1 ==============================
2 LLVM Language Reference Manual
3 ==============================
12 This document is a reference manual for the LLVM assembly language. LLVM
13 is a Static Single Assignment (SSA) based representation that provides
14 type safety, low-level operations, flexibility, and the capability of
15 representing 'all' high-level languages cleanly. It is the common code
16 representation used throughout all phases of the LLVM compilation
22 The LLVM code representation is designed to be used in three different
23 forms: as an in-memory compiler IR, as an on-disk bitcode representation
24 (suitable for fast loading by a Just-In-Time compiler), and as a human
25 readable assembly language representation. This allows LLVM to provide a
26 powerful intermediate representation for efficient compiler
27 transformations and analysis, while providing a natural means to debug
28 and visualize the transformations. The three different forms of LLVM are
29 all equivalent. This document describes the human readable
30 representation and notation.
32 The LLVM representation aims to be light-weight and low-level while
33 being expressive, typed, and extensible at the same time. It aims to be
34 a "universal IR" of sorts, by being at a low enough level that
35 high-level ideas may be cleanly mapped to it (similar to how
36 microprocessors are "universal IR's", allowing many source languages to
37 be mapped to them). By providing type information, LLVM can be used as
38 the target of optimizations: for example, through pointer analysis, it
39 can be proven that a C automatic variable is never accessed outside of
40 the current function, allowing it to be promoted to a simple SSA value
41 instead of a memory location.
48 It is important to note that this document describes 'well formed' LLVM
49 assembly language. There is a difference between what the parser accepts
50 and what is considered 'well formed'. For example, the following
51 instruction is syntactically okay, but not well formed:
57 because the definition of ``%x`` does not dominate all of its uses. The
58 LLVM infrastructure provides a verification pass that may be used to
59 verify that an LLVM module is well formed. This pass is automatically
60 run by the parser after parsing input assembly and by the optimizer
61 before it outputs bitcode. The violations pointed out by the verifier
62 pass indicate bugs in transformation passes or input to the parser.
69 LLVM identifiers come in two basic types: global and local. Global
70 identifiers (functions, global variables) begin with the ``'@'``
71 character. Local identifiers (register names, types) begin with the
72 ``'%'`` character. Additionally, there are three different formats for
73 identifiers, for different purposes:
75 #. Named values are represented as a string of characters with their
76 prefix. For example, ``%foo``, ``@DivisionByZero``,
77 ``%a.really.long.identifier``. The actual regular expression used is
78 '``[%@][a-zA-Z$._][a-zA-Z$._0-9]*``'. Identifiers which require other
79 characters in their names can be surrounded with quotes. Special
80 characters may be escaped using ``"\xx"`` where ``xx`` is the ASCII
81 code for the character in hexadecimal. In this way, any character can
82 be used in a name value, even quotes themselves.
83 #. Unnamed values are represented as an unsigned numeric value with
84 their prefix. For example, ``%12``, ``@2``, ``%44``.
85 #. Constants, which are described in the section Constants_ below.
87 LLVM requires that values start with a prefix for two reasons: Compilers
88 don't need to worry about name clashes with reserved words, and the set
89 of reserved words may be expanded in the future without penalty.
90 Additionally, unnamed identifiers allow a compiler to quickly come up
91 with a temporary variable without having to avoid symbol table
94 Reserved words in LLVM are very similar to reserved words in other
95 languages. There are keywords for different opcodes ('``add``',
96 '``bitcast``', '``ret``', etc...), for primitive type names ('``void``',
97 '``i32``', etc...), and others. These reserved words cannot conflict
98 with variable names, because none of them start with a prefix character
101 Here is an example of LLVM code to multiply the integer variable
108 %result = mul i32 %X, 8
110 After strength reduction:
114 %result = shl i32 %X, 3
120 %0 = add i32 %X, %X ; yields i32:%0
121 %1 = add i32 %0, %0 ; yields i32:%1
122 %result = add i32 %1, %1
124 This last way of multiplying ``%X`` by 8 illustrates several important
125 lexical features of LLVM:
127 #. Comments are delimited with a '``;``' and go until the end of line.
128 #. Unnamed temporaries are created when the result of a computation is
129 not assigned to a named value.
130 #. Unnamed temporaries are numbered sequentially (using a per-function
131 incrementing counter, starting with 0). Note that basic blocks are
132 included in this numbering. For example, if the entry basic block is not
133 given a label name, then it will get number 0.
135 It also shows a convention that we follow in this document. When
136 demonstrating instructions, we will follow an instruction with a comment
137 that defines the type and name of value produced.
145 LLVM programs are composed of ``Module``'s, each of which is a
146 translation unit of the input programs. Each module consists of
147 functions, global variables, and symbol table entries. Modules may be
148 combined together with the LLVM linker, which merges function (and
149 global variable) definitions, resolves forward declarations, and merges
150 symbol table entries. Here is an example of the "hello world" module:
154 ; Declare the string constant as a global constant.
155 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
157 ; External declaration of the puts function
158 declare i32 @puts(i8* nocapture) nounwind
160 ; Definition of main function
161 define i32 @main() { ; i32()*
162 ; Convert [13 x i8]* to i8 *...
163 %cast210 = getelementptr [13 x i8]* @.str, i64 0, i64 0
165 ; Call puts function to write out the string to stdout.
166 call i32 @puts(i8* %cast210)
171 !1 = metadata !{i32 42}
174 This example is made up of a :ref:`global variable <globalvars>` named
175 "``.str``", an external declaration of the "``puts``" function, a
176 :ref:`function definition <functionstructure>` for "``main``" and
177 :ref:`named metadata <namedmetadatastructure>` "``foo``".
179 In general, a module is made up of a list of global values (where both
180 functions and global variables are global values). Global values are
181 represented by a pointer to a memory location (in this case, a pointer
182 to an array of char, and a pointer to a function), and have one of the
183 following :ref:`linkage types <linkage>`.
190 All Global Variables and Functions have one of the following types of
194 Global values with "``private``" linkage are only directly
195 accessible by objects in the current module. In particular, linking
196 code into a module with an private global value may cause the
197 private to be renamed as necessary to avoid collisions. Because the
198 symbol is private to the module, all references can be updated. This
199 doesn't show up in any symbol table in the object file.
201 Similar to private, but the value shows as a local symbol
202 (``STB_LOCAL`` in the case of ELF) in the object file. This
203 corresponds to the notion of the '``static``' keyword in C.
204 ``available_externally``
205 Globals with "``available_externally``" linkage are never emitted
206 into the object file corresponding to the LLVM module. They exist to
207 allow inlining and other optimizations to take place given knowledge
208 of the definition of the global, which is known to be somewhere
209 outside the module. Globals with ``available_externally`` linkage
210 are allowed to be discarded at will, and are otherwise the same as
211 ``linkonce_odr``. This linkage type is only allowed on definitions,
214 Globals with "``linkonce``" linkage are merged with other globals of
215 the same name when linkage occurs. This can be used to implement
216 some forms of inline functions, templates, or other code which must
217 be generated in each translation unit that uses it, but where the
218 body may be overridden with a more definitive definition later.
219 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
220 that ``linkonce`` linkage does not actually allow the optimizer to
221 inline the body of this function into callers because it doesn't
222 know if this definition of the function is the definitive definition
223 within the program or whether it will be overridden by a stronger
224 definition. To enable inlining and other optimizations, use
225 "``linkonce_odr``" linkage.
227 "``weak``" linkage has the same merging semantics as ``linkonce``
228 linkage, except that unreferenced globals with ``weak`` linkage may
229 not be discarded. This is used for globals that are declared "weak"
232 "``common``" linkage is most similar to "``weak``" linkage, but they
233 are used for tentative definitions in C, such as "``int X;``" at
234 global scope. Symbols with "``common``" linkage are merged in the
235 same way as ``weak symbols``, and they may not be deleted if
236 unreferenced. ``common`` symbols may not have an explicit section,
237 must have a zero initializer, and may not be marked
238 ':ref:`constant <globalvars>`'. Functions and aliases may not have
241 .. _linkage_appending:
244 "``appending``" linkage may only be applied to global variables of
245 pointer to array type. When two global variables with appending
246 linkage are linked together, the two global arrays are appended
247 together. This is the LLVM, typesafe, equivalent of having the
248 system linker append together "sections" with identical names when
251 The semantics of this linkage follow the ELF object file model: the
252 symbol is weak until linked, if not linked, the symbol becomes null
253 instead of being an undefined reference.
254 ``linkonce_odr``, ``weak_odr``
255 Some languages allow differing globals to be merged, such as two
256 functions with different semantics. Other languages, such as
257 ``C++``, ensure that only equivalent globals are ever merged (the
258 "one definition rule" --- "ODR"). Such languages can use the
259 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
260 global will only be merged with equivalent globals. These linkage
261 types are otherwise the same as their non-``odr`` versions.
263 If none of the above identifiers are used, the global is externally
264 visible, meaning that it participates in linkage and can be used to
265 resolve external symbol references.
267 It is illegal for a function *declaration* to have any linkage type
268 other than ``external`` or ``extern_weak``.
275 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
276 :ref:`invokes <i_invoke>` can all have an optional calling convention
277 specified for the call. The calling convention of any pair of dynamic
278 caller/callee must match, or the behavior of the program is undefined.
