVM does directly support these vector operations, many
6299 sophisticated algorithms will want to use target-specific intrinsics to
6300 take full advantage of a specific target.
6302 .. _i_extractelement:
6304 '``extractelement``' Instruction
6305 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6312 <result> = extractelement <n x <ty>> <val>, <ty2> <idx> ; yields <ty>
6317 The '``extractelement``' instruction extracts a single scalar element
6318 from a vector at a specified index.
6323 The first operand of an '``extractelement``' instruction is a value of
6324 :ref:`vector <t_vector>` type. The second operand is an index indicating
6325 the position from which to extract the element. The index may be a
6326 variable of any integer type.
6331 The result is a scalar of the same type as the element type of ``val``.
6332 Its value is the value at position ``idx`` of ``val``. If ``idx``
6333 exceeds the length of ``val``, the results are undefined.
6338 .. code-block:: llvm
6340 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
6342 .. _i_insertelement:
6344 '``insertelement``' Instruction
6345 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6352 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, <ty2> <idx> ; yields <n x <ty>>
6357 The '``insertelement``' instruction inserts a scalar element into a
6358 vector at a specified index.
6363 The first operand of an '``insertelement``' instruction is a value of
6364 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
6365 type must equal the element type of the first operand. The third operand
6366 is an index indicating the position at which to insert the value. The
6367 index may be a variable of any integer type.
6372 The result is a vector of the same type as ``val``. Its element values
6373 are those of ``val`` except at position ``idx``, where it gets the value
6374 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
6380 .. code-block:: llvm
6382 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
6384 .. _i_shufflevector:
6386 '``shufflevector``' Instruction
6387 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6394 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
6399 The '``shufflevector``' instruction constructs a permutation of elements
6400 from two input vectors, returning a vector with the same element type as
6401 the input and length that is the same as the shuffle mask.
6406 The first two operands of a '``shufflevector``' instruction are vectors
6407 with the same type. The third argument is a shuffle mask whose element
6408 type is always 'i32'. The result of the instruction is a vector whose
6409 length is the same as the shuffle mask and whose element type is the
6410 same as the element type of the first two operands.
6412 The shuffle mask operand is required to be a constant vector with either
6413 constant integer or undef values.
6418 The elements of the two input vectors are numbered from left to right
6419 across both of the vectors. The shuffle mask operand specifies, for each
6420 element of the result vector, which element of the two input vectors the
6421 result element gets. The element selector may be undef (meaning "don't
6422 care") and the second operand may be undef if performing a shuffle from
6428 .. code-block:: llvm
6430 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
6431 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
6432 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
6433 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
6434 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
6435 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
6436 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
6437 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
6439 Aggregate Operations
6440 --------------------
6442 LLVM supports several instructions for working with
6443 :ref:`aggregate <t_aggregate>` values.
6447 '``extractvalue``' Instruction
6448 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6455 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
6460 The '``extractvalue``' instruction extracts the value of a member field
6461 from an :ref:`aggregate <t_aggregate>` value.
6466 The first operand of an '``extractvalue``' instruction is a value of
6467 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
6468 constant indices to specify which value to extract in a similar manner
6469 as indices in a '``getelementptr``' instruction.
6471 The major differences to ``getelementptr`` indexing are:
6473 - Since the value being indexed is not a pointer, the first index is
6474 omitted and assumed to be zero.
6475 - At least one index must be specified.
6476 - Not only struct indices but also array indices must be in bounds.
6481 The result is the value at the position in the aggregate specified by
6487 .. code-block:: llvm
6489 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
6493 '``insertvalue``' Instruction
6494 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6501 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
6506 The '``insertvalue``' instruction inserts a value into a member field in
6507 an :ref:`aggregate <t_aggregate>` value.
6512 The first operand of an '``insertvalue``' instruction is a value of
6513 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
6514 a first-class value to insert. The following operands are constant
6515 indices indicating the position at which to insert the value in a
6516 similar manner as indices in a '``extractvalue``' instruction. The value
6517 to insert must have the same type as the value identified by the
6523 The result is an aggregate of the same type as ``val``. Its value is
6524 that of ``val`` except that the value at the position specified by the
6525 indices is that of ``elt``.
6530 .. code-block:: llvm
6532 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
6533 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
6534 %agg3 = insertvalue {i32, {float}} undef, float %val, 1, 0 ; yields {i32 undef, {float %val}}
6538 Memory Access and Addressing Operations
6539 ---------------------------------------
6541 A key design point of an SSA-based representation is how it represents
6542 memory. In LLVM, no memory locations are in SSA form, which makes things
6543 very simple. This section describes how to read, write, and allocate
6548 '``alloca``' Instruction
6549 ^^^^^^^^^^^^^^^^^^^^^^^^
6556 <result> = alloca [inalloca] <type> [, <ty> <NumElements>] [, align <alignment>] ; yields type*:result
6561 The '``alloca``' instruction allocates memory on the stack frame of the
6562 currently executing function, to be automatically released when this
6563 function returns to its caller. The object is always allocated in the
6564 generic address space (address space zero).
6569 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
6570 bytes of memory on the runtime stack, returning a pointer of the
6571 appropriate type to the program. If "NumElements" is specified, it is
6572 the number of elements allocated, otherwise "NumElements" is defaulted
6573 to be one. If a constant alignment is specified, the value result of the
6574 allocation is guaranteed to be aligned to at least that boundary. The
6575 alignment may not be greater than ``1 << 29``. If not specified, or if
6576 zero, the target can choose to align the allocation on any convenient
6577 boundary compatible with the type.
6579 '``type``' may be any sized type.
6584 Memory is allocated; a pointer is returned. The operation is undefined
6585 if there is insufficient stack space for the allocation. '``alloca``'d
6586 memory is automatically released when the function returns. The
6587 '``alloca``' instruction is commonly used to represent automatic
6588 variables that must have an address available. When the function returns
6589 (either with the ``ret`` or ``resume`` instructions), the memory is
6590 reclaimed. Allocating zero bytes is legal, but the result is undefined.
6591 The order in which memory is allocated (ie., which way the stack grows)
6597 .. code-block:: llvm
6599 %ptr = alloca i32 ; yields i32*:ptr
6600 %ptr = alloca i32, i32 4 ; yields i32*:ptr
6601 %ptr = alloca i32, i32 4, align 1024 ; yields i32*:ptr
6602 %ptr = alloca i32, align 1024 ; yields i32*:ptr
6606 '``load``' Instruction
6607 ^^^^^^^^^^^^^^^^^^^^^^
6614 <result> = load [volatile] <ty>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>][, !nonnull !<index>][, !dereferenceable !<index>][, !dereferenceable_or_null !<index>]
6615 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
6616 !<index> = !{ i32 1 }
6621 The '``load``' instruction is used to read from memory.
6626 The argument to the ``load`` instruction specifies the memory address
6627 from which to load. The type specified must be a :ref:`first
6628 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
6629 then the optimizer is not allowed to modify the number or order of
6630 execution of this ``load`` with other :ref:`volatile
6631 operations <volatile>`.
6633 If the ``load`` is marked as ``atomic``, it takes an extra
6634 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
6635 ``release`` and ``acq_rel`` orderings are not valid on ``load``
6636 instructions. Atomic loads produce :ref:`defined <memmodel>` results
6637 when they may see multiple atomic stores. The type of the pointee must
6638 be an integer type whose bit width is a power of two greater than or
6639 equal to eight and less than or equal to a target-specific size limit.
6640 ``align`` must be explicitly specified on atomic loads, and the load has
6641 undefined behavior if the alignment is not set to a value which is at
6642 least the size in bytes of the pointee. ``!nontemporal`` does not have
6643 any defined semantics for atomic loads.
6645 The optional constant ``align`` argument specifies the alignment of the
6646 operation (that is, the alignment of the memory address). A value of 0
6647 or an omitted ``align`` argument means that the operation has the ABI
6648 alignment for the target. It is the responsibility of the code emitter
6649 to ensure that the alignment information is correct. Overestimating the
6650 alignment results in undefined behavior. Underestimating the alignment
6651 may produce less efficient code. An alignment of 1 is always safe. The
6652 maximum possible alignment is ``1 << 29``.
6654 The optional ``!nontemporal`` metadata must reference a single
6655 metadata name ``<index>`` corresponding to a metadata node with one
6656 ``i32`` entry of value 1. The existence of the ``!nontemporal``
6657 metadata on the instruction tells the optimizer and code generator
6658 that this load is not expected to be reused in the cache. The code
6659 generator may select special instructions to save cache bandwidth, such
6660 as the ``MOVNT`` instruction on x86.
6662 The optional ``!invariant.load`` metadata must reference a single
6663 metadata name ``<index>`` corresponding to a metadata node with no
6664 entries. The existence of the ``!invariant.load`` metadata on the
6665 instruction tells the optimizer and code generator that the address
6666 operand to this load points to memory which can be assumed unchanged.
6667 Being invariant does not imply that a location is dereferenceable,
6668 but it does imply that once the location is known dereferenceable
6669 its value is henceforth unchanging.
6671 The optional ``!nonnull`` metadata must reference a single
6672 metadata name ``<index>`` corresponding to a metadata node with no
6673 entries. The existence of the ``!nonnull`` metadata on the
6674 instruction tells the optimizer that the value loaded is known to
6675 never be null. This is analogous to the ''nonnull'' attribute
6676 on parameters and return values. This metadata can only be applied
6677 to loads of a pointer type.
6679 The optional ``!dereferenceable`` metadata must reference a single
6680 metadata name ``<index>`` corresponding to a metadata node with one ``i64``
6681 entry. The existence of the ``!dereferenceable`` metadata on the instruction
6682 tells the optimizer that the value loaded is known to be dereferenceable.
6683 The number of bytes known to be dereferenceable is specified by the integer
6684 value in the metadata node. This is analogous to the ''dereferenceable''
6685 attribute on parameters and return values. This metadata can only be applied
6686 to loads of a pointer type.
6688 The optional ``!dereferenceable_or_null`` metadata must reference a single
6689 metadata name ``<index>`` corresponding to a metadata node with one ``i64``
6690 entry. The existence of the ``!dereferenceable_or_null`` metadata on the
6691 instruction tells the optimizer that the value loaded is known to be either
6692 dereferenceable or null.
6693 The number of bytes known to be dereferenceable is specified by the integer
6694 value in the metadata node. This is analogous to the ''dereferenceable_or_null''
6695 attribute on parameters and return values. This metadata can only be applied
6696 to loads of a pointer type.
6701 The location of memory pointed to is loaded. If the value being loaded
6702 is of scalar type then the number of bytes read does not exceed the
6703 minimum number of bytes needed to hold all bits of the type. For
6704 example, loading an ``i24`` reads at most three bytes. When loading a
6705 value of a type like ``i20`` with a size that is not an integral number
6706 of bytes, the result is undefined if the value was not originally
6707 written using a store of the same type.
6712 .. code-block:: llvm
6714 %ptr = alloca i32 ; yields i32*:ptr
6715 store i32 3, i32* %ptr ; yields void
6716 %val = load i32, i32* %ptr ; yields i32:val = i32 3
6720 '``store``' Instruction
6721 ^^^^^^^^^^^^^^^^^^^^^^^
6728 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields void
6729 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields void
6734 The '``store``' instruction is used to write to memory.
6739 There are two arguments to the ``store`` instruction: a value to store
6740 and an address at which to store it. The type of the ``<pointer>``
6741 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
6742 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
6743 then the optimizer is not allowed to modify the number or order of
6744 execution of this ``store`` with other :ref:`volatile
6745 operations <volatile>`.
6747 If the ``store`` is marked as ``atomic``, it takes an extra
6748 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
6749 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
6750 instructions. Atomic loads produce :ref:`defined <memmodel>` results
6751 when they may see multiple atomic stores. The type of the pointee must
6752 be an integer type whose bit width is a power of two greater than or
6753 equal to eight and less than or equal to a target-specific size limit.
6754 ``align`` must be explicitly specified on atomic stores, and the store
6755 has undefined behavior if the alignment is not set to a value which is
6756 at least the size in bytes of the pointee. ``!nontemporal`` does not
6757 have any defined semantics for atomic stores.
6759 The optional constant ``align`` argument specifies the alignment of the
6760 operation (that is, the alignment of the memory address). A value of 0
6761 or an omitted ``align`` argument means that the operation has the ABI
6762 alignment for the target. It is the responsibility of the code emitter
6763 to ensure that the alignment information is correct. Overestimating the
6764 alignment results in undefined behavior. Underestimating the
6765 alignment may produce less efficient code. An alignment of 1 is always
6766 safe. The maximum possible alignment is ``1 << 29``.
6768 The optional ``!nontemporal`` metadata must reference a single metadata
6769 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
6770 value 1. The existence of the ``!nontemporal`` metadata on the instruction
6771 tells the optimizer and code generator that this load is not expected to
6772 be reused in the cache. The code generator may select special
6773 instructions to save cache bandwidth, such as the MOVNT instruction on
6779 The contents of memory are updated to contain ``<value>`` at the
6780 location specified by the ``<pointer>`` operand. If ``<value>`` is
6781 of scalar type then the number of bytes written does not exceed the
6782 minimum number of bytes needed to hold all bits of the type. For
6783 example, storing an ``i24`` writes at most three bytes. When writing a
6784 value of a type like ``i20`` with a size that is not an integral number
6785 of bytes, it is unspecified what happens to the extra bits that do not
6786 belong to the type, but they will typically be overwritten.
6791 .. code-block:: llvm
6793 %ptr = alloca i32 ; yields i32*:ptr
6794 store i32 3, i32* %ptr ; yields void
6795 %val = load i32* %ptr ; yields i32:val = i32 3
6799 '``fence``' Instruction
6800 ^^^^^^^^^^^^^^^^^^^^^^^
6807 fence [singlethread] <ordering> ; yields void
6812 The '``fence``' instruction is used to introduce happens-before edges
6818 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
6819 defines what *synchronizes-with* edges they add. They can only be given
6820 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
6825 A fence A which has (at least) ``release`` ordering semantics
6826 *synchronizes with* a fence B with (at least) ``acquire`` ordering
6827 semantics if and only if there exist atomic operations X and Y, both
6828 operating on some atomic object M, such that A is sequenced before X, X
6829 modifies M (either directly or through some side effect of a sequence
6830 headed by X), Y is sequenced before B, and Y observes M. This provides a
6831 *happens-before* dependency between A and B. Rather than an explicit
6832 ``fence``, one (but not both) of the atomic operations X or Y might
6833 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
6834 still *synchronize-with* the explicit ``fence`` and establish the
6835 *happens-before* edge.
6837 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
6838 ``acquire`` and ``release`` semantics specified above, participates in
6839 the global program order of other ``seq_cst`` operations and/or fences.
6841 The optional ":ref:`singlethread <singlethread>`" argument specifies
6842 that the fence only synchronizes with other fences in the same thread.
6843 (This is useful for interacting with signal handlers.)
6848 .. code-block:: llvm
6850 fence acquire ; yields void
6851 fence singlethread seq_cst ; yields void
6855 '``cmpxchg``' Instruction
6856 ^^^^^^^^^^^^^^^^^^^^^^^^^
6863 cmpxchg [weak] [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <success ordering> <failure ordering> ; yields { ty, i1 }
6868 The '``cmpxchg``' instruction is used to atomically modify memory. It
6869 loads a value in memory and compares it to a given value. If they are
6870 equal, it tries to store a new value into the memory.