279 The following calling conventions are supported by LLVM, and more may be
282 "``ccc``" - The C calling convention
283 This calling convention (the default if no other calling convention
284 is specified) matches the target C calling conventions. This calling
285 convention supports varargs function calls and tolerates some
286 mismatch in the declared prototype and implemented declaration of
287 the function (as does normal C).
288 "``fastcc``" - The fast calling convention
289 This calling convention attempts to make calls as fast as possible
290 (e.g. by passing things in registers). This calling convention
291 allows the target to use whatever tricks it wants to produce fast
292 code for the target, without having to conform to an externally
293 specified ABI (Application Binary Interface). `Tail calls can only
294 be optimized when this, the GHC or the HiPE convention is
295 used. <CodeGenerator.html#id80>`_ This calling convention does not
296 support varargs and requires the prototype of all callees to exactly
297 match the prototype of the function definition.
298 "``coldcc``" - The cold calling convention
299 This calling convention attempts to make code in the caller as
300 efficient as possible under the assumption that the call is not
301 commonly executed. As such, these calls often preserve all registers
302 so that the call does not break any live ranges in the caller side.
303 This calling convention does not support varargs and requires the
304 prototype of all callees to exactly match the prototype of the
305 function definition. Furthermore the inliner doesn't consider such function
307 "``cc 10``" - GHC convention
308 This calling convention has been implemented specifically for use by
309 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
310 It passes everything in registers, going to extremes to achieve this
311 by disabling callee save registers. This calling convention should
312 not be used lightly but only for specific situations such as an
313 alternative to the *register pinning* performance technique often
314 used when implementing functional programming languages. At the
315 moment only X86 supports this convention and it has the following
318 - On *X86-32* only supports up to 4 bit type parameters. No
319 floating point types are supported.
320 - On *X86-64* only supports up to 10 bit type parameters and 6
321 floating point parameters.
323 This calling convention supports `tail call
324 optimization <CodeGenerator.html#id80>`_ but requires both the
325 caller and callee are using it.
326 "``cc 11``" - The HiPE calling convention
327 This calling convention has been implemented specifically for use by
328 the `High-Performance Erlang
329 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
330 native code compiler of the `Ericsson's Open Source Erlang/OTP
331 system <http://www.erlang.org/download.shtml>`_. It uses more
332 registers for argument passing than the ordinary C calling
333 convention and defines no callee-saved registers. The calling
334 convention properly supports `tail call
335 optimization <CodeGenerator.html#id80>`_ but requires that both the
336 caller and the callee use it. It uses a *register pinning*
337 mechanism, similar to GHC's convention, for keeping frequently
338 accessed runtime components pinned to specific hardware registers.
339 At the moment only X86 supports this convention (both 32 and 64
341 "``webkit_jscc``" - WebKit's JavaScript calling convention
342 This calling convention has been implemented for `WebKit FTL JIT
343 <https://trac.webkit.org/wiki/FTLJIT>`_. It passes arguments on the
344 stack right to left (as cdecl does), and returns a value in the
345 platform's customary return register.
346 "``anyregcc``" - Dynamic calling convention for code patching
347 This is a special convention that supports patching an arbitrary code
348 sequence in place of a call site. This convention forces the call
349 arguments into registers but allows them to be dynamcially
350 allocated. This can currently only be used with calls to
351 llvm.experimental.patchpoint because only this intrinsic records
352 the location of its arguments in a side table. See :doc:`StackMaps`.
353 "``preserve_mostcc``" - The `PreserveMost` calling convention
354 This calling convention attempts to make the code in the caller as little
355 intrusive as possible. This calling convention behaves identical to the `C`
356 calling convention on how arguments and return values are passed, but it
357 uses a different set of caller/callee-saved registers. This alleviates the
358 burden of saving and recovering a large register set before and after the
359 call in the caller. If the arguments are passed in callee-saved registers,
360 then they will be preserved by the callee across the call. This doesn't
361 apply for values returned in callee-saved registers.
363 - On X86-64 the callee preserves all general purpose registers, except for
364 R11. R11 can be used as a scratch register. Floating-point registers
365 (XMMs/YMMs) are not preserved and need to be saved by the caller.
367 The idea behind this convention is to support calls to runtime functions
368 that have a hot path and a cold path. The hot path is usually a small piece
369 of code that doesn't many registers. The cold path might need to call out to
370 another function and therefore only needs to preserve the caller-saved
371 registers, which haven't already been saved by the caller. The
372 `PreserveMost` calling convention is very similar to the `cold` calling
373 convention in terms of caller/callee-saved registers, but they are used for
374 different types of function calls. `coldcc` is for function calls that are
375 rarely executed, whereas `preserve_mostcc` function calls are intended to be
376 on the hot path and definitely executed a lot. Furthermore `preserve_mostcc`
377 doesn't prevent the inliner from inlining the function call.
379 This calling convention will be used by a future version of the ObjectiveC
380 runtime and should therefore still be considered experimental at this time.
381 Although this convention was created to optimize certain runtime calls to
382 the ObjectiveC runtime, it is not limited to this runtime and might be used
383 by other runtimes in the future too. The current implementation only
384 supports X86-64, but the intention is to support more architectures in the
386 "``preserve_allcc``" - The `PreserveAll` calling convention
387 This calling convention attempts to make the code in the caller even less
388 intrusive than the `PreserveMost` calling convention. This calling
389 convention also behaves identical to the `C` calling convention on how
390 arguments and return values are passed, but it uses a different set of
391 caller/callee-saved registers. This removes the burden of saving and
392 recovering a large register set before and after the call in the caller. If
393 the arguments are passed in callee-saved registers, then they will be
394 preserved by the callee across the call. This doesn't apply for values
395 returned in callee-saved registers.
397 - On X86-64 the callee preserves all general purpose registers, except for
398 R11. R11 can be used as a scratch register. Furthermore it also preserves
399 all floating-point registers (XMMs/YMMs).
401 The idea behind this convention is to support calls to runtime functions
402 that don't need to call out to any other functions.
404 This calling convention, like the `PreserveMost` calling convention, will be
405 used by a future version of the ObjectiveC runtime and should be considered
406 experimental at this time.
407 "``cc <n>``" - Numbered convention
408 Any calling convention may be specified by number, allowing
409 target-specific calling conventions to be used. Target specific
410 calling conventions start at 64.
412 More calling conventions can be added/defined on an as-needed basis, to
413 support Pascal conventions or any other well-known target-independent
416 .. _visibilitystyles:
421 All Global Variables and Functions have one of the following visibility
424 "``default``" - Default style
425 On targets that use the ELF object file format, default visibility
426 means that the declaration is visible to other modules and, in
427 shared libraries, means that the declared entity may be overridden.
428 On Darwin, default visibility means that the declaration is visible
429 to other modules. Default visibility corresponds to "external
430 linkage" in the language.
431 "``hidden``" - Hidden style
432 Two declarations of an object with hidden visibility refer to the
433 same object if they are in the same shared object. Usually, hidden
434 visibility indicates that the symbol will not be placed into the
435 dynamic symbol table, so no other module (executable or shared
436 library) can reference it directly.
437 "``protected``" - Protected style
438 On ELF, protected visibility indicates that the symbol will be
439 placed in the dynamic symbol table, but that references within the
440 defining module will bind to the local symbol. That is, the symbol
441 cannot be overridden by another module.
443 A symbol with ``internal`` or ``private`` linkage must have ``default``
451 All Global Variables, Functions and Aliases can have one of the following
455 "``dllimport``" causes the compiler to reference a function or variable via
456 a global pointer to a pointer that is set up by the DLL exporting the
457 symbol. On Microsoft Windows targets, the pointer name is formed by
458 combining ``__imp_`` and the function or variable name.
460 "``dllexport``" causes the compiler to provide a global pointer to a pointer
461 in a DLL, so that it can be referenced with the ``dllimport`` attribute. On
462 Microsoft Windows targets, the pointer name is formed by combining
463 ``__imp_`` and the function or variable name. Since this storage class
464 exists for defining a dll interface, the compiler, assembler and linker know
465 it is externally referenced and must refrain from deleting the symbol.
469 Thread Local Storage Models
470 ---------------------------
472 A variable may be defined as ``thread_local``, which means that it will
473 not be shared by threads (each thread will have a separated copy of the
474 variable). Not all targets support thread-local variables. Optionally, a
475 TLS model may be specified:
478 For variables that are only used within the current shared library.
480 For variables in modules that will not be loaded dynamically.
482 For variables defined in the executable and only used within it.
484 If no explicit model is given, the "general dynamic" model is used.
486 The models correspond to the ELF TLS models; see `ELF Handling For
487 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
488 more information on under which circumstances the different models may
489 be used. The target may choose a different TLS model if the specified
490 model is not supported, or if a better choice of model can be made.
492 A model can also be specified in a alias, but then it only governs how
493 the alias is accessed. It will not have any effect in the aliasee.
500 LLVM IR allows you to specify both "identified" and "literal" :ref:`structure
501 types <t_struct>`. Literal types are uniqued structurally, but identified types
502 are never uniqued. An :ref:`opaque structural type <t_opaque>` can also be used
503 to forward declare a type which is not yet available.