6875 There are three arguments to the '``cmpxchg``' instruction: an address
6876 to operate on, a value to compare to the value currently be at that
6877 address, and a new value to place at that address if the compared values
6878 are equal. The type of '<cmp>' must be an integer type whose bit width
6879 is a power of two greater than or equal to eight and less than or equal
6880 to a target-specific size limit. '<cmp>' and '<new>' must have the same
6881 type, and the type of '<pointer>' must be a pointer to that type. If the
6882 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
6883 to modify the number or order of execution of this ``cmpxchg`` with
6884 other :ref:`volatile operations <volatile>`.
6886 The success and failure :ref:`ordering <ordering>` arguments specify how this
6887 ``cmpxchg`` synchronizes with other atomic operations. Both ordering parameters
6888 must be at least ``monotonic``, the ordering constraint on failure must be no
6889 stronger than that on success, and the failure ordering cannot be either
6890 ``release`` or ``acq_rel``.
6892 The optional "``singlethread``" argument declares that the ``cmpxchg``
6893 is only atomic with respect to code (usually signal handlers) running in
6894 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
6895 respect to all other code in the system.
6897 The pointer passed into cmpxchg must have alignment greater than or
6898 equal to the size in memory of the operand.
6903 The contents of memory at the location specified by the '``<pointer>``' operand
6904 is read and compared to '``<cmp>``'; if the read value is the equal, the
6905 '``<new>``' is written. The original value at the location is returned, together
6906 with a flag indicating success (true) or failure (false).
6908 If the cmpxchg operation is marked as ``weak`` then a spurious failure is
6909 permitted: the operation may not write ``<new>`` even if the comparison
6912 If the cmpxchg operation is strong (the default), the i1 value is 1 if and only
6913 if the value loaded equals ``cmp``.
6915 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose of
6916 identifying release sequences. A failed ``cmpxchg`` is equivalent to an atomic
6917 load with an ordering parameter determined the second ordering parameter.
6922 .. code-block:: llvm
6925 %orig = atomic load i32, i32* %ptr unordered ; yields i32
6929 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
6930 %squared = mul i32 %cmp, %cmp
6931 %val_success = cmpxchg i32* %ptr, i32 %cmp, i32 %squared acq_rel monotonic ; yields { i32, i1 }
6932 %value_loaded = extractvalue { i32, i1 } %val_success, 0
6933 %success = extractvalue { i32, i1 } %val_success, 1
6934 br i1 %success, label %done, label %loop
6941 '``atomicrmw``' Instruction
6942 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6949 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields ty
6954 The '``atomicrmw``' instruction is used to atomically modify memory.
6959 There are three arguments to the '``atomicrmw``' instruction: an
6960 operation to apply, an address whose value to modify, an argument to the
6961 operation. The operation must be one of the following keywords:
6975 The type of '<value>' must be an integer type whose bit width is a power
6976 of two greater than or equal to eight and less than or equal to a
6977 target-specific size limit. The type of the '``<pointer>``' operand must
6978 be a pointer to that type. If the ``atomicrmw`` is marked as
6979 ``volatile``, then the optimizer is not allowed to modify the number or
6980 order of execution of this ``atomicrmw`` with other :ref:`volatile
6981 operations <volatile>`.
6986 The contents of memory at the location specified by the '``<pointer>``'
6987 operand are atomically read, modified, and written back. The original
6988 value at the location is returned. The modification is specified by the
6991 - xchg: ``*ptr = val``
6992 - add: ``*ptr = *ptr + val``
6993 - sub: ``*ptr = *ptr - val``
6994 - and: ``*ptr = *ptr & val``
6995 - nand: ``*ptr = ~(*ptr & val)``
6996 - or: ``*ptr = *ptr | val``
6997 - xor: ``*ptr = *ptr ^ val``
6998 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
6999 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
7000 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
7002 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
7008 .. code-block:: llvm
7010 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields i32
7012 .. _i_getelementptr:
7014 '``getelementptr``' Instruction
7015 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7022 <result> = getelementptr <ty>, <ty>* <ptrval>{, <ty> <idx>}*
7023 <result> = getelementptr inbounds <ty>, <ty>* <ptrval>{, <ty> <idx>}*
7024 <result> = getelementptr <ty>, <ptr vector> <ptrval>, <vector index type> <idx>
7029 The '``getelementptr``' instruction is used to get the address of a
7030 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
7031 address calculation only and does not access memory. The instruction can also
7032 be used to calculate a vector of such addresses.
7037 The first argument is always a type used as the basis for the calculations.
7038 The second argument is always a pointer or a vector of pointers, and is the
7039 base address to start from. The remaining arguments are indices
7040 that indicate which of the elements of the aggregate object are indexed.
7041 The interpretation of each index is dependent on the type being indexed
7042 into. The first index always indexes the pointer value given as the
7043 first argument, the second index indexes a value of the type pointed to
7044 (not necessarily the value directly pointed to, since the first index
7045 can be non-zero), etc. The first type indexed into must be a pointer
7046 value, subsequent types can be arrays, vectors, and structs. Note that
7047 subsequent types being indexed into can never be pointers, since that
7048 would require loading the pointer before continuing calculation.
7050 The type of each index argument depends on the type it is indexing into.
7051 When indexing into a (optionally packed) structure, only ``i32`` integer
7052 **constants** are allowed (when using a vector of indices they must all
7053 be the **same** ``i32`` integer constant). When indexing into an array,
7054 pointer or vector, integers of any width are allowed, and they are not
7055 required to be constant. These integers are treated as signed values
7058 For example, let's consider a C code fragment and how it gets compiled
7074 int *foo(struct ST *s) {
7075 return &s[1].Z.B[5][13];
7078 The LLVM code generated by Clang is:
7080 .. code-block:: llvm
7082 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
7083 %struct.ST = type { i32, double, %struct.RT }
7085 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
7087 %arrayidx = getelementptr inbounds %struct.ST, %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
7094 In the example above, the first index is indexing into the
7095 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
7096 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
7097 indexes into the third element of the structure, yielding a
7098 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
7099 structure. The third index indexes into the second element of the
7100 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
7101 dimensions of the array are subscripted into, yielding an '``i32``'
7102 type. The '``getelementptr``' instruction returns a pointer to this
7103 element, thus computing a value of '``i32*``' type.
7105 Note that it is perfectly legal to index partially through a structure,
7106 returning a pointer to an inner element. Because of this, the LLVM code
7107 for the given testcase is equivalent to:
7109 .. code-block:: llvm
7111 define i32* @foo(%struct.ST* %s) {
7112 %t1 = getelementptr %struct.ST, %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
7113 %t2 = getelementptr %struct.ST, %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
7114 %t3 = getelementptr %struct.RT, %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
7115 %t4 = getelementptr [10 x [20 x i32]], [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
7116 %t5 = getelementptr [20 x i32], [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
7120 If the ``inbounds`` keyword is present, the result value of the
7121 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
7122 pointer is not an *in bounds* address of an allocated object, or if any
7123 of the addresses that would be formed by successive addition of the
7124 offsets implied by the indices to the base address with infinitely
7125 precise signed arithmetic are not an *in bounds* address of that
7126 allocated object. The *in bounds* addresses for an allocated object are
7127 all the addresses that point into the object, plus the address one byte
7128 past the end. In cases where the base is a vector of pointers the
7129 ``inbounds`` keyword applies to each of the computations element-wise.
7131 If the ``inbounds`` keyword is not present, the offsets are added to the
7132 base address with silently-wrapping two's complement arithmetic. If the
7133 offsets have a different width from the pointer, they are sign-extended
7134 or truncated to the width of the pointer. The result value of the
7135 ``getelementptr`` may be outside the object pointed to by the base
7136 pointer. The result value may not necessarily be used to access memory
7137 though, even if it happens to point into allocated storage. See the
7138 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
7141 The getelementptr instruction is often confusing. For some more insight
7142 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
7147 .. code-block:: llvm
7149 ; yields [12 x i8]*:aptr
7150 %aptr = getelementptr {i32, [12 x i8]}, {i32, [12 x i8]}* %saptr, i64 0, i32 1
7152 %vptr = getelementptr {i32, <2 x i8>}, {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
7154 %eptr = getelementptr [12 x i8], [12 x i8]* %aptr, i64 0, i32 1
7156 %iptr = getelementptr [10 x i32], [10 x i32]* @arr, i16 0, i16 0
7161 The ``getelementptr`` returns a vector of pointers, instead of a single address,
7162 when one or more of its arguments is a vector. In such cases, all vector
7163 arguments should have the same number of elements, and every scalar argument
7164 will be effectively broadcast into a vector during address calculation.
7166 .. code-block:: llvm
7168 ; All arguments are vectors:
7169 ; A[i] = ptrs[i] + offsets[i]*sizeof(i8)
7170 %A = getelementptr i8, <4 x i8*> %ptrs, <4 x i64> %offsets
7172 ; Add the same scalar offset to each pointer of a vector:
7173 ; A[i] = ptrs[i] + offset*sizeof(i8)
7174 %A = getelementptr i8, <4 x i8*> %ptrs, i64 %offset
7176 ; Add distinct offsets to the same pointer:
7177 ; A[i] = ptr + offsets[i]*sizeof(i8)
7178 %A = getelementptr i8, i8* %ptr, <4 x i64> %offsets
7180 ; In all cases described above the type of the result is <4 x i8*>
7182 The two following instructions are equivalent:
7184 .. code-block:: llvm
7186 getelementptr %struct.ST, <4 x %struct.ST*> %s, <4 x i64> %ind1,
7187 <4 x i32> <i32 2, i32 2, i32 2, i32 2>,
7188 <4 x i32> <i32 1, i32 1, i32 1, i32 1>,
7190 <4 x i64> <i64 13, i64 13, i64 13, i64 13>
7192 getelementptr %struct.ST, <4 x %struct.ST*> %s, <4 x i64> %ind1,
7193 i32 2, i32 1, <4 x i32> %ind4, i64 13
7195 Let's look at the C code, where the vector version of ``getelementptr``
7200 // Let's assume that we vectorize the following loop:
7201 double *A, B; int *C;
7202 for (int i = 0; i < size; ++i) {
7206 .. code-block:: llvm
7208 ; get pointers for 8 elements from array B
7209 %ptrs = getelementptr double, double* %B, <8 x i32> %C
7210 ; load 8 elements from array B into A
7211 %A = call <8 x double> @llvm.masked.gather.v8f64(<8 x double*> %ptrs,
7212 i32 8, <8 x i1> %mask, <8 x double> %passthru)
7214 Conversion Operations
7215 ---------------------
7217 The instructions in this category are the conversion instructions
7218 (casting) which all take a single operand and a type. They perform
7219 various bit conversions on the operand.
7221 '``trunc .. to``' Instruction
7222 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7229 <result> = trunc <ty> <value> to <ty2> ; yields ty2
7234 The '``trunc``' instruction truncates its operand to the type ``ty2``.
7239 The '``trunc``' instruction takes a value to trunc, and a type to trunc
7240 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
7241 of the same number of integers. The bit size of the ``value`` must be
7242 larger than the bit size of the destination type, ``ty2``. Equal sized
7243 types are not allowed.
7248 The '``trunc``' instruction truncates the high order bits in ``value``
7249 and converts the remaining bits to ``ty2``. Since the source size must
7250 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
7251 It will always truncate bits.
7256 .. code-block:: llvm
7258 %X = trunc i32 257 to i8 ; yields i8:1
7259 %Y = trunc i32 123 to i1 ; yields i1:true
7260 %Z = trunc i32 122 to i1 ; yields i1:false
7261 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
7263 '``zext .. to``' Instruction
7264 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7271 <result> = zext <ty> <value> to <ty2> ; yields ty2
7276 The '``zext``' instruction zero extends its operand to type ``ty2``.
7281 The '``zext``' instruction takes a value to cast, and a type to cast it
7282 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
7283 the same number of integers. The bit size of the ``value`` must be
7284 smaller than the bit size of the destination type, ``ty2``.
7289 The ``zext`` fills the high order bits of the ``value`` with zero bits
7290 until it reaches the size of the destination type, ``ty2``.
7292 When zero extending from i1, the result will always be either 0 or 1.
7297 .. code-block:: llvm
7299 %X = zext i32 257 to i64 ; yields i64:257
7300 %Y = zext i1 true to i32 ; yields i32:1
7301 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
7303 '``sext .. to``' Instruction
7304 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7311 <result> = sext <ty> <value> to <ty2> ; yields ty2
7316 The '``sext``' sign extends ``value`` to the type ``ty2``.
7321 The '``sext``' instruction takes a value to cast, and a type to cast it
7322 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
7323 the same number of integers. The bit size of the ``value`` must be
7324 smaller than the bit size of the destination type, ``ty2``.
7329 The '``sext``' instruction performs a sign extension by copying the sign
7330 bit (highest order bit) of the ``value`` until it reaches the bit size
7331 of the type ``ty2``.
7333 When sign extending from i1, the extension always results in -1 or 0.
7338 .. code-block:: llvm
7340 %X = sext i8 -1 to i16 ; yields i16 :65535
7341 %Y = sext i1 true to i32 ; yields i32:-1
7342 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
7344 '``fptrunc .. to``' Instruction
7345 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7352 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
7357 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
7362 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
7363 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
7364 The size of ``value`` must be larger than the size of ``ty2``. This
7365 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
7370 The '``fptrunc``' instruction truncates a ``value`` from a larger
7371 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
7372 point <t_floating>` type. If the value cannot fit within the
7373 destination type, ``ty2``, then the results are undefined.
7378 .. code-block:: llvm
7380 %X = fptrunc double 123.0 to float ; yields float:123.0
7381 %Y = fptrunc double 1.0E+300 to float ; yields undefined
7383 '``fpext .. to``' Instruction
7384 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7391 <result> = fpext <ty> <value> to <ty2> ; yields ty2
7396 The '``fpext``' extends a floating point ``value`` to a larger floating
7402 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
7403 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
7404 to. The source type must be smaller than the destination type.
7409 The '``fpext``' instruction extends the ``value`` from a smaller
7410 :ref:`floating point <t_floating>` type to a larger :ref:`floating
7411 point <t_floating>` type. The ``fpext`` cannot be used to make a
7412 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
7413 *no-op cast* for a floating point cast.
7418 .. code-block:: llvm
7420 %X = fpext float 3.125 to double ; yields double:3.125000e+00
7421 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
7423 '``fptoui .. to``' Instruction
7424 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7431 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
7436 The '``fptoui``' converts a floating point ``value`` to its unsigned
7437 integer equivalent of type ``ty2``.
7442 The '``fptoui``' instruction takes a value to cast, which must be a
7443 scalar or vector :ref:`floating point <t_floating>` value, and a type to
7444 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
7445 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
7446 type with the same number of elements as ``ty``
7451 The '``fptoui``' instruction converts its :ref:`floating
7452 point <t_floating>` operand into the nearest (rounding towards zero)
7453 unsigned integer value. If the value cannot fit in ``ty2``, the results
7459 .. code-block:: llvm
7461 %X = fptoui double 123.0 to i32 ; yields i32:123
7462 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
7463 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
7465 '``fptosi .. to``' Instruction
7466 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7473 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
7478 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
7479 ``value`` to type ``ty2``.
7484 The '``fptosi``' instruction takes a value to cast, which must be a
7485 scalar or vector :ref:`floating point <t_floating>` value, and a type to
7486 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
7487 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
7488 type with the same number of elements as ``ty``
7493 The '``fptosi``' instruction converts its :ref:`floating
7494 point <t_floating>` operand into the nearest (rounding towards zero)
7495 signed integer value. If the value cannot fit in ``ty2``, the results
7501 .. code-block:: llvm
7503 %X = fptosi double -123.0 to i32 ; yields i32:-123
7504 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
7505 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
7507 '``uitofp .. to``' Instruction
7508 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7515 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
7520 The '``uitofp``' instruction regards ``value`` as an unsigned integer
7521 and converts that value to the ``ty2`` type.
7526 The '``uitofp``' instruction takes a value to cast, which must be a
7527 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
7528 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
7529 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
7530 type with the same number of elements as ``ty``
7535 The '``uitofp``' instruction interprets its operand as an unsigned
7536 integer quantity and converts it to the corresponding floating point
7537 value. If the value cannot fit in the floating point value, the results
7543 .. code-block:: llvm
7545 %X = uitofp i32 257 to float ; yields float:257.0
7546 %Y = uitofp i8 -1 to double ; yields double:255.0
7548 '``sitofp .. to``' Instruction
7549 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7556 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
7561 The '``sitofp``' instruction regards ``value`` as a signed integer and
7562 converts that value to the ``ty2`` type.
7567 The '``sitofp``' instruction takes a value to cast, which must be a
7568 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
7569 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
7570 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
7571 type with the same number of elements as ``ty``
7576 The '``sitofp``' instruction interprets its operand as a signed integer
7577 quantity and converts it to the corresponding floating point value. If
7578 the value cannot fit in the floating point value, the results are
7584 .. code-block:: llvm
7586 %X = sitofp i32 257 to float ; yields float:257.0
7587 %Y = sitofp i8 -1 to double ; yields double:-1.0
7591 '``ptrtoint .. to``' Instruction
7592 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7599 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
7604 The '``ptrtoint``' instruction converts the pointer or a vector of
7605 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
7610 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
7611 a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
7612 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
7613 a vector of integers type.
7618 The '``ptrtoint``' instruction converts ``value`` to integer type
7619 ``ty2`` by interpreting the pointer value as an integer and either
7620 truncating or zero extending that value to the size of the integer type.
7621 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
7622 ``value`` is larger than ``ty2`` then a truncation is done. If they are
7623 the same size, then nothing is done (*no-op cast*) other than a type
7629 .. code-block:: llvm
7631 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
7632 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
7633 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
7637 '``inttoptr .. to``' Instruction
7638 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7645 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
7650 The '``inttoptr``' instruction converts an integer ``value`` to a
7651 pointer type, ``ty2``.
7656 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
7657 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
7663 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
7664 applying either a zero extension or a truncation depending on the size
7665 of the integer ``value``. If ``value`` is larger than the size of a
7666 pointer then a truncation is done. If ``value`` is smaller than the size
7667 of a pointer then a zero extension is done. If they are the same size,
7668 nothing is done (*no-op cast*).
7673 .. code-block:: llvm
7675 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
7676 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
7677 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
7678 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
7682 '``bitcast .. to``' Instruction
7683 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7690 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
7695 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
7701 The '``bitcast``' instruction takes a value to cast, which must be a
7702 non-aggregate first class value, and a type to cast it to, which must
7703 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
7704 bit sizes of ``value`` and the destination type, ``ty2``, must be
7705 identical. If the source type is a pointer, the destination type must
7706 also be a pointer of the same size. This instruction supports bitwise
7707 conversion of vectors to integers and to vectors of other types (as
7708 long as they have the same size).
7713 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
7714 is always a *no-op cast* because no bits change with this
7715 conversion. The conversion is done as if the ``value`` had been stored
7716 to memory and read back as type ``ty2``. Pointer (or vector of
7717 pointers) types may only be converted to other pointer (or vector of
7718 pointers) types with the same address space through this instruction.
7719 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
7720 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
7725 .. code-block:: llvm
7727 %X = bitcast i8 255 to i8 ; yields i8 :-1
7728 %Y = bitcast i32* %x to sint* ; yields sint*:%x
7729 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
7730 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
7732 .. _i_addrspacecast:
7734 '``addrspacecast .. to``' Instruction
7735 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7742 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
7747 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
7748 address space ``n`` to type ``pty2`` in address space ``m``.
7753 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
7754 to cast and a pointer type to cast it to, which must have a different
7760 The '``addrspacecast``' instruction converts the pointer value
7761 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
7762 value modification, depending on the target and the address space
7763 pair. Pointer conversions within the same address space must be
7764 performed with the ``bitcast`` instruction. Note that if the address space
7765 conversion is legal then both result and operand refer to the same memory
7771 .. code-block:: llvm
7773 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
7774 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
7775 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
7782 The instructions in this category are the "miscellaneous" instructions,
7783 which defy better classification.
7787 '``icmp``' Instruction
7788 ^^^^^^^^^^^^^^^^^^^^^^
7795 <result> = icmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
7800 The '``icmp``' instruction returns a boolean value or a vector of
7801 boolean values based on comparison of its two integer, integer vector,
7802 pointer, or pointer vector operands.
7807 The '``icmp``' instruction takes three operands. The first operand is
7808 the condition code indicating the kind of comparison to perform. It is
7809 not a value, just a keyword. The possible condition code are:
7812 #. ``ne``: not equal
7813 #. ``ugt``: unsigned greater than
7814 #. ``uge``: unsigned greater or equal
7815 #. ``ult``: unsigned less than
7816 #. ``ule``: unsigned less or equal
7817 #. ``sgt``: signed greater than
7818 #. ``sge``: signed greater or equal
7819 #. ``slt``: signed less than
7820 #. ``sle``: signed less or equal
7822 The remaining two arguments must be :ref:`integer <t_integer>` or
7823 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
7824 must also be identical types.
7829 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
7830 code given as ``cond``. The comparison performed always yields either an
7831 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
7833 #. ``eq``: yields ``true`` if the operands are equal, ``false``
7834 otherwise. No sign interpretation is necessary or performed.
7835 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
7836 otherwise. No sign interpretation is necessary or performed.
7837 #. ``ugt``: interprets the operands as unsigned values and yields
7838 ``true`` if ``op1`` is greater than ``op2``.
7839 #. ``uge``: interprets the operands as unsigned values and yields
7840 ``true`` if ``op1`` is greater than or equal to ``op2``.
7841 #. ``ult``: interprets the operands as unsigned values and yields
7842 ``true`` if ``op1`` is less than ``op2``.
7843 #. ``ule``: interprets the operands as unsigned values and yields
7844 ``true`` if ``op1`` is less than or equal to ``op2``.
7845 #. ``sgt``: interprets the operands as signed values and yields ``true``
7846 if ``op1`` is greater than ``op2``.
7847 #. ``sge``: interprets the operands as signed values and yields ``true``
7848 if ``op1`` is greater than or equal to ``op2``.
7849 #. ``slt``: interprets the operands as signed values and yields ``true``
7850 if ``op1`` is less than ``op2``.
7851 #. ``sle``: interprets the operands as signed values and yields ``true``
7852 if ``op1`` is less than or equal to ``op2``.
7854 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
7855 are compared as if they were integers.
7857 If the operands are integer vectors, then they are compared element by
7858 element. The result is an ``i1`` vector with the same number of elements
7859 as the values being compared. Otherwise, the result is an ``i1``.
7864 .. code-block:: llvm
7866 <result> = icmp eq i32 4, 5 ; yields: result=false
7867 <result> = icmp ne float* %X, %X ; yields: result=false
7868 <result> = icmp ult i16 4, 5 ; yields: result=true
7869 <result> = icmp sgt i16 4, 5 ; yields: result=false
7870 <result> = icmp ule i16 -4, 5 ; yields: result=false
7871 <result> = icmp sge i16 4, 5 ; yields: result=false
7873 Note that the code generator does not yet support vector types with the
7874 ``icmp`` instruction.
7878 '``fcmp``' Instruction
7879 ^^^^^^^^^^^^^^^^^^^^^^
7886 <result> = fcmp [fast-math flags]* <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
7891 The '``fcmp``' instruction returns a boolean value or vector of boolean
7892 values based on comparison of its operands.
7894 If the operands are floating point scalars, then the result type is a
7895 boolean (:ref:`i1 <t_integer>`).
7897 If the operands are floating point vectors, then the result type is a
7898 vector of boolean with the same number of elements as the operands being
7904 The '``fcmp``' instruction takes three operands. The first operand is
7905 the condition code indicating the kind of comparison to perform. It is
7906 not a value, just a keyword. The possible condition code are:
7908 #. ``false``: no comparison, always returns false
7909 #. ``oeq``: ordered and equal
7910 #. ``ogt``: ordered and greater than
7911 #. ``oge``: ordered and greater than or equal
7912 #. ``olt``: ordered and less than
7913 #. ``ole``: ordered and less than or equal
7914 #. ``one``: ordered and not equal
7915 #. ``ord``: ordered (no nans)
7916 #. ``ueq``: unordered or equal
7917 #. ``ugt``: unordered or greater than
7918 #. ``uge``: unordered or greater than or equal
7919 #. ``ult``: unordered or less than
7920 #. ``ule``: unordered or less than or equal
7921 #. ``une``: unordered or not equal
7922 #. ``uno``: unordered (either nans)
7923 #. ``true``: no comparison, always returns true
7925 *Ordered* means that neither operand is a QNAN while *unordered* means
7926 that either operand may be a QNAN.
7928 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
7929 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
7930 type. They must have identical types.
7935 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
7936 condition code given as ``cond``. If the operands are vectors, then the
7937 vectors are compared element by element. Each comparison performed
7938 always yields an :ref:`i1 <t_integer>` result, as follows:
7940 #. ``false``: always yields ``false``, regardless of operands.
7941 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
7942 is equal to ``op2``.
7943 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
7944 is greater than ``op2``.
7945 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
7946 is greater than or equal to ``op2``.
7947 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
7948 is less than ``op2``.
7949 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
7950 is less than or equal to ``op2``.
7951 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
7952 is not equal to ``op2``.
7953 #. ``ord``: yields ``true`` if both operands are not a QNAN.
7954 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
7956 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
7957 greater than ``op2``.
7958 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
7959 greater than or equal to ``op2``.
7960 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
7962 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
7963 less than or equal to ``op2``.
7964 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
7965 not equal to ``op2``.
7966 #. ``uno``: yields ``true`` if either operand is a QNAN.
7967 #. ``true``: always yields ``true``, regardless of operands.
7969 The ``fcmp`` instruction can also optionally take any number of
7970 :ref:`fast-math flags <fastmath>`, which are optimization hints to enable
7971 otherwise unsafe floating point optimizations.
7973 Any set of fast-math flags are legal on an ``fcmp`` instruction, but the
7974 only flags that have any effect on its semantics are those that allow
7975 assumptions to be made about the values of input arguments; namely
7976 ``nnan``, ``ninf``, and ``nsz``. See :ref:`fastmath` for more information.
7981 .. code-block:: llvm
7983 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
7984 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
7985 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
7986 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
7988 Note that the code generator does not yet support vector types with the
7989 ``fcmp`` instruction.
7993 '``phi``' Instruction
7994 ^^^^^^^^^^^^^^^^^^^^^
8001 <result> = phi <ty> [ <val0>, <label0>], ...
8006 The '``phi``' instruction is used to implement the φ node in the SSA
8007 graph representing the function.
8012 The type of the incoming values is specified with the first type field.
8013 After this, the '``phi``' instruction takes a list of pairs as
8014 arguments, with one pair for each predecessor basic block of the current
8015 block. Only values of :ref:`first class <t_firstclass>` type may be used as
8016 the value arguments to the PHI node. Only labels may be used as the
8019 There must be no non-phi instructions between the start of a basic block
8020 and the PHI instructions: i.e. PHI instructions must be first in a basic
8023 For the purposes of the SSA form, the use of each incoming value is
8024 deemed to occur on the edge from the corresponding predecessor block to
8025 the current block (but after any definition of an '``invoke``'
8026 instruction's return value on the same edge).
8031 At runtime, the '``phi``' instruction logically takes on the value
8032 specified by the pair corresponding to the predecessor basic block that
8033 executed just prior to the current block.
8038 .. code-block:: llvm
8040 Loop: ; Infinite loop that counts from 0 on up...
8041 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
8042 %nextindvar = add i32 %indvar, 1
8047 '``select``' Instruction
8048 ^^^^^^^^^^^^^^^^^^^^^^^^
8055 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
8057 selty is either i1 or {<N x i1>}
8062 The '``select``' instruction is used to choose one value based on a
8063 condition, without IR-level branching.
8068 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
8069 values indicating the condition, and two values of the same :ref:`first
8070 class <t_firstclass>` type.
8075 If the condition is an i1 and it evaluates to 1, the instruction returns
8076 the first value argument; otherwise, it returns the second value
8079 If the condition is a vector of i1, then the value arguments must be
8080 vectors of the same size, and the selection is done element by element.
8082 If the condition is an i1 and the value arguments are vectors of the
8083 same size, then an entire vector is selected.
8088 .. code-block:: llvm
8090 %X = select i1 true, i8 17, i8 42 ; yields i8:17
8094 '``call``' Instruction
8095 ^^^^^^^^^^^^^^^^^^^^^^
8102 <result> = [tail | musttail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
8107 The '``call``' instruction represents a simple function call.
8112 This instruction requires several arguments:
8114 #. The optional ``tail`` and ``musttail`` markers indicate that the optimizers
8115 should perform tail call optimization. The ``tail`` marker is a hint that
8116 `can be ignored <CodeGenerator.html#sibcallopt>`_. The ``musttail`` marker
8117 means that the call must be tail call optimized in order for the program to
8118 be correct. The ``musttail`` marker provides these guarantees:
8120 #. The call will not cause unbounded stack growth if it is part of a
8121 recursive cycle in the call graph.
8122 #. Arguments with the :ref:`inalloca <attr_inalloca>` attribute are
8125 Both markers imply that the callee does not access allocas or varargs from
8126 the caller. Calls marked ``musttail`` must obey the following additional
8129 - The call must immediately precede a :ref:`ret <i_ret>` instruction,
8130 or a pointer bitcast followed by a ret instruction.
8131 - The ret instruction must return the (possibly bitcasted) value
8132 produced by the call or void.
8133 - The caller and callee prototypes must match. Pointer types of
8134 parameters or return types may differ in pointee type, but not
8136 - The calling conventions of the caller and callee must match.
8137 - All ABI-impacting function attributes, such as sret, byval, inreg,
8138 returned, and inalloca, must match.
8139 - The callee must be varargs iff the caller is varargs. Bitcasting a
8140 non-varargs function to the appropriate varargs type is legal so
8141 long as the non-varargs prefixes obey the other rules.
8143 Tail call optimization for calls marked ``tail`` is guaranteed to occur if
8144 the following conditions are met:
8146 - Caller and callee both have the calling convention ``fastcc``.
8147 - The call is in tail position (ret immediately follows call and ret
8148 uses value of call or is void).
8149 - Option ``-tailcallopt`` is enabled, or
8150 ``llvm::GuaranteedTailCallOpt`` is ``true``.
8151 - `Platform-specific constraints are
8152 met. <CodeGenerator.html#tailcallopt>`_
8154 #. The optional "cconv" marker indicates which :ref:`calling
8155 convention <callingconv>` the call should use. If none is
8156 specified, the call defaults to using C calling conventions. The
8157 calling convention of the call must match the calling convention of
8158 the target function, or else the behavior is undefined.