505 An example of a identified structure specification is:
509 %mytype = type { %mytype*, i32 }
511 Prior to the LLVM 3.0 release, identified types were structurally uniqued. Only
512 literal types are uniqued in recent versions of LLVM.
519 Global variables define regions of memory allocated at compilation time
522 Global variables definitions must be initialized.
524 Global variables in other translation units can also be declared, in which
525 case they don't have an initializer.
527 Either global variable definitions or declarations may have an explicit section
528 to be placed in and may have an optional explicit alignment specified.
530 A variable may be defined as a global ``constant``, which indicates that
531 the contents of the variable will **never** be modified (enabling better
532 optimization, allowing the global data to be placed in the read-only
533 section of an executable, etc). Note that variables that need runtime
534 initialization cannot be marked ``constant`` as there is a store to the
537 LLVM explicitly allows *declarations* of global variables to be marked
538 constant, even if the final definition of the global is not. This
539 capability can be used to enable slightly better optimization of the
540 program, but requires the language definition to guarantee that
541 optimizations based on the 'constantness' are valid for the translation
542 units that do not include the definition.
544 As SSA values, global variables define pointer values that are in scope
545 (i.e. they dominate) all basic blocks in the program. Global variables
546 always define a pointer to their "content" type because they describe a
547 region of memory, and all memory objects in LLVM are accessed through
550 Global variables can be marked with ``unnamed_addr`` which indicates
551 that the address is not significant, only the content. Constants marked
552 like this can be merged with other constants if they have the same
553 initializer. Note that a constant with significant address *can* be
554 merged with a ``unnamed_addr`` constant, the result being a constant
555 whose address is significant.
557 A global variable may be declared to reside in a target-specific
558 numbered address space. For targets that support them, address spaces
559 may affect how optimizations are performed and/or what target
560 instructions are used to access the variable. The default address space
561 is zero. The address space qualifier must precede any other attributes.
563 LLVM allows an explicit section to be specified for globals. If the
564 target supports it, it will emit globals to the section specified.
565 Additionally, the global can placed in a comdat if the target has the necessary
568 By default, global initializers are optimized by assuming that global
569 variables defined within the module are not modified from their
570 initial values before the start of the global initializer. This is
571 true even for variables potentially accessible from outside the
572 module, including those with external linkage or appearing in
573 ``@llvm.used`` or dllexported variables. This assumption may be suppressed
574 by marking the variable with ``externally_initialized``.
576 An explicit alignment may be specified for a global, which must be a
577 power of 2. If not present, or if the alignment is set to zero, the
578 alignment of the global is set by the target to whatever it feels
579 convenient. If an explicit alignment is specified, the global is forced
580 to have exactly that alignment. Targets and optimizers are not allowed
581 to over-align the global if the global has an assigned section. In this
582 case, the extra alignment could be observable: for example, code could
583 assume that the globals are densely packed in their section and try to
584 iterate over them as an array, alignment padding would break this
585 iteration. The maximum alignment is ``1 << 29``.
587 Globals can also have a :ref:`DLL storage class <dllstorageclass>`.
589 Variables and aliasaes can have a
590 :ref:`Thread Local Storage Model <tls_model>`.
594 [@<GlobalVarName> =] [Linkage] [Visibility] [DLLStorageClass] [ThreadLocal]
595 [unnamed_addr] [AddrSpace] [ExternallyInitialized]
596 <global | constant> <Type> [<InitializerConstant>]
597 [, section "name"] [, align <Alignment>]
599 For example, the following defines a global in a numbered address space
600 with an initializer, section, and alignment:
604 @G = addrspace(5) constant float 1.0, section "foo", align 4
606 The following example just declares a global variable
610 @G = external global i32
612 The following example defines a thread-local global with the
613 ``initialexec`` TLS model:
617 @G = thread_local(initialexec) global i32 0, align 4
619 .. _functionstructure:
624 LLVM function definitions consist of the "``define``" keyword, an
625 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
626 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
627 an optional :ref:`calling convention <callingconv>`,
628 an optional ``unnamed_addr`` attribute, a return type, an optional
629 :ref:`parameter attribute <paramattrs>` for the return type, a function
630 name, a (possibly empty) argument list (each with optional :ref:`parameter
631 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
632 an optional section, an optional alignment,
633 an optional :ref:`comdat <langref_comdats>`,
634 an optional :ref:`garbage collector name <gc>`, an optional :ref:`prefix <prefixdata>`, an opening
635 curly brace, a list of basic blocks, and a closing curly brace.
637 LLVM function declarations consist of the "``declare``" keyword, an
638 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
639 style <visibility>`, an optional :ref:`DLL storage class <dllstorageclass>`,
640 an optional :ref:`calling convention <callingconv>`,
641 an optional ``unnamed_addr`` attribute, a return type, an optional
642 :ref:`parameter attribute <paramattrs>` for the return type, a function
643 name, a possibly empty list of arguments, an optional alignment, an optional
644 :ref:`garbage collector name <gc>` and an optional :ref:`prefix <prefixdata>`.
646 A function definition contains a list of basic blocks, forming the CFG (Control
647 Flow Graph) for the function. Each basic block may optionally start with a label
648 (giving the basic block a symbol table entry), contains a list of instructions,
649 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
650 function return). If an explicit label is not provided, a block is assigned an
651 implicit numbered label, using the next value from the same counter as used for
652 unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
653 entry block does not have an explicit label, it will be assigned label "%0",
654 then the first unnamed temporary in that block will be "%1", etc.
656 The first basic block in a function is special in two ways: it is
657 immediately executed on entrance to the function, and it is not allowed
658 to have predecessor basic blocks (i.e. there can not be any branches to
659 the entry block of a function). Because the block can have no
660 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
662 LLVM allows an explicit section to be specified for functions. If the
663 target supports it, it will emit functions to the section specified.
664 Additionally, the function can placed in a COMDAT.
666 An explicit alignment may be specified for a function. If not present,
667 or if the alignment is set to zero, the alignment of the function is set
668 by the target to whatever it feels convenient. If an explicit alignment
669 is specified, the function is forced to have at least that much
670 alignment. All alignments must be a power of 2.
672 If the ``unnamed_addr`` attribute is given, the address is know to not
673 be significant and two identical functions can be merged.
677 define [linkage] [visibility] [DLLStorageClass]
679 <ResultType> @<FunctionName> ([argument list])
680 [unnamed_addr] [fn Attrs] [section "name"] [comdat $<ComdatName>]
681 [align N] [gc] [prefix Constant] { ... }
688 Aliases, unlike function or variables, don't create any new data. They
689 are just a new symbol and metadata for an existing position.
691 Aliases have a name and an aliasee that is either a global value or a
694 Aliases may have an optional :ref:`linkage type <linkage>`, an optional
695 :ref:`visibility style <visibility>`, an optional :ref:`DLL storage class
696 <dllstorageclass>` and an optional :ref:`tls model <tls_model>`.
700 @<Name> = [Visibility] [DLLStorageClass] [ThreadLocal] [unnamed_addr] alias [Linkage] <AliaseeTy> @<Aliasee>
702 The linkage must be one of ``private``, ``internal``, ``linkonce``, ``weak``,
703 ``linkonce_odr``, ``weak_odr``, ``external``. Note that some system linkers
704 might not correctly handle dropping a weak symbol that is aliased.
706 Alias that are not ``unnamed_addr`` are guaranteed to have the same address as
707 the aliasee expression. ``unnamed_addr`` ones are only guaranteed to point
710 Since aliases are only a second name, some restrictions apply, of which
711 some can only be checked when producing an object file:
713 * The expression defining the aliasee must be computable at assembly
714 time. Since it is just a name, no relocations can be used.
716 * No alias in the expression can be weak as the possibility of the
717 intermediate alias being overridden cannot be represented in an
720 * No global value in the expression can be a declaration, since that
721 would require a relocation, which is not possible.
728 Comdat IR provides access to COFF and ELF object file COMDAT functionality.
730 Comdats have a name which represents the COMDAT key. All global objects which
731 specify this key will only end up in the final object file if the linker chooses
732 that key over some other key. Aliases are placed in the same COMDAT that their
733 aliasee computes to, if any.
735 Comdats have a selection kind to provide input on how the linker should
736 choose between keys in two different object files.
740 $<Name> = comdat SelectionKind
742 The selection kind must be one of the following:
745 The linker may choose any COMDAT key, the choice is arbitrary.
747 The linker may choose any COMDAT key but the sections must contain the
750 The linker will choose the section containing the largest COMDAT key.
752 The linker requires that only section with this COMDAT key exist.
754 The linker may choose any COMDAT key but the sections must contain the
757 Note that the Mach-O platform doesn't support COMDATs and ELF only supports
758 ``any`` as a selection kind.
760 Here is an example of a COMDAT group where a function will only be selected if
761 the COMDAT key's section is the largest:
765 $foo = comdat largest
766 @foo = global i32 2, comdat $foo
768 define void @bar() comdat $foo {
772 In a COFF object file, this will create a COMDAT section with selection kind
773 ``IMAGE_COMDAT_SELECT_LARGEST`` containing the contents of the ``@foo`` symbol
774 and another COMDAT section with selection kind
775 ``IMAGE_COMDAT_SELECT_ASSOCIATIVE`` which is associated with the first COMDAT
776 section and contains the contents of the ``@baz`` symbol.