8159 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
8160 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
8162 #. '``ty``': the type of the call instruction itself which is also the
8163 type of the return value. Functions that return no value are marked
8165 #. '``fnty``': shall be the signature of the pointer to function value
8166 being invoked. The argument types must match the types implied by
8167 this signature. This type can be omitted if the function is not
8168 varargs and if the function type does not return a pointer to a
8170 #. '``fnptrval``': An LLVM value containing a pointer to a function to
8171 be invoked. In most cases, this is a direct function invocation, but
8172 indirect ``call``'s are just as possible, calling an arbitrary pointer
8174 #. '``function args``': argument list whose types match the function
8175 signature argument types and parameter attributes. All arguments must
8176 be of :ref:`first class <t_firstclass>` type. If the function signature
8177 indicates the function accepts a variable number of arguments, the
8178 extra arguments can be specified.
8179 #. The optional :ref:`function attributes <fnattrs>` list. Only
8180 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
8181 attributes are valid here.
8186 The '``call``' instruction is used to cause control flow to transfer to
8187 a specified function, with its incoming arguments bound to the specified
8188 values. Upon a '``ret``' instruction in the called function, control
8189 flow continues with the instruction after the function call, and the
8190 return value of the function is bound to the result argument.
8195 .. code-block:: llvm
8197 %retval = call i32 @test(i32 %argc)
8198 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
8199 %X = tail call i32 @foo() ; yields i32
8200 %Y = tail call fastcc i32 @foo() ; yields i32
8201 call void %foo(i8 97 signext)
8203 %struct.A = type { i32, i8 }
8204 %r = call %struct.A @foo() ; yields { i32, i8 }
8205 %gr = extractvalue %struct.A %r, 0 ; yields i32
8206 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
8207 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
8208 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
8210 llvm treats calls to some functions with names and arguments that match
8211 the standard C99 library as being the C99 library functions, and may
8212 perform optimizations or generate code for them under that assumption.
8213 This is something we'd like to change in the future to provide better
8214 support for freestanding environments and non-C-based languages.
8218 '``va_arg``' Instruction
8219 ^^^^^^^^^^^^^^^^^^^^^^^^
8226 <resultval> = va_arg <va_list*> <arglist>, <argty>
8231 The '``va_arg``' instruction is used to access arguments passed through
8232 the "variable argument" area of a function call. It is used to implement
8233 the ``va_arg`` macro in C.
8238 This instruction takes a ``va_list*`` value and the type of the
8239 argument. It returns a value of the specified argument type and
8240 increments the ``va_list`` to point to the next argument. The actual
8241 type of ``va_list`` is target specific.
8246 The '``va_arg``' instruction loads an argument of the specified type
8247 from the specified ``va_list`` and causes the ``va_list`` to point to
8248 the next argument. For more information, see the variable argument
8249 handling :ref:`Intrinsic Functions <int_varargs>`.
8251 It is legal for this instruction to be called in a function which does
8252 not take a variable number of arguments, for example, the ``vfprintf``
8255 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
8256 function <intrinsics>` because it takes a type as an argument.
8261 See the :ref:`variable argument processing <int_varargs>` section.
8263 Note that the code generator does not yet fully support va\_arg on many
8264 targets. Also, it does not currently support va\_arg with aggregate
8265 types on any target.
8269 '``landingpad``' Instruction
8270 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8277 <resultval> = landingpad <resultty> <clause>+
8278 <resultval> = landingpad <resultty> cleanup <clause>*
8280 <clause> := catch <type> <value>
8281 <clause> := filter <array constant type> <array constant>
8286 The '``landingpad``' instruction is used by `LLVM's exception handling
8287 system <ExceptionHandling.html#overview>`_ to specify that a basic block
8288 is a landing pad --- one where the exception lands, and corresponds to the
8289 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
8290 defines values supplied by the :ref:`personality function <personalityfn>` upon
8291 re-entry to the function. The ``resultval`` has the type ``resultty``.
8297 ``cleanup`` flag indicates that the landing pad block is a cleanup.
8299 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
8300 contains the global variable representing the "type" that may be caught
8301 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
8302 clause takes an array constant as its argument. Use
8303 "``[0 x i8**] undef``" for a filter which cannot throw. The
8304 '``landingpad``' instruction must contain *at least* one ``clause`` or
8305 the ``cleanup`` flag.
8310 The '``landingpad``' instruction defines the values which are set by the
8311 :ref:`personality function <personalityfn>` upon re-entry to the function, and
8312 therefore the "result type" of the ``landingpad`` instruction. As with
8313 calling conventions, how the personality function results are
8314 represented in LLVM IR is target specific.
8316 The clauses are applied in order from top to bottom. If two
8317 ``landingpad`` instructions are merged together through inlining, the
8318 clauses from the calling function are appended to the list of clauses.
8319 When the call stack is being unwound due to an exception being thrown,
8320 the exception is compared against each ``clause`` in turn. If it doesn't
8321 match any of the clauses, and the ``cleanup`` flag is not set, then
8322 unwinding continues further up the call stack.
8324 The ``landingpad`` instruction has several restrictions:
8326 - A landing pad block is a basic block which is the unwind destination
8327 of an '``invoke``' instruction.
8328 - A landing pad block must have a '``landingpad``' instruction as its
8329 first non-PHI instruction.
8330 - There can be only one '``landingpad``' instruction within the landing
8332 - A basic block that is not a landing pad block may not include a
8333 '``landingpad``' instruction.
8338 .. code-block:: llvm
8340 ;; A landing pad which can catch an integer.
8341 %res = landingpad { i8*, i32 }
8343 ;; A landing pad that is a cleanup.
8344 %res = landingpad { i8*, i32 }
8346 ;; A landing pad which can catch an integer and can only throw a double.
8347 %res = landingpad { i8*, i32 }
8349 filter [1 x i8**] [@_ZTId]
8353 '``cleanuppad``' Instruction
8354 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8361 <resultval> = cleanuppad <resultty> [<args>*]
8366 The '``cleanuppad``' instruction is used by `LLVM's exception handling
8367 system <ExceptionHandling.html#overview>`_ to specify that a basic block
8368 is a cleanup block --- one where a personality routine attempts to
8369 transfer control to run cleanup actions.
8370 The ``args`` correspond to whatever additional
8371 information the :ref:`personality function <personalityfn>` requires to
8372 execute the cleanup.
8373 The ``resultval`` has the type ``resultty``.
8378 The instruction takes a list of arbitrary values which are interpreted
8379 by the :ref:`personality function <personalityfn>`.
8384 The '``cleanuppad``' instruction defines the values which are set by the
8385 :ref:`personality function <personalityfn>` upon re-entry to the function, and
8386 therefore the "result type" of the ``cleanuppad`` instruction. As with
8387 calling conventions, how the personality function results are
8388 represented in LLVM IR is target specific.
8390 When the call stack is being unwound due to an exception being thrown,
8391 the :ref:`personality function <personalityfn>` transfers control to the
8392 ``cleanuppad`` with the aid of the personality-specific arguments.
8394 The ``cleanuppad`` instruction has several restrictions:
8396 - A cleanup block is a basic block which is the unwind destination of
8397 an exceptional instruction.
8398 - A cleanup block must have a '``cleanuppad``' instruction as its
8399 first non-PHI instruction.
8400 - There can be only one '``cleanuppad``' instruction within the
8402 - A basic block that is not a cleanup block may not include a
8403 '``cleanuppad``' instruction.
8404 - It is undefined behavior for control to transfer from a ``cleanuppad`` to a
8405 ``catchret`` without first executing a ``cleanupret`` and a subsequent
8407 - It is undefined behavior for control to transfer from a ``cleanuppad`` to a
8408 ``ret`` without first executing a ``cleanupret``.
8413 .. code-block:: llvm
8415 %res = cleanuppad { i8*, i32 } [label %nextaction]
8422 LLVM supports the notion of an "intrinsic function". These functions
8423 have well known names and semantics and are required to follow certain
8424 restrictions. Overall, these intrinsics represent an extension mechanism
8425 for the LLVM language that does not require changing all of the
8426 transformations in LLVM when adding to the language (or the bitcode
8427 reader/writer, the parser, etc...).
8429 Intrinsic function names must all start with an "``llvm.``" prefix. This
8430 prefix is reserved in LLVM for intrinsic names; thus, function names may
8431 not begin with this prefix. Intrinsic functions must always be external
8432 functions: you cannot define the body of intrinsic functions. Intrinsic
8433 functions may only be used in call or invoke instructions: it is illegal
8434 to take the address of an intrinsic function. Additionally, because
8435 intrinsic functions are part of the LLVM language, it is required if any
8436 are added that they be documented here.
8438 Some intrinsic functions can be overloaded, i.e., the intrinsic
8439 represents a family of functions that perform the same operation but on
8440 different data types. Because LLVM can represent over 8 million
8441 different integer types, overloading is used commonly to allow an
8442 intrinsic function to operate on any integer type. One or more of the
8443 argument types or the result type can be overloaded to accept any
8444 integer type. Argument types may also be defined as exactly matching a
8445 previous argument's type or the result type. This allows an intrinsic
8446 function which accepts multiple arguments, but needs all of them to be
8447 of the same type, to only be overloaded with respect to a single
8448 argument or the result.
8450 Overloaded intrinsics will have the names of its overloaded argument
8451 types encoded into its function name, each preceded by a period. Only
8452 those types which are overloaded result in a name suffix. Arguments
8453 whose type is matched against another type do not. For example, the
8454 ``llvm.ctpop`` function can take an integer of any width and returns an
8455 integer of exactly the same integer width. This leads to a family of
8456 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
8457 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
8458 overloaded, and only one type suffix is required. Because the argument's
8459 type is matched against the return type, it does not require its own
8462 To learn how to add an intrinsic function, please see the `Extending
8463 LLVM Guide <ExtendingLLVM.html>`_.
8467 Variable Argument Handling Intrinsics
8468 -------------------------------------
8470 Variable argument support is defined in LLVM with the
8471 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
8472 functions. These functions are related to the similarly named macros
8473 defined in the ``<stdarg.h>`` header file.
8475 All of these functions operate on arguments that use a target-specific
8476 value type "``va_list``". The LLVM assembly language reference manual
8477 does not define what this type is, so all transformations should be
8478 prepared to handle these functions regardless of the type used.
8480 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
8481 variable argument handling intrinsic functions are used.
8483 .. code-block:: llvm
8485 ; This struct is different for every platform. For most platforms,
8486 ; it is merely an i8*.
8487 %struct.va_list = type { i8* }
8489 ; For Unix x86_64 platforms, va_list is the following struct:
8490 ; %struct.va_list = type { i32, i32, i8*, i8* }
8492 define i32 @test(i32 %X, ...) {
8493 ; Initialize variable argument processing
8494 %ap = alloca %struct.va_list
8495 %ap2 = bitcast %struct.va_list* %ap to i8*
8496 call void @llvm.va_start(i8* %ap2)
8498 ; Read a single integer argument
8499 %tmp = va_arg i8* %ap2, i32
8501 ; Demonstrate usage of llvm.va_copy and llvm.va_end
8503 %aq2 = bitcast i8** %aq to i8*
8504 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
8505 call void @llvm.va_end(i8* %aq2)
8507 ; Stop processing of arguments.
8508 call void @llvm.va_end(i8* %ap2)
8512 declare void @llvm.va_start(i8*)
8513 declare void @llvm.va_copy(i8*, i8*)
8514 declare void @llvm.va_end(i8*)
8518 '``llvm.va_start``' Intrinsic
8519 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8526 declare void @llvm.va_start(i8* <arglist>)
8531 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
8532 subsequent use by ``va_arg``.
8537 The argument is a pointer to a ``va_list`` element to initialize.
8542 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
8543 available in C. In a target-dependent way, it initializes the
8544 ``va_list`` element to which the argument points, so that the next call
8545 to ``va_arg`` will produce the first variable argument passed to the
8546 function. Unlike the C ``va_start`` macro, this intrinsic does not need
8547 to know the last argument of the function as the compiler can figure
8550 '``llvm.va_end``' Intrinsic
8551 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8558 declare void @llvm.va_end(i8* <arglist>)
8563 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
8564 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
8569 The argument is a pointer to a ``va_list`` to destroy.
8574 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
8575 available in C. In a target-dependent way, it destroys the ``va_list``
8576 element to which the argument points. Calls to
8577 :ref:`llvm.va_start <int_va_start>` and
8578 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
8583 '``llvm.va_copy``' Intrinsic
8584 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8591 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
8596 The '``llvm.va_copy``' intrinsic copies the current argument position
8597 from the source argument list to the destination argument list.
8602 The first argument is a pointer to a ``va_list`` element to initialize.
8603 The second argument is a pointer to a ``va_list`` element to copy from.
8608 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
8609 available in C. In a target-dependent way, it copies the source
8610 ``va_list`` element into the destination ``va_list`` element. This
8611 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
8612 arbitrarily complex and require, for example, memory allocation.
8614 Accurate Garbage Collection Intrinsics
8615 --------------------------------------
8617 LLVM's support for `Accurate Garbage Collection <GarbageCollection.html>`_
8618 (GC) requires the frontend to generate code containing appropriate intrinsic
8619 calls and select an appropriate GC strategy which knows how to lower these
8620 intrinsics in a manner which is appropriate for the target collector.
8622 These intrinsics allow identification of :ref:`GC roots on the
8623 stack <int_gcroot>`, as well as garbage collector implementations that
8624 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
8625 Frontends for type-safe garbage collected languages should generate
8626 these intrinsics to make use of the LLVM garbage collectors. For more
8627 details, see `Garbage Collection with LLVM <GarbageCollection.html>`_.
8629 Experimental Statepoint Intrinsics
8630 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8632 LLVM provides an second experimental set of intrinsics for describing garbage
8633 collection safepoints in compiled code. These intrinsics are an alternative
8634 to the ``llvm.gcroot`` intrinsics, but are compatible with the ones for
8635 :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers. The
8636 differences in approach are covered in the `Garbage Collection with LLVM
8637 <GarbageCollection.html>`_ documentation. The intrinsics themselves are
8638 described in :doc:`Statepoints`.
8642 '``llvm.gcroot``' Intrinsic
8643 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8650 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
8655 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
8656 the code generator, and allows some metadata to be associated with it.
8661 The first argument specifies the address of a stack object that contains
8662 the root pointer. The second pointer (which must be either a constant or
8663 a global value address) contains the meta-data to be associated with the
8669 At runtime, a call to this intrinsic stores a null pointer into the
8670 "ptrloc" location. At compile-time, the code generator generates
8671 information to allow the runtime to find the pointer at GC safe points.
8672 The '``llvm.gcroot``' intrinsic may only be used in a function which
8673 :ref:`specifies a GC algorithm <gc>`.
8677 '``llvm.gcread``' Intrinsic
8678 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8685 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
8690 The '``llvm.gcread``' intrinsic identifies reads of references from heap
8691 locations, allowing garbage collector implementations that require read
8697 The second argument is the address to read from, which should be an
8698 address allocated from the garbage collector. The first object is a
8699 pointer to the start of the referenced object, if needed by the language
8700 runtime (otherwise null).
8705 The '``llvm.gcread``' intrinsic has the same semantics as a load
8706 instruction, but may be replaced with substantially more complex code by
8707 the garbage collector runtime, as needed. The '``llvm.gcread``'
8708 intrinsic may only be used in a function which :ref:`specifies a GC
8713 '``llvm.gcwrite``' Intrinsic
8714 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8721 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
8726 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
8727 locations, allowing garbage collector implementations that require write
8728 barriers (such as generational or reference counting collectors).
8733 The first argument is the reference to store, the second is the start of
8734 the object to store it to, and the third is the address of the field of
8735 Obj to store to. If the runtime does not require a pointer to the
8736 object, Obj may be null.