778 There are some restrictions on the properties of the global object.
779 It, or an alias to it, must have the same name as the COMDAT group when
781 The contents and size of this object may be used during link-time to determine
782 which COMDAT groups get selected depending on the selection kind.
783 Because the name of the object must match the name of the COMDAT group, the
784 linkage of the global object must not be local; local symbols can get renamed
785 if a collision occurs in the symbol table.
787 The combined use of COMDATS and section attributes may yield surprising results.
794 @g1 = global i32 42, section "sec", comdat $foo
795 @g2 = global i32 42, section "sec", comdat $bar
797 From the object file perspective, this requires the creation of two sections
798 with the same name. This is necessary because both globals belong to different
799 COMDAT groups and COMDATs, at the object file level, are represented by
802 Note that certain IR constructs like global variables and functions may create
803 COMDATs in the object file in addition to any which are specified using COMDAT
804 IR. This arises, for example, when a global variable has linkonce_odr linkage.
806 .. _namedmetadatastructure:
811 Named metadata is a collection of metadata. :ref:`Metadata
812 nodes <metadata>` (but not metadata strings) are the only valid
813 operands for a named metadata.
817 ; Some unnamed metadata nodes, which are referenced by the named metadata.
818 !0 = metadata !{metadata !"zero"}
819 !1 = metadata !{metadata !"one"}
820 !2 = metadata !{metadata !"two"}
822 !name = !{!0, !1, !2}
829 The return type and each parameter of a function type may have a set of
830 *parameter attributes* associated with them. Parameter attributes are
831 used to communicate additional information about the result or
832 parameters of a function. Parameter attributes are considered to be part
833 of the function, not of the function type, so functions with different
834 parameter attributes can have the same function type.
836 Parameter attributes are simple keywords that follow the type specified.
837 If multiple parameter attributes are needed, they are space separated.
842 declare i32 @printf(i8* noalias nocapture, ...)
843 declare i32 @atoi(i8 zeroext)
844 declare signext i8 @returns_signed_char()
846 Note that any attributes for the function result (``nounwind``,
847 ``readonly``) come immediately after the argument list.
849 Currently, only the following parameter attributes are defined:
852 This indicates to the code generator that the parameter or return
853 value should be zero-extended to the extent required by the target's
854 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
855 the caller (for a parameter) or the callee (for a return value).
857 This indicates to the code generator that the parameter or return
858 value should be sign-extended to the extent required by the target's
859 ABI (which is usually 32-bits) by the caller (for a parameter) or
860 the callee (for a return value).
862 This indicates that this parameter or return value should be treated
863 in a special target-dependent fashion during while emitting code for
864 a function call or return (usually, by putting it in a register as
865 opposed to memory, though some targets use it to distinguish between
866 two different kinds of registers). Use of this attribute is
869 This indicates that the pointer parameter should really be passed by
870 value to the function. The attribute implies that a hidden copy of
871 the pointee is made between the caller and the callee, so the callee
872 is unable to modify the value in the caller. This attribute is only
873 valid on LLVM pointer arguments. It is generally used to pass
874 structs and arrays by value, but is also valid on pointers to
875 scalars. The copy is considered to belong to the caller not the
876 callee (for example, ``readonly`` functions should not write to
877 ``byval`` parameters). This is not a valid attribute for return
880 The byval attribute also supports specifying an alignment with the
881 align attribute. It indicates the alignment of the stack slot to
882 form and the known alignment of the pointer specified to the call
883 site. If the alignment is not specified, then the code generator
884 makes a target-specific assumption.
890 The ``inalloca`` argument attribute allows the caller to take the
891 address of outgoing stack arguments. An ``inalloca`` argument must
892 be a pointer to stack memory produced by an ``alloca`` instruction.
893 The alloca, or argument allocation, must also be tagged with the
894 inalloca keyword. Only the last argument may have the ``inalloca``
895 attribute, and that argument is guaranteed to be passed in memory.
897 An argument allocation may be used by a call at most once because
898 the call may deallocate it. The ``inalloca`` attribute cannot be
899 used in conjunction with other attributes that affect argument
900 storage, like ``inreg``, ``nest``, ``sret``, or ``byval``. The
901 ``inalloca`` attribute also disables LLVM's implicit lowering of
902 large aggregate return values, which means that frontend authors
903 must lower them with ``sret`` pointers.
905 When the call site is reached, the argument allocation must have
906 been the most recent stack allocation that is still live, or the
907 results are undefined. It is possible to allocate additional stack
908 space after an argument allocation and before its call site, but it
909 must be cleared off with :ref:`llvm.stackrestore
912 See :doc:`InAlloca` for more information on how to use this
916 This indicates that the pointer parameter specifies the address of a
917 structure that is the return value of the function in the source
918 program. This pointer must be guaranteed by the caller to be valid:
919 loads and stores to the structure may be assumed by the callee
920 not to trap and to be properly aligned. This may only be applied to
921 the first parameter. This is not a valid attribute for return
925 This indicates that the pointer value may be assumed by the optimizer to
926 have the specified alignment.
928 Note that this attribute has additional semantics when combined with the
934 This indicates that pointer values :ref:`based <pointeraliasing>` on
935 the argument or return value do not alias pointer values which are
936 not *based* on it, ignoring certain "irrelevant" dependencies. For a
937 call to the parent function, dependencies between memory references
938 from before or after the call and from those during the call are
939 "irrelevant" to the ``noalias`` keyword for the arguments and return
940 value used in that call. The caller shares the responsibility with
941 the callee for ensuring that these requirements are met. For further
942 details, please see the discussion of the NoAlias response in :ref:`alias
943 analysis <Must, May, or No>`.
945 Note that this definition of ``noalias`` is intentionally similar
946 to the definition of ``restrict`` in C99 for function arguments,
947 though it is slightly weaker.
949 For function return values, C99's ``restrict`` is not meaningful,
950 while LLVM's ``noalias`` is.
952 This indicates that the callee does not make any copies of the
953 pointer that outlive the callee itself. This is not a valid
954 attribute for return values.
959 This indicates that the pointer parameter can be excised using the
960 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
961 attribute for return values and can only be applied to one parameter.
964 This indicates that the function always returns the argument as its return
965 value. This is an optimization hint to the code generator when generating
966 the caller, allowing tail call optimization and omission of register saves
967 and restores in some cases; it is not checked or enforced when generating
968 the callee. The parameter and the function return type must be valid
969 operands for the :ref:`bitcast instruction <i_bitcast>`. This is not a
970 valid attribute for return values and can only be applied to one parameter.
973 This indicates that the parameter or return pointer is not null. This
974 attribute may only be applied to pointer typed parameters. This is not
975 checked or enforced by LLVM, the caller must ensure that the pointer
976 passed in is non-null, or the callee must ensure that the returned pointer
979 ``dereferenceable(<n>)``
980 This indicates that the parameter or return pointer is dereferenceable. This
981 attribute may only be applied to pointer typed parameters. A pointer that
982 is dereferenceable can be loaded from speculatively without a risk of
983 trapping. The number of bytes known to be dereferenceable must be provided
984 in parentheses. It is legal for the number of bytes to be less than the
985 size of the pointee type. The ``nonnull`` attribute does not imply
986 dereferenceability (consider a pointer to one element past the end of an
987 array), however ``dereferenceable(<n>)`` does imply ``nonnull`` in
988 ``addrspace(0)`` (which is the default address space).
992 Garbage Collector Names
993 -----------------------
995 Each function may specify a garbage collector name, which is simply a
1000 define void @f() gc "name" { ... }
1002 The compiler declares the supported values of *name*. Specifying a
1003 collector which will cause the compiler to alter its output in order to
1004 support the named garbage collection algorithm.
1011 Prefix data is data associated with a function which the code generator
1012 will emit immediately before the function body. The purpose of this feature
1013 is to allow frontends to associate language-specific runtime metadata with
1014 specific functions and make it available through the function pointer while
1015 still allowing the function pointer to be called. To access the data for a
1016 given function, a program may bitcast the function pointer to a pointer to
1017 the constant's type. This implies that the IR symbol points to the start
1020 To maintain the semantics of ordinary function calls, the prefix data must
1021 have a particular format. Specifically, it must begin with a sequence of
1022 bytes which decode to a sequence of machine instructions, valid for the
1023 module's target, which transfer control to the point immediately succeeding
1024 the prefix data, without performing any other visible action. This allows
1025 the inliner and other passes to reason about the semantics of the function
1026 definition without needing to reason about the prefix data. Obviously this
1027 makes the format of the prefix data highly target dependent.