8741 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
8742 instruction, but may be replaced with substantially more complex code by
8743 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
8744 intrinsic may only be used in a function which :ref:`specifies a GC
8747 Code Generator Intrinsics
8748 -------------------------
8750 These intrinsics are provided by LLVM to expose special features that
8751 may only be implemented with code generator support.
8753 '``llvm.returnaddress``' Intrinsic
8754 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8761 declare i8 *@llvm.returnaddress(i32 <level>)
8766 The '``llvm.returnaddress``' intrinsic attempts to compute a
8767 target-specific value indicating the return address of the current
8768 function or one of its callers.
8773 The argument to this intrinsic indicates which function to return the
8774 address for. Zero indicates the calling function, one indicates its
8775 caller, etc. The argument is **required** to be a constant integer
8781 The '``llvm.returnaddress``' intrinsic either returns a pointer
8782 indicating the return address of the specified call frame, or zero if it
8783 cannot be identified. The value returned by this intrinsic is likely to
8784 be incorrect or 0 for arguments other than zero, so it should only be
8785 used for debugging purposes.
8787 Note that calling this intrinsic does not prevent function inlining or
8788 other aggressive transformations, so the value returned may not be that
8789 of the obvious source-language caller.
8791 '``llvm.frameaddress``' Intrinsic
8792 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8799 declare i8* @llvm.frameaddress(i32 <level>)
8804 The '``llvm.frameaddress``' intrinsic attempts to return the
8805 target-specific frame pointer value for the specified stack frame.
8810 The argument to this intrinsic indicates which function to return the
8811 frame pointer for. Zero indicates the calling function, one indicates
8812 its caller, etc. The argument is **required** to be a constant integer
8818 The '``llvm.frameaddress``' intrinsic either returns a pointer
8819 indicating the frame address of the specified call frame, or zero if it
8820 cannot be identified. The value returned by this intrinsic is likely to
8821 be incorrect or 0 for arguments other than zero, so it should only be
8822 used for debugging purposes.
8824 Note that calling this intrinsic does not prevent function inlining or
8825 other aggressive transformations, so the value returned may not be that
8826 of the obvious source-language caller.
8828 '``llvm.localescape``' and '``llvm.localrecover``' Intrinsics
8829 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8836 declare void @llvm.localescape(...)
8837 declare i8* @llvm.localrecover(i8* %func, i8* %fp, i32 %idx)
8842 The '``llvm.localescape``' intrinsic escapes offsets of a collection of static
8843 allocas, and the '``llvm.localrecover``' intrinsic applies those offsets to a
8844 live frame pointer to recover the address of the allocation. The offset is
8845 computed during frame layout of the caller of ``llvm.localescape``.
8850 All arguments to '``llvm.localescape``' must be pointers to static allocas or
8851 casts of static allocas. Each function can only call '``llvm.localescape``'
8852 once, and it can only do so from the entry block.
8854 The ``func`` argument to '``llvm.localrecover``' must be a constant
8855 bitcasted pointer to a function defined in the current module. The code
8856 generator cannot determine the frame allocation offset of functions defined in
8859 The ``fp`` argument to '``llvm.localrecover``' must be a frame pointer of a
8860 call frame that is currently live. The return value of '``llvm.localaddress``'
8861 is one way to produce such a value, but various runtimes also expose a suitable
8862 pointer in platform-specific ways.
8864 The ``idx`` argument to '``llvm.localrecover``' indicates which alloca passed to
8865 '``llvm.localescape``' to recover. It is zero-indexed.
8870 These intrinsics allow a group of functions to share access to a set of local
8871 stack allocations of a one parent function. The parent function may call the
8872 '``llvm.localescape``' intrinsic once from the function entry block, and the
8873 child functions can use '``llvm.localrecover``' to access the escaped allocas.
8874 The '``llvm.localescape``' intrinsic blocks inlining, as inlining changes where
8875 the escaped allocas are allocated, which would break attempts to use
8876 '``llvm.localrecover``'.
8878 .. _int_read_register:
8879 .. _int_write_register:
8881 '``llvm.read_register``' and '``llvm.write_register``' Intrinsics
8882 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8889 declare i32 @llvm.read_register.i32(metadata)
8890 declare i64 @llvm.read_register.i64(metadata)
8891 declare void @llvm.write_register.i32(metadata, i32 @value)
8892 declare void @llvm.write_register.i64(metadata, i64 @value)
8898 The '``llvm.read_register``' and '``llvm.write_register``' intrinsics
8899 provides access to the named register. The register must be valid on
8900 the architecture being compiled to. The type needs to be compatible
8901 with the register being read.
8906 The '``llvm.read_register``' intrinsic returns the current value of the
8907 register, where possible. The '``llvm.write_register``' intrinsic sets
8908 the current value of the register, where possible.
8910 This is useful to implement named register global variables that need
8911 to always be mapped to a specific register, as is common practice on
8912 bare-metal programs including OS kernels.
8914 The compiler doesn't check for register availability or use of the used
8915 register in surrounding code, including inline assembly. Because of that,
8916 allocatable registers are not supported.
8918 Warning: So far it only works with the stack pointer on selected
8919 architectures (ARM, AArch64, PowerPC and x86_64). Significant amount of
8920 work is needed to support other registers and even more so, allocatable
8925 '``llvm.stacksave``' Intrinsic
8926 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8933 declare i8* @llvm.stacksave()
8938 The '``llvm.stacksave``' intrinsic is used to remember the current state
8939 of the function stack, for use with
8940 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
8941 implementing language features like scoped automatic variable sized
8947 This intrinsic returns a opaque pointer value that can be passed to
8948 :ref:`llvm.stackrestore <int_stackrestore>`. When an
8949 ``llvm.stackrestore`` intrinsic is executed with a value saved from
8950 ``llvm.stacksave``, it effectively restores the state of the stack to
8951 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
8952 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
8953 were allocated after the ``llvm.stacksave`` was executed.
8955 .. _int_stackrestore:
8957 '``llvm.stackrestore``' Intrinsic
8958 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8965 declare void @llvm.stackrestore(i8* %ptr)
8970 The '``llvm.stackrestore``' intrinsic is used to restore the state of
8971 the function stack to the state it was in when the corresponding
8972 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
8973 useful for implementing language features like scoped automatic variable
8974 sized arrays in C99.
8979 See the description for :ref:`llvm.stacksave <int_stacksave>`.
8981 '``llvm.prefetch``' Intrinsic
8982 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8989 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
8994 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
8995 insert a prefetch instruction if supported; otherwise, it is a noop.
8996 Prefetches have no effect on the behavior of the program but can change
8997 its performance characteristics.
9002 ``address`` is the address to be prefetched, ``rw`` is the specifier
9003 determining if the fetch should be for a read (0) or write (1), and
9004 ``locality`` is a temporal locality specifier ranging from (0) - no
9005 locality, to (3) - extremely local keep in cache. The ``cache type``
9006 specifies whether the prefetch is performed on the data (1) or
9007 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
9008 arguments must be constant integers.
9013 This intrinsic does not modify the behavior of the program. In
9014 particular, prefetches cannot trap and do not produce a value. On
9015 targets that support this intrinsic, the prefetch can provide hints to
9016 the processor cache for better performance.
9018 '``llvm.pcmarker``' Intrinsic
9019 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9026 declare void @llvm.pcmarker(i32 <id>)
9031 The '``llvm.pcmarker``' intrinsic is a method to export a Program
9032 Counter (PC) in a region of code to simulators and other tools. The
9033 method is target specific, but it is expected that the marker will use
9034 exported symbols to transmit the PC of the marker. The marker makes no
9035 guarantees that it will remain with any specific instruction after
9036 optimizations. It is possible that the presence of a marker will inhibit
9037 optimizations. The intended use is to be inserted after optimizations to
9038 allow correlations of simulation runs.
9043 ``id`` is a numerical id identifying the marker.
9048 This intrinsic does not modify the behavior of the program. Backends
9049 that do not support this intrinsic may ignore it.
9051 '``llvm.readcyclecounter``' Intrinsic
9052 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9059 declare i64 @llvm.readcyclecounter()
9064 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
9065 counter register (or similar low latency, high accuracy clocks) on those
9066 targets that support it. On X86, it should map to RDTSC. On Alpha, it
9067 should map to RPCC. As the backing counters overflow quickly (on the
9068 order of 9 seconds on alpha), this should only be used for small
9074 When directly supported, reading the cycle counter should not modify any
9075 memory. Implementations are allowed to either return a application
9076 specific value or a system wide value. On backends without support, this
9077 is lowered to a constant 0.
9079 Note that runtime support may be conditional on the privilege-level code is
9080 running at and the host platform.
9082 '``llvm.clear_cache``' Intrinsic
9083 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9090 declare void @llvm.clear_cache(i8*, i8*)
9095 The '``llvm.clear_cache``' intrinsic ensures visibility of modifications
9096 in the specified range to the execution unit of the processor. On
9097 targets with non-unified instruction and data cache, the implementation
9098 flushes the instruction cache.
9103 On platforms with coherent instruction and data caches (e.g. x86), this
9104 intrinsic is a nop. On platforms with non-coherent instruction and data
9105 cache (e.g. ARM, MIPS), the intrinsic is lowered either to appropriate
9106 instructions or a system call, if cache flushing requires special
9109 The default behavior is to emit a call to ``__clear_cache`` from the run
9112 This instrinsic does *not* empty the instruction pipeline. Modifications
9113 of the current function are outside the scope of the intrinsic.
9115 '``llvm.instrprof_increment``' Intrinsic
9116 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9123 declare void @llvm.instrprof_increment(i8* <name>, i64 <hash>,
9124 i32 <num-counters>, i32 <index>)
9129 The '``llvm.instrprof_increment``' intrinsic can be emitted by a
9130 frontend for use with instrumentation based profiling. These will be
9131 lowered by the ``-instrprof`` pass to generate execution counts of a
9137 The first argument is a pointer to a global variable containing the
9138 name of the entity being instrumented. This should generally be the
9139 (mangled) function name for a set of counters.
9141 The second argument is a hash value that can be used by the consumer
9142 of the profile data to detect changes to the instrumented source, and
9143 the third is the number of counters associated with ``name``. It is an
9144 error if ``hash`` or ``num-counters`` differ between two instances of
9145 ``instrprof_increment`` that refer to the same name.
9147 The last argument refers to which of the counters for ``name`` should
9148 be incremented. It should be a value between 0 and ``num-counters``.
9153 This intrinsic represents an increment of a profiling counter. It will
9154 cause the ``-instrprof`` pass to generate the appropriate data
9155 structures and the code to increment the appropriate value, in a
9156 format that can be written out by a compiler runtime and consumed via
9157 the ``llvm-profdata`` tool.
9159 Standard C Library Intrinsics
9160 -----------------------------
9162 LLVM provides intrinsics for a few important standard C library
9163 functions. These intrinsics allow source-language front-ends to pass
9164 information about the alignment of the pointer arguments to the code
9165 generator, providing opportunity for more efficient code generation.
9169 '``llvm.memcpy``' Intrinsic
9170 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9175 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
9176 integer bit width and for different address spaces. Not all targets
9177 support all bit widths however.
9181 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
9182 i32 <len>, i32 <align>, i1 <isvolatile>)
9183 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
9184 i64 <len>, i32 <align>, i1 <isvolatile>)
9189 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
9190 source location to the destination location.
9192 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
9193 intrinsics do not return a value, takes extra alignment/isvolatile
9194 arguments and the pointers can be in specified address spaces.
9199 The first argument is a pointer to the destination, the second is a
9200 pointer to the source. The third argument is an integer argument
9201 specifying the number of bytes to copy, the fourth argument is the
9202 alignment of the source and destination locations, and the fifth is a
9203 boolean indicating a volatile access.
9205 If the call to this intrinsic has an alignment value that is not 0 or 1,
9206 then the caller guarantees that both the source and destination pointers
9207 are aligned to that boundary.
9209 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
9210 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
9211 very cleanly specified and it is unwise to depend on it.
9216 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
9217 source location to the destination location, which are not allowed to
9218 overlap. It copies "len" bytes of memory over. If the argument is known
9219 to be aligned to some boundary, this can be specified as the fourth
9220 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
9222 '``llvm.memmove``' Intrinsic
9223 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9228 This is an overloaded intrinsic. You can use llvm.memmove on any integer
9229 bit width and for different address space. Not all targets support all
9234 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
9235 i32 <len>, i32 <align>, i1 <isvolatile>)
9236 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
9237 i64 <len>, i32 <align>, i1 <isvolatile>)
9242 The '``llvm.memmove.*``' intrinsics move a block of memory from the
9243 source location to the destination location. It is similar to the
9244 '``llvm.memcpy``' intrinsic but allows the two memory locations to
9247 Note that, unlike the standard libc function, the ``llvm.memmove.*``
9248 intrinsics do not return a value, takes extra alignment/isvolatile
9249 arguments and the pointers can be in specified address spaces.
9254 The first argument is a pointer to the destination, the second is a
9255 pointer to the source. The third argument is an integer argument
9256 specifying the number of bytes to copy, the fourth argument is the
9257 alignment of the source and destination locations, and the fifth is a
9258 boolean indicating a volatile access.
9260 If the call to this intrinsic has an alignment value that is not 0 or 1,
9261 then the caller guarantees that the source and destination pointers are
9262 aligned to that boundary.
9264 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
9265 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
9266 not very cleanly specified and it is unwise to depend on it.
9271 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
9272 source location to the destination location, which may overlap. It
9273 copies "len" bytes of memory over. If the argument is known to be
9274 aligned to some boundary, this can be specified as the fourth argument,
9275 otherwise it should be set to 0 or 1 (both meaning no alignment).
9277 '``llvm.memset.*``' Intrinsics
9278 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9283 This is an overloaded intrinsic. You can use llvm.memset on any integer
9284 bit width and for different address spaces. However, not all targets
9285 support all bit widths.
9289 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
9290 i32 <len>, i32 <align>, i1 <isvolatile>)
9291 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
9292 i64 <len>, i32 <align>, i1 <isvolatile>)
9297 The '``llvm.memset.*``' intrinsics fill a block of memory with a
9298 particular byte value.
9300 Note that, unlike the standard libc function, the ``llvm.memset``
9301 intrinsic does not return a value and takes extra alignment/volatile
9302 arguments. Also, the destination can be in an arbitrary address space.
9307 The first argument is a pointer to the destination to fill, the second
9308 is the byte value with which to fill it, the third argument is an
9309 integer argument specifying the number of bytes to fill, and the fourth
9310 argument is the known alignment of the destination location.
9312 If the call to this intrinsic has an alignment value that is not 0 or 1,
9313 then the caller guarantees that the destination pointer is aligned to
9316 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
9317 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
9318 very cleanly specified and it is unwise to depend on it.
9323 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
9324 at the destination location. If the argument is known to be aligned to
9325 some boundary, this can be specified as the fourth argument, otherwise
9326 it should be set to 0 or 1 (both meaning no alignment).
9328 '``llvm.sqrt.*``' Intrinsic
9329 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9334 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
9335 floating point or vector of floating point type. Not all targets support
9340 declare float @llvm.sqrt.f32(float %Val)
9341 declare double @llvm.sqrt.f64(double %Val)
9342 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
9343 declare fp128 @llvm.sqrt.f128(fp128 %Val)
9344 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
9349 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
9350 returning the same value as the libm '``sqrt``' functions would. Unlike
9351 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
9352 negative numbers other than -0.0 (which allows for better optimization,
9353 because there is no need to worry about errno being set).