1029 Prefix data is laid out as if it were an initializer for a global variable
1030 of the prefix data's type. No padding is automatically placed between the
1031 prefix data and the function body. If padding is required, it must be part
1034 A trivial example of valid prefix data for the x86 architecture is ``i8 144``,
1035 which encodes the ``nop`` instruction:
1037 .. code-block:: llvm
1039 define void @f() prefix i8 144 { ... }
1041 Generally prefix data can be formed by encoding a relative branch instruction
1042 which skips the metadata, as in this example of valid prefix data for the
1043 x86_64 architecture, where the first two bytes encode ``jmp .+10``:
1045 .. code-block:: llvm
1047 %0 = type <{ i8, i8, i8* }>
1049 define void @f() prefix %0 <{ i8 235, i8 8, i8* @md}> { ... }
1051 A function may have prefix data but no body. This has similar semantics
1052 to the ``available_externally`` linkage in that the data may be used by the
1053 optimizers but will not be emitted in the object file.
1060 Attribute groups are groups of attributes that are referenced by objects within
1061 the IR. They are important for keeping ``.ll`` files readable, because a lot of
1062 functions will use the same set of attributes. In the degenerative case of a
1063 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
1064 group will capture the important command line flags used to build that file.
1066 An attribute group is a module-level object. To use an attribute group, an
1067 object references the attribute group's ID (e.g. ``#37``). An object may refer
1068 to more than one attribute group. In that situation, the attributes from the
1069 different groups are merged.
1071 Here is an example of attribute groups for a function that should always be
1072 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
1074 .. code-block:: llvm
1076 ; Target-independent attributes:
1077 attributes #0 = { alwaysinline alignstack=4 }
1079 ; Target-dependent attributes:
1080 attributes #1 = { "no-sse" }
1082 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
1083 define void @f() #0 #1 { ... }
1090 Function attributes are set to communicate additional information about
1091 a function. Function attributes are considered to be part of the
1092 function, not of the function type, so functions with different function
1093 attributes can have the same function type.
1095 Function attributes are simple keywords that follow the type specified.
1096 If multiple attributes are needed, they are space separated. For
1099 .. code-block:: llvm
1101 define void @f() noinline { ... }
1102 define void @f() alwaysinline { ... }
1103 define void @f() alwaysinline optsize { ... }
1104 define void @f() optsize { ... }
1107 This attribute indicates that, when emitting the prologue and
1108 epilogue, the backend should forcibly align the stack pointer.
1109 Specify the desired alignment, which must be a power of two, in
1112 This attribute indicates that the inliner should attempt to inline
1113 this function into callers whenever possible, ignoring any active
1114 inlining size threshold for this caller.
1116 This indicates that the callee function at a call site should be
1117 recognized as a built-in function, even though the function's declaration
1118 uses the ``nobuiltin`` attribute. This is only valid at call sites for
1119 direct calls to functions which are declared with the ``nobuiltin``
1122 This attribute indicates that this function is rarely called. When
1123 computing edge weights, basic blocks post-dominated by a cold
1124 function call are also considered to be cold; and, thus, given low
1127 This attribute indicates that the source code contained a hint that
1128 inlining this function is desirable (such as the "inline" keyword in
1129 C/C++). It is just a hint; it imposes no requirements on the
1132 This attribute indicates that the function should be added to a
1133 jump-instruction table at code-generation time, and that all address-taken
1134 references to this function should be replaced with a reference to the
1135 appropriate jump-instruction-table function pointer. Note that this creates
1136 a new pointer for the original function, which means that code that depends
1137 on function-pointer identity can break. So, any function annotated with
1138 ``jumptable`` must also be ``unnamed_addr``.
1140 This attribute suggests that optimization passes and code generator
1141 passes make choices that keep the code size of this function as small
1142 as possible and perform optimizations that may sacrifice runtime
1143 performance in order to minimize the size of the generated code.
1145 This attribute disables prologue / epilogue emission for the
1146 function. This can have very system-specific consequences.
1148 This indicates that the callee function at a call site is not recognized as
1149 a built-in function. LLVM will retain the original call and not replace it
1150 with equivalent code based on the semantics of the built-in function, unless
1151 the call site uses the ``builtin`` attribute. This is valid at call sites
1152 and on function declarations and definitions.
1154 This attribute indicates that calls to the function cannot be
1155 duplicated. A call to a ``noduplicate`` function may be moved
1156 within its parent function, but may not be duplicated within
1157 its parent function.
1159 A function containing a ``noduplicate`` call may still
1160 be an inlining candidate, provided that the call is not
1161 duplicated by inlining. That implies that the function has
1162 internal linkage and only has one call site, so the original
1163 call is dead after inlining.
1165 This attributes disables implicit floating point instructions.
1167 This attribute indicates that the inliner should never inline this
1168 function in any situation. This attribute may not be used together
1169 with the ``alwaysinline`` attribute.
1171 This attribute suppresses lazy symbol binding for the function. This
1172 may make calls to the function faster, at the cost of extra program
1173 startup time if the function is not called during program startup.
1175 This attribute indicates that the code generator should not use a
1176 red zone, even if the target-specific ABI normally permits it.
1178 This function attribute indicates that the function never returns
1179 normally. This produces undefined behavior at runtime if the
1180 function ever does dynamically return.
1182 This function attribute indicates that the function never returns
1183 with an unwind or exceptional control flow. If the function does
1184 unwind, its runtime behavior is undefined.
1186 This function attribute indicates that the function is not optimized
1187 by any optimization or code generator passes with the
1188 exception of interprocedural optimization passes.
1189 This attribute cannot be used together with the ``alwaysinline``
1190 attribute; this attribute is also incompatible
1191 with the ``minsize`` attribute and the ``optsize`` attribute.
1193 This attribute requires the ``noinline`` attribute to be specified on
1194 the function as well, so the function is never inlined into any caller.
1195 Only functions with the ``alwaysinline`` attribute are valid
1196 candidates for inlining into the body of this function.
1198 This attribute suggests that optimization passes and code generator
1199 passes make choices that keep the code size of this function low,
1200 and otherwise do optimizations specifically to reduce code size as
1201 long as they do not significantly impact runtime performance.
1203 On a function, this attribute indicates that the function computes its
1204 result (or decides to unwind an exception) based strictly on its arguments,
1205 without dereferencing any pointer arguments or otherwise accessing
1206 any mutable state (e.g. memory, control registers, etc) visible to
1207 caller functions. It does not write through any pointer arguments
1208 (including ``byval`` arguments) and never changes any state visible
1209 to callers. This means that it cannot unwind exceptions by calling
1210 the ``C++`` exception throwing methods.
1212 On an argument, this attribute indicates that the function does not
1213 dereference that pointer argument, even though it may read or write the
1214 memory that the pointer points to if accessed through other pointers.
1216 On a function, this attribute indicates that the function does not write
1217 through any pointer arguments (including ``byval`` arguments) or otherwise
1218 modify any state (e.g. memory, control registers, etc) visible to
1219 caller functions. It may dereference pointer arguments and read
1220 state that may be set in the caller. A readonly function always
1221 returns the same value (or unwinds an exception identically) when
1222 called with the same set of arguments and global state. It cannot
1223 unwind an exception by calling the ``C++`` exception throwing
1226 On an argument, this attribute indicates that the function does not write
1227 through this pointer argument, even though it may write to the memory that
1228 the pointer points to.
1230 This attribute indicates that this function can return twice. The C
1231 ``setjmp`` is an example of such a function. The compiler disables
1232 some optimizations (like tail calls) in the caller of these
1234 ``sanitize_address``
1235 This attribute indicates that AddressSanitizer checks
1236 (dynamic address safety analysis) are enabled for this function.
1238 This attribute indicates that MemorySanitizer checks (dynamic detection
1239 of accesses to uninitialized memory) are enabled for this function.
1241 This attribute indicates that ThreadSanitizer checks
1242 (dynamic thread safety analysis) are enabled for this function.
1244 This attribute indicates that the function should emit a stack
1245 smashing protector. It is in the form of a "canary" --- a random value
1246 placed on the stack before the local variables that's checked upon
1247 return from the function to see if it has been overwritten. A
1248 heuristic is used to determine if a function needs stack protectors
1249 or not. The heuristic used will enable protectors for functions with:
1251 - Character arrays larger than ``ssp-buffer-size`` (default 8).
1252 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
1253 - Calls to alloca() with variable sizes or constant sizes greater than
1254 ``ssp-buffer-size``.
1256 Variables that are identified as requiring a protector will be arranged
1257 on the stack such that they are adjacent to the stack protector guard.
1259 If a function that has an ``ssp`` attribute is inlined into a
1260 function that doesn't have an ``ssp`` attribute, then the resulting
1261 function will have an ``ssp`` attribute.
1263 This attribute indicates that the function should *always* emit a
1264 stack smashing protector. This overrides the ``ssp`` function
1267 Variables that are identified as requiring a protector will be arranged
1268 on the stack such that they are adjacent to the stack protector guard.
1269 The specific layout rules are:
1271 #. Large arrays and structures containing large arrays
1272 (``>= ssp-buffer-size``) are closest to the stack protector.
1273 #. Small arrays and structures containing small arrays
1274 (``< ssp-buffer-size``) are 2nd closest to the protector.
1275 #. Variables that have had their address taken are 3rd closest to the
1278 If a function that has an ``sspreq`` attribute is inlined into a
1279 function that doesn't have an ``sspreq`` attribute or which has an
1280 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
1281 an ``sspreq`` attribute.
1283 This attribute indicates that the function should emit a stack smashing
1284 protector. This attribute causes a strong heuristic to be used when
1285 determining if a function needs stack protectors. The strong heuristic
1286 will enable protectors for functions with:
1288 - Arrays of any size and type
1289 - Aggregates containing an array of any size and type.