9354 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
9359 The argument and return value are floating point numbers of the same
9365 This function returns the sqrt of the specified operand if it is a
9366 nonnegative floating point number.
9368 '``llvm.powi.*``' Intrinsic
9369 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9374 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
9375 floating point or vector of floating point type. Not all targets support
9380 declare float @llvm.powi.f32(float %Val, i32 %power)
9381 declare double @llvm.powi.f64(double %Val, i32 %power)
9382 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
9383 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
9384 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
9389 The '``llvm.powi.*``' intrinsics return the first operand raised to the
9390 specified (positive or negative) power. The order of evaluation of
9391 multiplications is not defined. When a vector of floating point type is
9392 used, the second argument remains a scalar integer value.
9397 The second argument is an integer power, and the first is a value to
9398 raise to that power.
9403 This function returns the first value raised to the second power with an
9404 unspecified sequence of rounding operations.
9406 '``llvm.sin.*``' Intrinsic
9407 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9412 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
9413 floating point or vector of floating point type. Not all targets support
9418 declare float @llvm.sin.f32(float %Val)
9419 declare double @llvm.sin.f64(double %Val)
9420 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
9421 declare fp128 @llvm.sin.f128(fp128 %Val)
9422 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
9427 The '``llvm.sin.*``' intrinsics return the sine of the operand.
9432 The argument and return value are floating point numbers of the same
9438 This function returns the sine of the specified operand, returning the
9439 same values as the libm ``sin`` functions would, and handles error
9440 conditions in the same way.
9442 '``llvm.cos.*``' Intrinsic
9443 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9448 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
9449 floating point or vector of floating point type. Not all targets support
9454 declare float @llvm.cos.f32(float %Val)
9455 declare double @llvm.cos.f64(double %Val)
9456 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
9457 declare fp128 @llvm.cos.f128(fp128 %Val)
9458 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
9463 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
9468 The argument and return value are floating point numbers of the same
9474 This function returns the cosine of the specified operand, returning the
9475 same values as the libm ``cos`` functions would, and handles error
9476 conditions in the same way.
9478 '``llvm.pow.*``' Intrinsic
9479 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9484 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
9485 floating point or vector of floating point type. Not all targets support
9490 declare float @llvm.pow.f32(float %Val, float %Power)
9491 declare double @llvm.pow.f64(double %Val, double %Power)
9492 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
9493 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
9494 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
9499 The '``llvm.pow.*``' intrinsics return the first operand raised to the
9500 specified (positive or negative) power.
9505 The second argument is a floating point power, and the first is a value
9506 to raise to that power.
9511 This function returns the first value raised to the second power,
9512 returning the same values as the libm ``pow`` functions would, and
9513 handles error conditions in the same way.
9515 '``llvm.exp.*``' Intrinsic
9516 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9521 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
9522 floating point or vector of floating point type. Not all targets support
9527 declare float @llvm.exp.f32(float %Val)
9528 declare double @llvm.exp.f64(double %Val)
9529 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
9530 declare fp128 @llvm.exp.f128(fp128 %Val)
9531 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
9536 The '``llvm.exp.*``' intrinsics perform the exp function.
9541 The argument and return value are floating point numbers of the same
9547 This function returns the same values as the libm ``exp`` functions
9548 would, and handles error conditions in the same way.
9550 '``llvm.exp2.*``' Intrinsic
9551 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9556 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
9557 floating point or vector of floating point type. Not all targets support
9562 declare float @llvm.exp2.f32(float %Val)
9563 declare double @llvm.exp2.f64(double %Val)
9564 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
9565 declare fp128 @llvm.exp2.f128(fp128 %Val)
9566 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
9571 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
9576 The argument and return value are floating point numbers of the same
9582 This function returns the same values as the libm ``exp2`` functions
9583 would, and handles error conditions in the same way.
9585 '``llvm.log.*``' Intrinsic
9586 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9591 This is an overloaded intrinsic. You can use ``llvm.log`` on any
9592 floating point or vector of floating point type. Not all targets support
9597 declare float @llvm.log.f32(float %Val)
9598 declare double @llvm.log.f64(double %Val)
9599 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
9600 declare fp128 @llvm.log.f128(fp128 %Val)
9601 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
9606 The '``llvm.log.*``' intrinsics perform the log function.
9611 The argument and return value are floating point numbers of the same
9617 This function returns the same values as the libm ``log`` functions
9618 would, and handles error conditions in the same way.
9620 '``llvm.log10.*``' Intrinsic
9621 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9626 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
9627 floating point or vector of floating point type. Not all targets support
9632 declare float @llvm.log10.f32(float %Val)
9633 declare double @llvm.log10.f64(double %Val)
9634 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
9635 declare fp128 @llvm.log10.f128(fp128 %Val)
9636 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
9641 The '``llvm.log10.*``' intrinsics perform the log10 function.
9646 The argument and return value are floating point numbers of the same
9652 This function returns the same values as the libm ``log10`` functions
9653 would, and handles error conditions in the same way.
9655 '``llvm.log2.*``' Intrinsic
9656 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9661 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
9662 floating point or vector of floating point type. Not all targets support
9667 declare float @llvm.log2.f32(float %Val)
9668 declare double @llvm.log2.f64(double %Val)
9669 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
9670 declare fp128 @llvm.log2.f128(fp128 %Val)
9671 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
9676 The '``llvm.log2.*``' intrinsics perform the log2 function.
9681 The argument and return value are floating point numbers of the same
9687 This function returns the same values as the libm ``log2`` functions
9688 would, and handles error conditions in the same way.
9690 '``llvm.fma.*``' Intrinsic
9691 ^^^^^^^^^^^^^^^^^^^^^^^^^^
9696 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
9697 floating point or vector of floating point type. Not all targets support
9702 declare float @llvm.fma.f32(float %a, float %b, float %c)
9703 declare double @llvm.fma.f64(double %a, double %b, double %c)
9704 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
9705 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
9706 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
9711 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
9717 The argument and return value are floating point numbers of the same
9723 This function returns the same values as the libm ``fma`` functions
9724 would, and does not set errno.
9726 '``llvm.fabs.*``' Intrinsic
9727 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9732 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
9733 floating point or vector of floating point type. Not all targets support
9738 declare float @llvm.fabs.f32(float %Val)
9739 declare double @llvm.fabs.f64(double %Val)
9740 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
9741 declare fp128 @llvm.fabs.f128(fp128 %Val)
9742 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
9747 The '``llvm.fabs.*``' intrinsics return the absolute value of the
9753 The argument and return value are floating point numbers of the same
9759 This function returns the same values as the libm ``fabs`` functions
9760 would, and handles error conditions in the same way.
9762 '``llvm.minnum.*``' Intrinsic
9763 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9768 This is an overloaded intrinsic. You can use ``llvm.minnum`` on any
9769 floating point or vector of floating point type. Not all targets support
9774 declare float @llvm.minnum.f32(float %Val0, float %Val1)
9775 declare double @llvm.minnum.f64(double %Val0, double %Val1)
9776 declare x86_fp80 @llvm.minnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
9777 declare fp128 @llvm.minnum.f128(fp128 %Val0, fp128 %Val1)
9778 declare ppc_fp128 @llvm.minnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
9783 The '``llvm.minnum.*``' intrinsics return the minimum of the two
9790 The arguments and return value are floating point numbers of the same
9796 Follows the IEEE-754 semantics for minNum, which also match for libm's
9799 If either operand is a NaN, returns the other non-NaN operand. Returns
9800 NaN only if both operands are NaN. If the operands compare equal,
9801 returns a value that compares equal to both operands. This means that
9802 fmin(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
9804 '``llvm.maxnum.*``' Intrinsic
9805 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9810 This is an overloaded intrinsic. You can use ``llvm.maxnum`` on any
9811 floating point or vector of floating point type. Not all targets support
9816 declare float @llvm.maxnum.f32(float %Val0, float %Val1l)
9817 declare double @llvm.maxnum.f64(double %Val0, double %Val1)
9818 declare x86_fp80 @llvm.maxnum.f80(x86_fp80 %Val0, x86_fp80 %Val1)
9819 declare fp128 @llvm.maxnum.f128(fp128 %Val0, fp128 %Val1)
9820 declare ppc_fp128 @llvm.maxnum.ppcf128(ppc_fp128 %Val0, ppc_fp128 %Val1)
9825 The '``llvm.maxnum.*``' intrinsics return the maximum of the two
9832 The arguments and return value are floating point numbers of the same
9837 Follows the IEEE-754 semantics for maxNum, which also match for libm's
9840 If either operand is a NaN, returns the other non-NaN operand. Returns
9841 NaN only if both operands are NaN. If the operands compare equal,
9842 returns a value that compares equal to both operands. This means that
9843 fmax(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
9845 '``llvm.copysign.*``' Intrinsic
9846 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9851 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
9852 floating point or vector of floating point type. Not all targets support
9857 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
9858 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
9859 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
9860 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
9861 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
9866 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
9867 first operand and the sign of the second operand.
9872 The arguments and return value are floating point numbers of the same
9878 This function returns the same values as the libm ``copysign``
9879 functions would, and handles error conditions in the same way.
9881 '``llvm.floor.*``' Intrinsic
9882 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9887 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
9888 floating point or vector of floating point type. Not all targets support
9893 declare float @llvm.floor.f32(float %Val)
9894 declare double @llvm.floor.f64(double %Val)
9895 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
9896 declare fp128 @llvm.floor.f128(fp128 %Val)
9897 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
9902 The '``llvm.floor.*``' intrinsics return the floor of the operand.
9907 The argument and return value are floating point numbers of the same
9913 This function returns the same values as the libm ``floor`` functions
9914 would, and handles error conditions in the same way.
9916 '``llvm.ceil.*``' Intrinsic
9917 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9922 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
9923 floating point or vector of floating point type. Not all targets support
9928 declare float @llvm.ceil.f32(float %Val)
9929 declare double @llvm.ceil.f64(double %Val)
9930 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
9931 declare fp128 @llvm.ceil.f128(fp128 %Val)
9932 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
9937 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
9942 The argument and return value are floating point numbers of the same
9948 This function returns the same values as the libm ``ceil`` functions
9949 would, and handles error conditions in the same way.
9951 '``llvm.trunc.*``' Intrinsic
9952 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9957 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
9958 floating point or vector of floating point type. Not all targets support
9963 declare float @llvm.trunc.f32(float %Val)
9964 declare double @llvm.trunc.f64(double %Val)
9965 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
9966 declare fp128 @llvm.trunc.f128(fp128 %Val)
9967 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
9972 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
9973 nearest integer not larger in magnitude than the operand.
9978 The argument and return value are floating point numbers of the same
9984 This function returns the same values as the libm ``trunc`` functions
9985 would, and handles error conditions in the same way.
9987 '``llvm.rint.*``' Intrinsic
9988 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9993 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
9994 floating point or vector of floating point type. Not all targets support
9999 declare float @llvm.rint.f32(float %Val)
10000 declare double @llvm.rint.f64(double %Val)
10001 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
10002 declare fp128 @llvm.rint.f128(fp128 %Val)
10003 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
10008 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
10009 nearest integer. It may raise an inexact floating-point exception if the
10010 operand isn't an integer.
10015 The argument and return value are floating point numbers of the same
10021 This function returns the same values as the libm ``rint`` functions
10022 would, and handles error conditions in the same way.
10024 '``llvm.nearbyint.*``' Intrinsic
10025 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10030 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
10031 floating point or vector of floating point type. Not all targets support
10036 declare float @llvm.nearbyint.f32(float %Val)
10037 declare double @llvm.nearbyint.f64(double %Val)
10038 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
10039 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
10040 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
10045 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
10051 The argument and return value are floating point numbers of the same
10057 This function returns the same values as the libm ``nearbyint``
10058 functions would, and handles error conditions in the same way.
10060 '``llvm.round.*``' Intrinsic
10061 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10066 This is an overloaded intrinsic. You can use ``llvm.round`` on any
10067 floating point or vector of floating point type. Not all targets support
10072 declare float @llvm.round.f32(float %Val)
10073 declare double @llvm.round.f64(double %Val)
10074 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
10075 declare fp128 @llvm.round.f128(fp128 %Val)
10076 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
10081 The '``llvm.round.*``' intrinsics returns the operand rounded to the
10087 The argument and return value are floating point numbers of the same
10093 This function returns the same values as the libm ``round``
10094 functions would, and handles error conditions in the same way.
10096 Bit Manipulation Intrinsics
10097 ---------------------------
10099 LLVM provides intrinsics for a few important bit manipulation
10100 operations. These allow efficient code generation for some algorithms.
10102 '``llvm.bswap.*``' Intrinsics
10103 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10108 This is an overloaded intrinsic function. You can use bswap on any
10109 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
10113 declare i16 @llvm.bswap.i16(i16 <id>)
10114 declare i32 @llvm.bswap.i32(i32 <id>)
10115 declare i64 @llvm.bswap.i64(i64 <id>)
10120 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
10121 values with an even number of bytes (positive multiple of 16 bits).
10122 These are useful for performing operations on data that is not in the
10123 target's native byte order.
10128 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
10129 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
10130 intrinsic returns an i32 value that has the four bytes of the input i32
10131 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
10132 returned i32 will have its bytes in 3, 2, 1, 0 order. The
10133 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
10134 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
10137 '``llvm.ctpop.*``' Intrinsic
10138 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10143 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
10144 bit width, or on any vector with integer elements. Not all targets
10145 support all bit widths or vector types, however.
10149 declare i8 @llvm.ctpop.i8(i8 <src>)
10150 declare i16 @llvm.ctpop.i16(i16 <src>)
10151 declare i32 @llvm.ctpop.i32(i32 <src>)
10152 declare i64 @llvm.ctpop.i64(i64 <src>)
10153 declare i256 @llvm.ctpop.i256(i256 <src>)
10154 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
10159 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
10165 The only argument is the value to be counted. The argument may be of any
10166 integer type, or a vector with integer elements. The return type must
10167 match the argument type.
10172 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
10173 each element of a vector.
10175 '``llvm.ctlz.*``' Intrinsic
10176 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10181 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
10182 integer bit width, or any vector whose elements are integers. Not all
10183 targets support all bit widths or vector types, however.
10187 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
10188 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
10189 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
10190 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
10191 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
10192 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
10197 The '``llvm.ctlz``' family of intrinsic functions counts the number of
10198 leading zeros in a variable.
10203 The first argument is the value to be counted. This argument may be of
10204 any integer type, or a vector with integer element type. The return
10205 type must match the first argument type.
10207 The second argument must be a constant and is a flag to indicate whether
10208 the intrinsic should ensure that a zero as the first argument produces a
10209 defined result. Historically some architectures did not provide a
10210 defined result for zero values as efficiently, and many algorithms are
10211 now predicated on avoiding zero-value inputs.
10216 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
10217 zeros in a variable, or within each element of the vector. If
10218 ``src == 0`` then the result is the size in bits of the type of ``src``
10219 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
10220 ``llvm.ctlz(i32 2) = 30``.
10222 '``llvm.cttz.*``' Intrinsic
10223 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
10228 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
10229 integer bit width, or any vector of integer elements. Not all targets
10230 support all bit widths or vector types, however.
10234 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
10235 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
10236 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
10237 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
10238 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
10239 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
10244 The '``llvm.cttz``' family of intrinsic functions counts the number of
10250 The first argument is the value to be counted. This argument may be of
10251 any integer type, or a vector with integer element type. The return
10252 type must match the first argument type.