1290 - Calls to alloca().
1291 - Local variables that have had their address taken.
1293 Variables that are identified as requiring a protector will be arranged
1294 on the stack such that they are adjacent to the stack protector guard.
1295 The specific layout rules are:
1297 #. Large arrays and structures containing large arrays
1298 (``>= ssp-buffer-size``) are closest to the stack protector.
1299 #. Small arrays and structures containing small arrays
1300 (``< ssp-buffer-size``) are 2nd closest to the protector.
1301 #. Variables that have had their address taken are 3rd closest to the
1304 This overrides the ``ssp`` function attribute.
1306 If a function that has an ``sspstrong`` attribute is inlined into a
1307 function that doesn't have an ``sspstrong`` attribute, then the
1308 resulting function will have an ``sspstrong`` attribute.
1310 This attribute indicates that the ABI being targeted requires that
1311 an unwind table entry be produce for this function even if we can
1312 show that no exceptions passes by it. This is normally the case for
1313 the ELF x86-64 abi, but it can be disabled for some compilation
1318 Module-Level Inline Assembly
1319 ----------------------------
1321 Modules may contain "module-level inline asm" blocks, which corresponds
1322 to the GCC "file scope inline asm" blocks. These blocks are internally
1323 concatenated by LLVM and treated as a single unit, but may be separated
1324 in the ``.ll`` file if desired. The syntax is very simple:
1326 .. code-block:: llvm
1328 module asm "inline asm code goes here"
1329 module asm "more can go here"
1331 The strings can contain any character by escaping non-printable
1332 characters. The escape sequence used is simply "\\xx" where "xx" is the
1333 two digit hex code for the number.
1335 The inline asm code is simply printed to the machine code .s file when
1336 assembly code is generated.
1338 .. _langref_datalayout:
1343 A module may specify a target specific data layout string that specifies
1344 how data is to be laid out in memory. The syntax for the data layout is
1347 .. code-block:: llvm
1349 target datalayout = "layout specification"
1351 The *layout specification* consists of a list of specifications
1352 separated by the minus sign character ('-'). Each specification starts
1353 with a letter and may include other information after the letter to
1354 define some aspect of the data layout. The specifications accepted are
1358 Specifies that the target lays out data in big-endian form. That is,
1359 the bits with the most significance have the lowest address
1362 Specifies that the target lays out data in little-endian form. That
1363 is, the bits with the least significance have the lowest address
1366 Specifies the natural alignment of the stack in bits. Alignment
1367 promotion of stack variables is limited to the natural stack
1368 alignment to avoid dynamic stack realignment. The stack alignment
1369 must be a multiple of 8-bits. If omitted, the natural stack
1370 alignment defaults to "unspecified", which does not prevent any
1371 alignment promotions.
1372 ``p[n]:<size>:<abi>:<pref>``
1373 This specifies the *size* of a pointer and its ``<abi>`` and
1374 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1375 bits. The address space, ``n`` is optional, and if not specified,
1376 denotes the default address space 0. The value of ``n`` must be
1377 in the range [1,2^23).
1378 ``i<size>:<abi>:<pref>``
1379 This specifies the alignment for an integer type of a given bit
1380 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1381 ``v<size>:<abi>:<pref>``
1382 This specifies the alignment for a vector type of a given bit
1384 ``f<size>:<abi>:<pref>``
1385 This specifies the alignment for a floating point type of a given bit
1386 ``<size>``. Only values of ``<size>`` that are supported by the target
1387 will work. 32 (float) and 64 (double) are supported on all targets; 80
1388 or 128 (different flavors of long double) are also supported on some
1391 This specifies the alignment for an object of aggregate type.
1393 If present, specifies that llvm names are mangled in the output. The
1396 * ``e``: ELF mangling: Private symbols get a ``.L`` prefix.
1397 * ``m``: Mips mangling: Private symbols get a ``$`` prefix.
1398 * ``o``: Mach-O mangling: Private symbols get ``L`` prefix. Other
1399 symbols get a ``_`` prefix.
1400 * ``w``: Windows COFF prefix: Similar to Mach-O, but stdcall and fastcall
1401 functions also get a suffix based on the frame size.
1402 ``n<size1>:<size2>:<size3>...``
1403 This specifies a set of native integer widths for the target CPU in
1404 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1405 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1406 this set are considered to support most general arithmetic operations
1409 On every specification that takes a ``<abi>:<pref>``, specifying the
1410 ``<pref>`` alignment is optional. If omitted, the preceding ``:``
1411 should be omitted too and ``<pref>`` will be equal to ``<abi>``.
1413 When constructing the data layout for a given target, LLVM starts with a
1414 default set of specifications which are then (possibly) overridden by
1415 the specifications in the ``datalayout`` keyword. The default
1416 specifications are given in this list:
1418 - ``E`` - big endian
1419 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment.
1420 - ``p[n]:64:64:64`` - Other address spaces are assumed to be the
1421 same as the default address space.
1422 - ``S0`` - natural stack alignment is unspecified
1423 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1424 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1425 - ``i16:16:16`` - i16 is 16-bit aligned
1426 - ``i32:32:32`` - i32 is 32-bit aligned
1427 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1428 alignment of 64-bits
1429 - ``f16:16:16`` - half is 16-bit aligned
1430 - ``f32:32:32`` - float is 32-bit aligned
1431 - ``f64:64:64`` - double is 64-bit aligned
1432 - ``f128:128:128`` - quad is 128-bit aligned
1433 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1434 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1435 - ``a:0:64`` - aggregates are 64-bit aligned
1437 When LLVM is determining the alignment for a given type, it uses the
1440 #. If the type sought is an exact match for one of the specifications,
1441 that specification is used.
1442 #. If no match is found, and the type sought is an integer type, then
1443 the smallest integer type that is larger than the bitwidth of the
1444 sought type is used. If none of the specifications are larger than
1445 the bitwidth then the largest integer type is used. For example,
1446 given the default specifications above, the i7 type will use the
1447 alignment of i8 (next largest) while both i65 and i256 will use the
1448 alignment of i64 (largest specified).
1449 #. If no match is found, and the type sought is a vector type, then the
1450 largest vector type that is smaller than the sought vector type will
1451 be used as a fall back. This happens because <128 x double> can be
1452 implemented in terms of 64 <2 x double>, for example.
1454 The function of the data layout string may not be what you expect.
1455 Notably, this is not a specification from the frontend of what alignment
1456 the code generator should use.
1458 Instead, if specified, the target data layout is required to match what
1459 the ultimate *code generator* expects. This string is used by the
1460 mid-level optimizers to improve code, and this only works if it matches
1461 what the ultimate code generator uses. If you would like to generate IR
1462 that does not embed this target-specific detail into the IR, then you
1463 don't have to specify the string. This will disable some optimizations
1464 that require precise layout information, but this also prevents those
1465 optimizations from introducing target specificity into the IR.
1472 A module may specify a target triple string that describes the target
1473 host. The syntax for the target triple is simply:
1475 .. code-block:: llvm
1477 target triple = "x86_64-apple-macosx10.7.0"
1479 The *target triple* string consists of a series of identifiers delimited
1480 by the minus sign character ('-'). The canonical forms are:
1484 ARCHITECTURE-VENDOR-OPERATING_SYSTEM
1485 ARCHITECTURE-VENDOR-OPERATING_SYSTEM-ENVIRONMENT
1487 This information is passed along to the backend so that it generates
1488 code for the proper architecture. It's possible to override this on the
1489 command line with the ``-mtriple`` command line option.
1491 .. _pointeraliasing:
1493 Pointer Aliasing Rules
1494 ----------------------
1496 Any memory access must be done through a pointer value associated with
1497 an address range of the memory access, otherwise the behavior is
1498 undefined. Pointer values are associated with address ranges according
1499 to the following rules:
1501 - A pointer value is associated with the addresses associated with any
1502 value it is *based* on.
1503 - An address of a global variable is associated with the address range
1504 of the variable's storage.
1505 - The result value of an allocation instruction is associated with the
1506 address range of the allocated storage.
1507 - A null pointer in the default address-space is associated with no
1509 - An integer constant other than zero or a pointer value returned from
1510 a function not defined within LLVM may be associated with address
1511 ranges allocated through mechanisms other than those provided by
1512 LLVM. Such ranges shall not overlap with any ranges of addresses
1513 allocated by mechanisms provided by LLVM.
1515 A pointer value is *based* on another pointer value according to the
1518 - A pointer value formed from a ``getelementptr`` operation is *based*
1519 on the first operand of the ``getelementptr``.
1520 - The result value of a ``bitcast`` is *based* on the operand of the
1522 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1523 values that contribute (directly or indirectly) to the computation of
1524 the pointer's value.
1525 - The "*based* on" relationship is transitive.
1527 Note that this definition of *"based"* is intentionally similar to the
1528 definition of *"based"* in C99, though it is slightly weaker.
1530 LLVM IR does not associate types with memory. The result type of a
1531 ``load`` merely indicates the size and alignment of the memory from
1532 which to load, as well as the interpretation of the value. The first
1533 operand type of a ``store`` similarly only indicates the size and
1534 alignment of the store.
1536 Consequently, type-based alias analysis, aka TBAA, aka
1537 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1538 :ref:`Metadata <metadata>` may be used to encode additional information
1539 which specialized optimization passes may use to implement type-based
1544 Volatile Memory Accesses
1545 ------------------------
1547 Certain memory accesses, such as :ref:`load <i_load>`'s,
1548 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1549 marked ``volatile``. The optimizers must not change the number of
1550 volatile operations or change their order of execution relative to other
1551 volatile operations. The optimizers *may* change the order of volatile
1552 operations relative to non-volatile operations. This is not Java's
1553 "volatile" and has no cross-thread synchronization behavior.
1555 IR-level volatile loads and stores cannot safely be optimized into
1556 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1557 flagged volatile. Likewise, the backend should never split or merge
1558 target-legal volatile load/store instructions.
1560 .. admonition:: Rationale
1562 Platforms may rely on volatile loads and stores of natively supported
1563 data width to be executed as single instruction. For example, in C
1564 this holds for an l-value of volatile primitive type with native
1565 hardware support, but not necessarily for aggregate types. The
1566 frontend upholds these expectations, which are intentionally
1567 unspecified in the IR. The rules above ensure that IR transformation
1568 do not violate the frontend's contract with the language.
1572 Memory Model for Concurrent Operations
1573 --------------------------------------
1575 The LLVM IR does not define any way to start parallel threads of
1576 execution or to register signal handlers. Nonetheless, there are
1577 platform-specific ways to create them, and we define LLVM IR's behavior
1578 in their presence. This model is inspired by the C++0x memory model.
1580 For a more informal introduction to this model, see the :doc:`Atomics`.
1582 We define a *happens-before* partial order as the least partial order
1585 - Is a superset of single-thread program order, and
1586 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1587 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1588 techniques, like pthread locks, thread creation, thread joining,
1589 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1590 Constraints <ordering>`).
1592 Note that program order does not introduce *happens-before* edges
1593 between a thread and signals executing inside that thread.
1595 Every (defined) read operation (load instructions, memcpy, atomic
1596 loads/read-modify-writes, etc.) R reads a series of bytes written by
1597 (defined) write operations (store instructions, atomic
1598 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1599 section, initialized globals are considered to have a write of the
1600 initializer which is atomic and happens before any other read or write
1601 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1602 may see any write to the same byte, except:
1604 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1605 write\ :sub:`2` happens before R\ :sub:`byte`, then
1606 R\ :sub:`byte` does not see write\ :sub:`1`.
1607 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1608 R\ :sub:`byte` does not see write\ :sub:`3`.
1610 Given that definition, R\ :sub:`byte` is defined as follows:
1612 - If R is volatile, the result is target-dependent. (Volatile is
1613 supposed to give guarantees which can support ``sig_atomic_t`` in
1614 C/C++, and may be used for accesses to addresses which do not behave
1615 like normal memory. It does not generally provide cross-thread
1617 - Otherwise, if there is no write to the same byte that happens before
1618 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1619 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1620 R\ :sub:`byte` returns the value written by that write.
1621 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1622 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1623 Memory Ordering Constraints <ordering>` section for additional
1624 constraints on how the choice is made.
1625 - Otherwise R\ :sub:`byte` returns ``undef``.
1627 R returns the value composed of the series of bytes it read. This
1628 implies that some bytes within the value may be ``undef`` **without**
1629 the entire value being ``undef``. Note that this only defines the
1630 semantics of the operation; it doesn't mean that targets will emit more
1631 than one instruction to read the series of bytes.
1633 Note that in cases where none of the atomic intrinsics are used, this
1634 model places only one restriction on IR transformations on top of what
1635 is required for single-threaded execution: introducing a store to a byte
1636 which might not otherwise be stored is not allowed in general.
1637 (Specifically, in the case where another thread might write to and read
1638 from an address, introducing a store can change a load that may see
1639 exactly one write into a load that may see multiple writes.)
1643 Atomic Memory Ordering Constraints
1644 ----------------------------------
1646 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1647 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1648 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1649 ordering parameters that determine which other atomic instructions on
1650 the same address they *synchronize with*. These semantics are borrowed
1651 from Java and C++0x, but are somewhat more colloquial. If these
1652 descriptions aren't precise enough, check those specs (see spec
1653 references in the :doc:`atomics guide <Atomics>`).
1654 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1655 differently since they don't take an address. See that instruction's
1656 documentation for details.
1658 For a simpler introduction to the ordering constraints, see the
1662 The set of values that can be read is governed by the happens-before
1663 partial order. A value cannot be read unless some operation wrote
1664 it. This is intended to provide a guarantee strong enough to model
1665 Java's non-volatile shared variables. This ordering cannot be
1666 specified for read-modify-write operations; it is not strong enough
1667 to make them atomic in any interesting way.
1669 In addition to the guarantees of ``unordered``, there is a single
1670 total order for modifications by ``monotonic`` operations on each
1671 address. All modification orders must be compatible with the
1672 happens-before order. There is no guarantee that the modification
1673 orders can be combined to a global total order for the whole program
1674 (and this often will not be possible). The read in an atomic
1675 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1676 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1677 order immediately before the value it writes. If one atomic read
1678 happens before another atomic read of the same address, the later
1679 read must see the same value or a later value in the address's
1680 modification order. This disallows reordering of ``monotonic`` (or
1681 stronger) operations on the same address. If an address is written
1682 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1683 read that address repeatedly, the other threads must eventually see
1684 the write. This corresponds to the C++0x/C1x
1685 ``memory_order_relaxed``.
1687 In addition to the guarantees of ``monotonic``, a
1688 *synchronizes-with* edge may be formed with a ``release`` operation.
1689 This is intended to model C++'s ``memory_order_acquire``.
1691 In addition to the guarantees of ``monotonic``, if this operation
1692 writes a value which is subsequently read by an ``acquire``
1693 operation, it *synchronizes-with* that operation. (This isn't a
1694 complete description; see the C++0x definition of a release
1695 sequence.) This corresponds to the C++0x/C1x
1696 ``memory_order_release``.
1697 ``acq_rel`` (acquire+release)
1698 Acts as both an ``acquire`` and ``release`` operation on its
1699 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1700 ``seq_cst`` (sequentially consistent)
1701 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1702 operation which only reads, ``release`` for an operation which only
1703 writes), there is a global total order on all
1704 sequentially-consistent operations on all addresses, which is
1705 consistent with the *happens-before* partial order and with the
1706 modification orders of all the affected addresses. Each
1707 sequentially-consistent read sees the last preceding write to the
1708 same address in this global order. This corresponds to the C++0x/C1x
1709 ``memory_order_seq_cst`` and Java volatile.
1713 If an atomic operation is marked ``singlethread``, it only *synchronizes
1714 with* or participates in modification and seq\_cst total orderings with
1715 other operations running in the same thread (for example, in signal
1723 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1724 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1725 :ref:`frem <i_frem>`) have the following flags that can set to enable
1726 otherwise unsafe floating point operations
1729 No NaNs - Allow optimizations to assume the arguments and result are not
1730 NaN. Such optimizations are required to retain defined behavior over
1731 NaNs, but the value of the result is undefined.
1734 No Infs - Allow optimizations to assume the arguments and result are not
1735 +/-Inf. Such optimizations are required to retain defined behavior over
1736 +/-Inf, but the value of the result is undefined.
1739 No Signed Zeros - Allow optimizations to treat the sign of a zero
1740 argument or result as insignificant.
1743 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1744 argument rather than perform division.
1747 Fast - Allow algebraically equivalent transformations that may
1748 dramatically change results in floating point (e.g. reassociate). This
1749 flag implies all the others.
1756 The LLVM type system is one of the most important features of the
1757 intermediate representation. Being typed enables a number of
1758 optimizations to be performed on the intermediate representation
1759 directly, without having to do extra analyses on the side before the
1760 transformation. A strong type system makes it easier to read the
1761 generated code and enables novel analyses and transformations that are
1762 not feasible to perform on normal three address code representations.
1772 The void type does not represent any value and has no size.
1790 The function type can be thought of as a function signature. It consists of a
1791 return type and a list of formal parameter types. The return type of a function
1792 type is a void type or first class type --- except for :ref:`label <t_label>`
1793 and :ref:`metadata <t_metadata>` types.
1799 <returntype> (<parameter list>)
1801 ...where '``<parameter list>``' is a comma-separated list of type
1802 specifiers. Optionally, the parameter list may include a type ``...``, which
1803 indicates that the function takes a variable number of arguments. Variable
1804 argument functions can access their arguments with the :ref:`variable argument
1805 handling intrinsic <int_varargs>` functions. '``<returntype>``' is any type
1806 except :ref:`label <t_label>` and :ref:`metadata <t_metadata>`.