10254 The second argument must be a constant and is a flag to indicate whether
10255 the intrinsic should ensure that a zero as the first argument produces a
10256 defined result. Historically some architectures did not provide a
10257 defined result for zero values as efficiently, and many algorithms are
10258 now predicated on avoiding zero-value inputs.
10263 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
10264 zeros in a variable, or within each element of a vector. If ``src == 0``
10265 then the result is the size in bits of the type of ``src`` if
10266 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
10267 ``llvm.cttz(2) = 1``.
10271 Arithmetic with Overflow Intrinsics
10272 -----------------------------------
10274 LLVM provides intrinsics for some arithmetic with overflow operations.
10276 '``llvm.sadd.with.overflow.*``' Intrinsics
10277 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10282 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
10283 on any integer bit width.
10287 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
10288 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
10289 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
10294 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
10295 a signed addition of the two arguments, and indicate whether an overflow
10296 occurred during the signed summation.
10301 The arguments (%a and %b) and the first element of the result structure
10302 may be of integer types of any bit width, but they must have the same
10303 bit width. The second element of the result structure must be of type
10304 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
10310 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
10311 a signed addition of the two variables. They return a structure --- the
10312 first element of which is the signed summation, and the second element
10313 of which is a bit specifying if the signed summation resulted in an
10319 .. code-block:: llvm
10321 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
10322 %sum = extractvalue {i32, i1} %res, 0
10323 %obit = extractvalue {i32, i1} %res, 1
10324 br i1 %obit, label %overflow, label %normal
10326 '``llvm.uadd.with.overflow.*``' Intrinsics
10327 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10332 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
10333 on any integer bit width.
10337 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
10338 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
10339 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
10344 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
10345 an unsigned addition of the two arguments, and indicate whether a carry
10346 occurred during the unsigned summation.
10351 The arguments (%a and %b) and the first element of the result structure
10352 may be of integer types of any bit width, but they must have the same
10353 bit width. The second element of the result structure must be of type
10354 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
10360 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
10361 an unsigned addition of the two arguments. They return a structure --- the
10362 first element of which is the sum, and the second element of which is a
10363 bit specifying if the unsigned summation resulted in a carry.
10368 .. code-block:: llvm
10370 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
10371 %sum = extractvalue {i32, i1} %res, 0
10372 %obit = extractvalue {i32, i1} %res, 1
10373 br i1 %obit, label %carry, label %normal
10375 '``llvm.ssub.with.overflow.*``' Intrinsics
10376 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10381 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
10382 on any integer bit width.
10386 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
10387 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
10388 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
10393 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
10394 a signed subtraction of the two arguments, and indicate whether an
10395 overflow occurred during the signed subtraction.
10400 The arguments (%a and %b) and the first element of the result structure
10401 may be of integer types of any bit width, but they must have the same
10402 bit width. The second element of the result structure must be of type
10403 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
10409 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
10410 a signed subtraction of the two arguments. They return a structure --- the
10411 first element of which is the subtraction, and the second element of
10412 which is a bit specifying if the signed subtraction resulted in an
10418 .. code-block:: llvm
10420 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
10421 %sum = extractvalue {i32, i1} %res, 0
10422 %obit = extractvalue {i32, i1} %res, 1
10423 br i1 %obit, label %overflow, label %normal
10425 '``llvm.usub.with.overflow.*``' Intrinsics
10426 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10431 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
10432 on any integer bit width.
10436 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
10437 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
10438 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
10443 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
10444 an unsigned subtraction of the two arguments, and indicate whether an
10445 overflow occurred during the unsigned subtraction.
10450 The arguments (%a and %b) and the first element of the result structure
10451 may be of integer types of any bit width, but they must have the same
10452 bit width. The second element of the result structure must be of type
10453 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
10459 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
10460 an unsigned subtraction of the two arguments. They return a structure ---
10461 the first element of which is the subtraction, and the second element of
10462 which is a bit specifying if the unsigned subtraction resulted in an
10468 .. code-block:: llvm
10470 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
10471 %sum = extractvalue {i32, i1} %res, 0
10472 %obit = extractvalue {i32, i1} %res, 1
10473 br i1 %obit, label %overflow, label %normal
10475 '``llvm.smul.with.overflow.*``' Intrinsics
10476 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10481 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
10482 on any integer bit width.
10486 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
10487 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
10488 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
10493 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
10494 a signed multiplication of the two arguments, and indicate whether an
10495 overflow occurred during the signed multiplication.
10500 The arguments (%a and %b) and the first element of the result structure
10501 may be of integer types of any bit width, but they must have the same
10502 bit width. The second element of the result structure must be of type
10503 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
10509 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
10510 a signed multiplication of the two arguments. They return a structure ---
10511 the first element of which is the multiplication, and the second element
10512 of which is a bit specifying if the signed multiplication resulted in an
10518 .. code-block:: llvm
10520 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
10521 %sum = extractvalue {i32, i1} %res, 0
10522 %obit = extractvalue {i32, i1} %res, 1
10523 br i1 %obit, label %overflow, label %normal
10525 '``llvm.umul.with.overflow.*``' Intrinsics
10526 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10531 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
10532 on any integer bit width.
10536 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
10537 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
10538 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
10543 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
10544 a unsigned multiplication of the two arguments, and indicate whether an
10545 overflow occurred during the unsigned multiplication.
10550 The arguments (%a and %b) and the first element of the result structure
10551 may be of integer types of any bit width, but they must have the same
10552 bit width. The second element of the result structure must be of type
10553 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
10559 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
10560 an unsigned multiplication of the two arguments. They return a structure ---
10561 the first element of which is the multiplication, and the second
10562 element of which is a bit specifying if the unsigned multiplication
10563 resulted in an overflow.
10568 .. code-block:: llvm
10570 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
10571 %sum = extractvalue {i32, i1} %res, 0
10572 %obit = extractvalue {i32, i1} %res, 1
10573 br i1 %obit, label %overflow, label %normal
10575 Specialised Arithmetic Intrinsics
10576 ---------------------------------
10578 '``llvm.canonicalize.*``' Intrinsic
10579 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10586 declare float @llvm.canonicalize.f32(float %a)
10587 declare double @llvm.canonicalize.f64(double %b)
10592 The '``llvm.canonicalize.*``' intrinsic returns the platform specific canonical
10593 encoding of a floating point number. This canonicalization is useful for
10594 implementing certain numeric primitives such as frexp. The canonical encoding is
10595 defined by IEEE-754-2008 to be:
10599 2.1.8 canonical encoding: The preferred encoding of a floating-point
10600 representation in a format. Applied to declets, significands of finite
10601 numbers, infinities, and NaNs, especially in decimal formats.
10603 This operation can also be considered equivalent to the IEEE-754-2008
10604 conversion of a floating-point value to the same format. NaNs are handled
10605 according to section 6.2.
10607 Examples of non-canonical encodings:
10609 - x87 pseudo denormals, pseudo NaNs, pseudo Infinity, Unnormals. These are
10610 converted to a canonical representation per hardware-specific protocol.
10611 - Many normal decimal floating point numbers have non-canonical alternative
10613 - Some machines, like GPUs or ARMv7 NEON, do not support subnormal values.
10614 These are treated as non-canonical encodings of zero and with be flushed to
10615 a zero of the same sign by this operation.
10617 Note that per IEEE-754-2008 6.2, systems that support signaling NaNs with
10618 default exception handling must signal an invalid exception, and produce a
10621 This function should always be implementable as multiplication by 1.0, provided
10622 that the compiler does not constant fold the operation. Likewise, division by
10623 1.0 and ``llvm.minnum(x, x)`` are possible implementations. Addition with
10624 -0.0 is also sufficient provided that the rounding mode is not -Infinity.
10626 ``@llvm.canonicalize`` must preserve the equality relation. That is:
10628 - ``(@llvm.canonicalize(x) == x)`` is equivalent to ``(x == x)``
10629 - ``(@llvm.canonicalize(x) == @llvm.canonicalize(y))`` is equivalent to
10632 Additionally, the sign of zero must be conserved:
10633 ``@llvm.canonicalize(-0.0) = -0.0`` and ``@llvm.canonicalize(+0.0) = +0.0``
10635 The payload bits of a NaN must be conserved, with two exceptions.
10636 First, environments which use only a single canonical representation of NaN
10637 must perform said canonicalization. Second, SNaNs must be quieted per the
10640 The canonicalization operation may be optimized away if:
10642 - The input is known to be canonical. For example, it was produced by a
10643 floating-point operation that is required by the standard to be canonical.
10644 - The result is consumed only by (or fused with) other floating-point
10645 operations. That is, the bits of the floating point value are not examined.
10647 '``llvm.fmuladd.*``' Intrinsic
10648 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10655 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
10656 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
10661 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
10662 expressions that can be fused if the code generator determines that (a) the
10663 target instruction set has support for a fused operation, and (b) that the
10664 fused operation is more efficient than the equivalent, separate pair of mul
10665 and add instructions.
10670 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
10671 multiplicands, a and b, and an addend c.
10680 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
10682 is equivalent to the expression a \* b + c, except that rounding will
10683 not be performed between the multiplication and addition steps if the
10684 code generator fuses the operations. Fusion is not guaranteed, even if
10685 the target platform supports it. If a fused multiply-add is required the
10686 corresponding llvm.fma.\* intrinsic function should be used
10687 instead. This never sets errno, just as '``llvm.fma.*``'.
10692 .. code-block:: llvm
10694 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields float:r2 = (a * b) + c
10697 '``llvm.uabsdiff.*``' and '``llvm.sabsdiff.*``' Intrinsics
10698 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10702 This is an overloaded intrinsic. The loaded data is a vector of any integer bit width.
10704 .. code-block:: llvm
10706 declare <4 x integer> @llvm.uabsdiff.v4i32(<4 x integer> %a, <4 x integer> %b)
10712 The ``llvm.uabsdiff`` intrinsic returns a vector result of the absolute difference of the two operands,
10713 treating them both as unsigned integers.
10715 The ``llvm.sabsdiff`` intrinsic returns a vector result of the absolute difference of the two operands,
10716 treating them both as signed integers.
10720 These intrinsics are primarily used during the code generation stage of compilation.
10721 They are generated by compiler passes such as the Loop and SLP vectorizers.it is not
10722 recommended for users to create them manually.
10727 Both intrinsics take two integer of the same bitwidth.
10734 call <4 x i32> @llvm.uabsdiff.v4i32(<4 x i32> %a, <4 x i32> %b)
10738 %sub = sub <4 x i32> %a, %b
10739 %ispos = icmp ugt <4 x i32> %sub, <i32 -1, i32 -1, i32 -1, i32 -1>
10740 %neg = sub <4 x i32> zeroinitializer, %sub
10741 %1 = select <4 x i1> %ispos, <4 x i32> %sub, <4 x i32> %neg
10743 Similarly the expression::
10745 call <4 x i32> @llvm.sabsdiff.v4i32(<4 x i32> %a, <4 x i32> %b)
10749 %sub = sub nsw <4 x i32> %a, %b
10750 %ispos = icmp sgt <4 x i32> %sub, <i32 -1, i32 -1, i32 -1, i32 -1>
10751 %neg = sub nsw <4 x i32> zeroinitializer, %sub
10752 %1 = select <4 x i1> %ispos, <4 x i32> %sub, <4 x i32> %neg
10755 Half Precision Floating Point Intrinsics
10756 ----------------------------------------
10758 For most target platforms, half precision floating point is a
10759 storage-only format. This means that it is a dense encoding (in memory)
10760 but does not support computation in the format.
10762 This means that code must first load the half-precision floating point
10763 value as an i16, then convert it to float with
10764 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
10765 then be performed on the float value (including extending to double
10766 etc). To store the value back to memory, it is first converted to float
10767 if needed, then converted to i16 with
10768 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
10771 .. _int_convert_to_fp16:
10773 '``llvm.convert.to.fp16``' Intrinsic
10774 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10781 declare i16 @llvm.convert.to.fp16.f32(float %a)
10782 declare i16 @llvm.convert.to.fp16.f64(double %a)
10787 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
10788 conventional floating point type to half precision floating point format.
10793 The intrinsic function contains single argument - the value to be
10799 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
10800 conventional floating point format to half precision floating point format. The
10801 return value is an ``i16`` which contains the converted number.
10806 .. code-block:: llvm
10808 %res = call i16 @llvm.convert.to.fp16.f32(float %a)
10809 store i16 %res, i16* @x, align 2
10811 .. _int_convert_from_fp16:
10813 '``llvm.convert.from.fp16``' Intrinsic
10814 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10821 declare float @llvm.convert.from.fp16.f32(i16 %a)
10822 declare double @llvm.convert.from.fp16.f64(i16 %a)
10827 The '``llvm.convert.from.fp16``' intrinsic function performs a
10828 conversion from half precision floating point format to single precision
10829 floating point format.
10834 The intrinsic function contains single argument - the value to be
10840 The '``llvm.convert.from.fp16``' intrinsic function performs a
10841 conversion from half single precision floating point format to single
10842 precision floating point format. The input half-float value is
10843 represented by an ``i16`` value.
10848 .. code-block:: llvm
10850 %a = load i16, i16* @x, align 2
10851 %res = call float @llvm.convert.from.fp16(i16 %a)
10853 .. _dbg_intrinsics:
10855 Debugger Intrinsics
10856 -------------------
10858 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
10859 prefix), are described in the `LLVM Source Level
10860 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
10863 Exception Handling Intrinsics
10864 -----------------------------
10866 The LLVM exception handling intrinsics (which all start with
10867 ``llvm.eh.`` prefix), are described in the `LLVM Exception
10868 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
10870 .. _int_trampoline:
10872 Trampoline Intrinsics
10873 ---------------------
10875 These intrinsics make it possible to excise one parameter, marked with
10876 the :ref:`nest <nest>` attribute, from a function. The result is a
10877 callable function pointer lacking the nest parameter - the caller does
10878 not need to provide a value for it. Instead, the value to use is stored
10879 in advance in a "trampoline", a block of memory usually allocated on the
10880 stack, which also contains code to splice the nest value into the
10881 argument list. This is used to implement the GCC nested function address
10884 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
10885 then the resulting function pointer has signature ``i32 (i32, i32)*``.
10886 It can be created as follows:
10888 .. code-block:: llvm
10890 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
10891 %tramp1 = getelementptr [10 x i8], [10 x i8]* %tramp, i32 0, i32 0
10892 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
10893 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
10894 %fp = bitcast i8* %p to i32 (i32, i32)*
10896 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
10897 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
10901 '``llvm.init.trampoline``' Intrinsic
10902 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10909 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
10914 This fills the memory pointed to by ``tramp`` with executable code,
10915 turning it into a trampoline.
10920 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
10921 pointers. The ``tramp`` argument must point to a sufficiently large and
10922 sufficiently aligned block of memory; this memory is written to by the
10923 intrinsic. Note that the size and the alignment are target-specific -
10924 LLVM currently provides no portable way of determining them, so a
10925 front-end that generates this intrinsic needs to have some
10926 target-specific knowledge. The ``func`` argument must hold a function
10927 bitcast to an ``i8*``.
10932 The block of memory pointed to by ``tramp`` is filled with target
10933 dependent code, turning it into a function. Then ``tramp`` needs to be
10934 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
10935 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
10936 function's signature is the same as that of ``func`` with any arguments
10937 marked with the ``nest`` attribute removed. At most one such ``nest``
10938 argument is allowed, and it must be of pointer type. Calling the new
10939 function is equivalent to calling ``func`` with the same argument list,
10940 but with ``nval`` used for the missing ``nest`` argument. If, after
10941 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
10942 modified, then the effect of any later call to the returned function
10943 pointer is undefined.
10947 '``llvm.adjust.trampoline``' Intrinsic
10948 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10955 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
10960 This performs any required machine-specific adjustment to the address of
10961 a trampoline (passed as ``tramp``).
10966 ``tramp`` must point to a block of memory which already has trampoline
10967 code filled in by a previous call to
10968 :ref:`llvm.init.trampoline <int_it>`.
10973 On some architectures the address of the code to be executed needs to be
10974 different than the address where the trampoline is actually stored. This
10975 intrinsic returns the executable address corresponding to ``tramp``
10976 after performing the required machine specific adjustments. The pointer
10977 returned can then be :ref:`bitcast and executed <int_trampoline>`.
10979 .. _int_mload_mstore:
10981 Masked Vector Load and Store Intrinsics
10982 ---------------------------------------
10984 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.
10988 '``llvm.masked.load.*``' Intrinsics
10989 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10993 This is an overloaded intrinsic. The loaded data is a vector of any integer or floating point data type.
10997 declare <16 x float> @llvm.masked.load.v16f32 (<16 x float>* <ptr>, i32 <alignment>, <16 x i1> <mask>, <16 x float> <passthru>)
10998 declare <2 x double> @llvm.masked.load.v2f64 (<2 x double>* <ptr>, i32 <alignment>, <2 x i1> <mask>, <2 x double> <passthru>)
11003 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.
11009 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.
11015 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.
11016 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.
11021 %res = call <16 x float> @llvm.masked.load.v16f32 (<16 x float>* %ptr, i32 4, <16 x i1>%mask, <16 x float> %passthru)
11023 ;; The result of the two following instructions is identical aside from potential memory access exception
11024 %loadlal = load <16 x float>, <16 x float>* %ptr, align 4
11025 %res = select <16 x i1> %mask, <16 x float> %loadlal, <16 x float> %passthru
11029 '``llvm.masked.store.*``' Intrinsics
11030 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11034 This is an overloaded intrinsic. The data stored in memory is a vector of any integer or floating point data type.
11038 declare void @llvm.masked.store.v8i32 (<8 x i32> <value>, <8 x i32> * <ptr>, i32 <alignment>, <8 x i1> <mask>)
11039 declare void @llvm.masked.store.v16f32(<16 x i32> <value>, <16 x i32>* <ptr>, i32 <alignment>, <16 x i1> <mask>)
11044 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.
11049 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.
11055 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.
11056 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.
11060 call void @llvm.masked.store.v16f32(<16 x float> %value, <16 x float>* %ptr, i32 4, <16 x i1> %mask)
11062 ;; The result of the following instructions is identical aside from potential data races and memory access exceptions
11063 %oldval = load <16 x float>, <16 x float>* %ptr, align 4
11064 %res = select <16 x i1> %mask, <16 x float> %value, <16 x float> %oldval
11065 store <16 x float> %res, <16 x float>* %ptr, align 4
11068 Masked Vector Gather and Scatter Intrinsics
11069 -------------------------------------------
11071 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.
11075 '``llvm.masked.gather.*``' Intrinsics
11076 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11080 This is an overloaded intrinsic. The loaded data are multiple scalar values of any integer or floating point data type gathered together into one vector.
11084 declare <16 x float> @llvm.masked.gather.v16f32 (<16 x float*> <ptrs>, i32 <alignment>, <16 x i1> <mask>, <16 x float> <passthru>)
11085 declare <2 x double> @llvm.masked.gather.v2f64 (<2 x double*> <ptrs>, i32 <alignment>, <2 x i1> <mask>, <2 x double> <passthru>)
11090 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.
11096 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.
11102 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.
11103 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.
11108 %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>)
11110 ;; The gather with all-true mask is equivalent to the following instruction sequence
11111 %ptr0 = extractelement <4 x double*> %ptrs, i32 0
11112 %ptr1 = extractelement <4 x double*> %ptrs, i32 1
11113 %ptr2 = extractelement <4 x double*> %ptrs, i32 2
11114 %ptr3 = extractelement <4 x double*> %ptrs, i32 3
11116 %val0 = load double, double* %ptr0, align 8
11117 %val1 = load double, double* %ptr1, align 8
11118 %val2 = load double, double* %ptr2, align 8
11119 %val3 = load double, double* %ptr3, align 8
11121 %vec0 = insertelement <4 x double>undef, %val0, 0
11122 %vec01 = insertelement <4 x double>%vec0, %val1, 1
11123 %vec012 = insertelement <4 x double>%vec01, %val2, 2
11124 %vec0123 = insertelement <4 x double>%vec012, %val3, 3
11128 '``llvm.masked.scatter.*``' Intrinsics
11129 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11133 This is an overloaded intrinsic. The data stored in memory is a vector of any integer or floating point data type. Each vector element is stored in an arbitrary memory addresses. Scatter with overlapping addresses is guaranteed to be ordered from least-significant to most-significant element.
11137 declare void @llvm.masked.scatter.v8i32 (<8 x i32> <value>, <8 x i32*> <ptrs>, i32 <alignment>, <8 x i1> <mask>)
11138 declare void @llvm.masked.scatter.v16f32(<16 x i32> <value>, <16 x i32*> <ptrs>, i32 <alignment>, <16 x i1> <mask>)
11143 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.
11148 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.
11154 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 divergency. Other targets may support this intrinsic differently, for example by lowering it into a sequence of branches that guard scalar store operations.
11158 ;; This instruction unconditionaly stores data vector in multiple addresses
11159 call @llvm.masked.scatter.v8i32 (<8 x i32> %value, <8 x i32*> %ptrs, i32 4, <8 x i1> <true, true, .. true>)
11161 ;; It is equivalent to a list of scalar stores
11162 %val0 = extractelement <8 x i32> %value, i32 0
11163 %val1 = extractelement <8 x i32> %value, i32 1
11165 %val7 = extractelement <8 x i32> %value, i32 7
11166 %ptr0 = extractelement <8 x i32*> %ptrs, i32 0
11167 %ptr1 = extractelement <8 x i32*> %ptrs, i32 1
11169 %ptr7 = extractelement <8 x i32*> %ptrs, i32 7
11170 ;; Note: the order of the following stores is important when they overlap:
11171 store i32 %val0, i32* %ptr0, align 4
11172 store i32 %val1, i32* %ptr1, align 4
11174 store i32 %val7, i32* %ptr7, align 4
11180 This class of intrinsics provides information about the lifetime of
11181 memory objects and ranges where variables are immutable.
11185 '``llvm.lifetime.start``' Intrinsic
11186 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11193 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
11198 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
11204 The first argument is a constant integer representing the size of the
11205 object, or -1 if it is variable sized. The second argument is a pointer
11211 This intrinsic indicates that before this point in the code, the value
11212 of the memory pointed to by ``ptr`` is dead. This means that it is known
11213 to never be used and has an undefined value. A load from the pointer
11214 that precedes this intrinsic can be replaced with ``'undef'``.
11218 '``llvm.lifetime.end``' Intrinsic
11219 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11226 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
11231 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
11237 The first argument is a constant integer representing the size of the
11238 object, or -1 if it is variable sized. The second argument is a pointer
11244 This intrinsic indicates that after this point in the code, the value of
11245 the memory pointed to by ``ptr`` is dead. This means that it is known to
11246 never be used and has an undefined value. Any stores into the memory
11247 object following this intrinsic may be removed as dead.
11249 '``llvm.invariant.start``' Intrinsic
11250 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11257 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
11262 The '``llvm.invariant.start``' intrinsic specifies that the contents of
11263 a memory object will not change.
11268 The first argument is a constant integer representing the size of the
11269 object, or -1 if it is variable sized. The second argument is a pointer
11275 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
11276 the return value, the referenced memory location is constant and
11279 '``llvm.invariant.end``' Intrinsic
11280 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11287 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
11292 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
11293 memory object are mutable.
11298 The first argument is the matching ``llvm.invariant.start`` intrinsic.
11299 The second argument is a constant integer representing the size of the
11300 object, or -1 if it is variable sized and the third argument is a
11301 pointer to the object.
11306 This intrinsic indicates that the memory is mutable again.
11311 This class of intrinsics is designed to be generic and has no specific
11314 '``llvm.var.annotation``' Intrinsic
11315 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11322 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
11327 The '``llvm.var.annotation``' intrinsic.
11332 The first argument is a pointer to a value, the second is a pointer to a
11333 global string, the third is a pointer to a global string which is the
11334 source file name, and the last argument is the line number.
11339 This intrinsic allows annotation of local variables with arbitrary
11340 strings. This can be useful for special purpose optimizations that want
11341 to look for these annotations. These have no other defined use; they are
11342 ignored by code generation and optimization.
11344 '``llvm.ptr.annotation.*``' Intrinsic
11345 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11350 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
11351 pointer to an integer of any width. *NOTE* you must specify an address space for
11352 the pointer. The identifier for the default address space is the integer
11357 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
11358 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
11359 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
11360 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
11361 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
11366 The '``llvm.ptr.annotation``' intrinsic.
11371 The first argument is a pointer to an integer value of arbitrary bitwidth
11372 (result of some expression), the second is a pointer to a global string, the
11373 third is a pointer to a global string which is the source file name, and the
11374 last argument is the line number. It returns the value of the first argument.
11379 This intrinsic allows annotation of a pointer to an integer with arbitrary
11380 strings. This can be useful for special purpose optimizations that want to look
11381 for these annotations. These have no other defined use; they are ignored by code
11382 generation and optimization.
11384 '``llvm.annotation.*``' Intrinsic
11385 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11390 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
11391 any integer bit width.
11395 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
11396 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
11397 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
11398 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
11399 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
11404 The '``llvm.annotation``' intrinsic.
11409 The first argument is an integer value (result of some expression), the
11410 second is a pointer to a global string, the third is a pointer to a
11411 global string which is the source file name, and the last argument is
11412 the line number. It returns the value of the first argument.
11417 This intrinsic allows annotations to be put on arbitrary expressions
11418 with arbitrary strings. This can be useful for special purpose
11419 optimizations that want to look for these annotations. These have no
11420 other defined use; they are ignored by code generation and optimization.
11422 '``llvm.trap``' Intrinsic
11423 ^^^^^^^^^^^^^^^^^^^^^^^^^
11430 declare void @llvm.trap() noreturn nounwind
11435 The '``llvm.trap``' intrinsic.
11445 This intrinsic is lowered to the target dependent trap instruction. If
11446 the target does not have a trap instruction, this intrinsic will be
11447 lowered to a call of the ``abort()`` function.
11449 '``llvm.debugtrap``' Intrinsic
11450 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11457 declare void @llvm.debugtrap() nounwind
11462 The '``llvm.debugtrap``' intrinsic.
11472 This intrinsic is lowered to code which is intended to cause an
11473 execution trap with the intention of requesting the attention of a
11476 '``llvm.stackprotector``' Intrinsic
11477 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11484 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
11489 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
11490 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
11491 is placed on the stack before local variables.
11496 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
11497 The first argument is the value loaded from the stack guard
11498 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
11499 enough space to hold the value of the guard.
11504 This intrinsic causes the prologue/epilogue inserter to force the position of
11505 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
11506 to ensure that if a local variable on the stack is overwritten, it will destroy
11507 the value of the guard. When the function exits, the guard on the stack is
11508 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
11509 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
11510 calling the ``__stack_chk_fail()`` function.
11512 '``llvm.stackprotectorcheck``' Intrinsic
11513 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11520 declare void @llvm.stackprotectorcheck(i8** <guard>)
11525 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
11526 created stack protector and if they are not equal calls the
11527 ``__stack_chk_fail()`` function.
11532 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
11533 the variable ``@__stack_chk_guard``.
11538 This intrinsic is provided to perform the stack protector check by comparing
11539 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
11540 values do not match call the ``__stack_chk_fail()`` function.
11542 The reason to provide this as an IR level intrinsic instead of implementing it
11543 via other IR operations is that in order to perform this operation at the IR
11544 level without an intrinsic, one would need to create additional basic blocks to
11545 handle the success/failure cases. This makes it difficult to stop the stack
11546 protector check from disrupting sibling tail calls in Codegen. With this
11547 intrinsic, we are able to generate the stack protector basic blocks late in
11548 codegen after the tail call decision has occurred.
11550 '``llvm.objectsize``' Intrinsic
11551 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11558 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
11559 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
11564 The ``llvm.objectsize`` intrinsic is designed to provide information to
11565 the optimizers to determine at compile time whether a) an operation
11566 (like memcpy) will overflow a buffer that corresponds to an object, or
11567 b) that a runtime check for overflow isn't necessary. An object in this
11568 context means an allocation of a specific class, structure, array, or
11574 The ``llvm.objectsize`` intrinsic takes two arguments. The first
11575 argument is a pointer to or into the ``object``. The second argument is
11576 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
11577 or -1 (if false) when the object size is unknown. The second argument
11578 only accepts constants.
11583 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
11584 the size of the object concerned. If the size cannot be determined at
11585 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
11586 on the ``min`` argument).
11588 '``llvm.expect``' Intrinsic
11589 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
11594 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
11599 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
11600 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
11601 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
11606 The ``llvm.expect`` intrinsic provides information about expected (the
11607 most probable) value of ``val``, which can be used by optimizers.
11612 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
11613 a value. The second argument is an expected value, this needs to be a
11614 constant value, variables are not allowed.
11619 This intrinsic is lowered to the ``val``.
11623 '``llvm.assume``' Intrinsic
11624 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11631 declare void @llvm.assume(i1 %cond)
11636 The ``llvm.assume`` allows the optimizer to assume that the provided
11637 condition is true. This information can then be used in simplifying other parts
11643 The condition which the optimizer may assume is always true.
11648 The intrinsic allows the optimizer to assume that the provided condition is
11649 always true whenever the control flow reaches the intrinsic call. No code is
11650 generated for this intrinsic, and instructions that contribute only to the
11651 provided condition are not used for code generation. If the condition is
11652 violated during execution, the behavior is undefined.
11654 Note that the optimizer might limit the transformations performed on values
11655 used by the ``llvm.assume`` intrinsic in order to preserve the instructions
11656 only used to form the intrinsic's input argument. This might prove undesirable
11657 if the extra information provided by the ``llvm.assume`` intrinsic does not cause
11658 sufficient overall improvement in code quality. For this reason,
11659 ``llvm.assume`` should not be used to document basic mathematical invariants
11660 that the optimizer can otherwise deduce or facts that are of little use to the
11665 '``llvm.bitset.test``' Intrinsic
11666 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11673 declare i1 @llvm.bitset.test(i8* %ptr, metadata %bitset) nounwind readnone
11679 The first argument is a pointer to be tested. The second argument is a
11680 metadata string containing the name of a :doc:`bitset <BitSets>`.
11685 The ``llvm.bitset.test`` intrinsic tests whether the given pointer is a
11686 member of the given bitset.
11688 '``llvm.donothing``' Intrinsic
11689 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11696 declare void @llvm.donothing() nounwind readnone
11701 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's one of only
11702 two intrinsics (besides ``llvm.experimental.patchpoint``) that can be called
11703 with an invoke instruction.
11713 This intrinsic does nothing, and it's removed by optimizers and ignored
11716 Stack Map Intrinsics
11717 --------------------
11719 LLVM provides experimental intrinsics to support runtime patching
11720 mechanisms commonly desired in dynamic language JITs. These intrinsics
11721 are described in :doc:`StackMaps`.