1810 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1811 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1812 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1813 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1814 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1815 | ``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. |
1816 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1817 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1818 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1825 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1826 Values of these types are the only ones which can be produced by
1834 These are the types that are valid in registers from CodeGen's perspective.
1843 The integer type is a very simple type that simply specifies an
1844 arbitrary bit width for the integer type desired. Any bit width from 1
1845 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1853 The number of bits the integer will occupy is specified by the ``N``
1859 +----------------+------------------------------------------------+
1860 | ``i1`` | a single-bit integer. |
1861 +----------------+------------------------------------------------+
1862 | ``i32`` | a 32-bit integer. |
1863 +----------------+------------------------------------------------+
1864 | ``i1942652`` | a really big integer of over 1 million bits. |
1865 +----------------+------------------------------------------------+
1869 Floating Point Types
1870 """"""""""""""""""""
1879 - 16-bit floating point value
1882 - 32-bit floating point value
1885 - 64-bit floating point value
1888 - 128-bit floating point value (112-bit mantissa)
1891 - 80-bit floating point value (X87)
1894 - 128-bit floating point value (two 64-bits)
1901 The x86_mmx type represents a value held in an MMX register on an x86
1902 machine. The operations allowed on it are quite limited: parameters and
1903 return values, load and store, and bitcast. User-specified MMX
1904 instructions are represented as intrinsic or asm calls with arguments
1905 and/or results of this type. There are no arrays, vectors or constants
1922 The pointer type is used to specify memory locations. Pointers are
1923 commonly used to reference objects in memory.
1925 Pointer types may have an optional address space attribute defining the
1926 numbered address space where the pointed-to object resides. The default
1927 address space is number zero. The semantics of non-zero address spaces
1928 are target-specific.
1930 Note that LLVM does not permit pointers to void (``void*``) nor does it
1931 permit pointers to labels (``label*``). Use ``i8*`` instead.
1941 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1942 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
1943 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1944 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
1945 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1946 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
1947 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1956 A vector type is a simple derived type that represents a vector of
1957 elements. Vector types are used when multiple primitive data are
1958 operated in parallel using a single instruction (SIMD). A vector type
1959 requires a size (number of elements) and an underlying primitive data
1960 type. Vector types are considered :ref:`first class <t_firstclass>`.
1966 < <# elements> x <elementtype> >
1968 The number of elements is a constant integer value larger than 0;
1969 elementtype may be any integer, floating point or pointer type. Vectors
1970 of size zero are not allowed.
1974 +-------------------+--------------------------------------------------+
1975 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
1976 +-------------------+--------------------------------------------------+
1977 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
1978 +-------------------+--------------------------------------------------+
1979 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
1980 +-------------------+--------------------------------------------------+
1981 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
1982 +-------------------+--------------------------------------------------+
1991 The label type represents code labels.
2006 The metadata type represents embedded metadata. No derived types may be
2007 created from metadata except for :ref:`function <t_function>` arguments.
2020 Aggregate Types are a subset of derived types that can contain multiple
2021 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
2022 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
2032 The array type is a very simple derived type that arranges elements
2033 sequentially in memory. The array type requires a size (number of
2034 elements) and an underlying data type.
2040 [<# elements> x <elementtype>]
2042 The number of elements is a constant integer value; ``elementtype`` may
2043 be any type with a size.
2047 +------------------+--------------------------------------+
2048 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
2049 +------------------+--------------------------------------+
2050 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
2051 +------------------+--------------------------------------+
2052 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
2053 +------------------+--------------------------------------+
2055 Here are some examples of multidimensional arrays:
2057 +-----------------------------+----------------------------------------------------------+
2058 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
2059 +-----------------------------+----------------------------------------------------------+
2060 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
2061 +-----------------------------+----------------------------------------------------------+
2062 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
2063 +-----------------------------+----------------------------------------------------------+
2065 There is no restriction on indexing beyond the end of the array implied
2066 by a static type (though there are restrictions on indexing beyond the
2067 bounds of an allocated object in some cases). This means that
2068 single-dimension 'variable sized array' addressing can be implemented in
2069 LLVM with a zero length array type. An implementation of 'pascal style
2070 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
2080 The structure type is used to represent a collection of data members
2081 together in memory. The elements of a structure may be any type that has
2084 Structures in memory are accessed using '``load``' and '``store``' by
2085 getting a pointer to a field with the '``getelementptr``' instruction.
2086 Structures in registers are accessed using the '``extractvalue``' and
2087 '``insertvalue``' instructions.
2089 Structures may optionally be "packed" structures, which indicate that
2090 the alignment of the struct is one byte, and that there is no padding
2091 between the elements. In non-packed structs, padding between field types
2092 is inserted as defined by the DataLayout string in the module, which is
2093 required to match what the underlying code generator expects.
2095 Structures can either be "literal" or "identified". A literal structure
2096 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
2097 identified types are always defined at the top level with a name.
2098 Literal types are uniqued by their contents and can never be recursive
2099 or opaque since there is no way to write one. Identified types can be
2100 recursive, can be opaqued, and are never uniqued.
2106 %T1 = type { <type list> } ; Identified normal struct type
2107 %T2 = type <{ <type list> }> ; Identified packed struct type
2111 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2112 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
2113 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2114 | ``{ 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``. |
2115 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2116 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
2117 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
2121 Opaque Structure Types
2122 """"""""""""""""""""""
2126 Opaque structure types are used to represent named structure types that
2127 do not have a body specified. This corresponds (for example) to the C
2128 notion of a forward declared structure.
2139 +--------------+-------------------+
2140 | ``opaque`` | An opaque type. |
2141 +--------------+-------------------+
2148 LLVM has several different basic types of constants. This section
2149 describes them all and their syntax.
2154 **Boolean constants**
2155 The two strings '``true``' and '``false``' are both valid constants
2157 **Integer constants**
2158 Standard integers (such as '4') are constants of the
2159 :ref:`integer <t_integer>` type. Negative numbers may be used with
2161 **Floating point constants**
2162 Floating point constants use standard decimal notation (e.g.
2163 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
2164 hexadecimal notation (see below). The assembler requires the exact
2165 decimal value of a floating-point constant. For example, the
2166 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
2167 decimal in binary. Floating point constants must have a :ref:`floating
2168 point <t_floating>` type.
2169 **Null pointer constants**
2170 The identifier '``null``' is recognized as a null pointer constant
2171 and must be of :ref:`pointer type <t_pointer>`.
2173 The one non-intuitive notation for constants is the hexadecimal form of
2174 floating point constants. For example, the form
2175 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
2176 than) '``double 4.5e+15``'. The only time hexadecimal floating point
2177 constants are required (and the only time that they are generated by the
2178 disassembler) is when a floating point constant must be emitted but it
2179 cannot be represented as a decimal floating point number in a reasonable
2180 number of digits. For example, NaN's, infinities, and other special
2181 values are represented in their IEEE hexadecimal format so that assembly
2182 and disassembly do not cause any bits to change in the constants.
2184 When using the hexadecimal form, constants of types half, float, and
2185 double are represented using the 16-digit form shown above (which
2186 matches the IEEE754 representation for double); half and float values
2187 must, however, be exactly representable as IEEE 754 half and single
2188 precision, respectively. Hexadecimal format is always used for long
2189 double, and there are three forms of long double. The 80-bit format used
2190 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
2191 128-bit format used by PowerPC (two adjacent doubles) is represented by
2192 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
2193 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
2194 will only work if they match the long double format on your target.
2195 The IEEE 16-bit format (half precision) is represented by ``0xH``
2196 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
2197 (sign bit at the left).
2199 There are no constants of type x86_mmx.
2201 .. _complexconstants:
2206 Complex constants are a (potentially recursive) combination of simple
2207 constants and smaller complex constants.
2209 **Structure constants**
2210 Structure constants are represented with notation similar to
2211 structure type definitions (a comma separated list of elements,
2212 surrounded by braces (``{}``)). For example:
2213 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
2214 "``@G = external global i32``". Structure constants must have
2215 :ref:`structure type <t_struct>`, and the number and types of elements
2216 must match those specified by the type.
2218 Array constants are represented with notation similar to array type
2219 definitions (a comma separated list of elements, surrounded by
2220 square brackets (``[]``)). For example:
2221 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
2222 :ref:`array type <t_array>`, and the number and types of elements must
2223 match those specified by the type.
2224 **Vector constants**
2225 Vector constants are represented with notation similar to vector
2226 type definitions (a comma separated list of elements, surrounded by
2227 less-than/greater-than's (``<>``)). For example:
2228 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
2229 must have :ref:`vector type <t_vector>`, and the number and types of
2230 elements must match those specified by the type.
2231 **Zero initialization**
2232 The string '``zeroinitializer``' can be used to zero initialize a
2233 value to zero of *any* type, including scalar and
2234 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
2235 having to print large zero initializers (e.g. for large arrays) and
2236 is always exactly equivalent to using explicit zero initializers.
2238 A metadata node is a structure-like constant with :ref:`metadata
2239 type <t_metadata>`. For example:
2240 "``metadata !{ i32 0, metadata !"test" }``". Unlike other
2241 constants that are meant to be interpreted as part of the
2242 instruction stream, metadata is a place to attach additional
2243 information such as debug info.
2245 Global Variable and Function Addresses
2246 --------------------------------------
2248 The addresses of :ref:`global variables <globalvars>` and
2249 :ref:`functions <functionstructure>` are always implicitly valid
2250 (link-time) constants. These constants are explicitly referenced when
2251 the :ref:`identifier for the global <identifiers>` is used and always have
2252 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
2255 .. code-block:: llvm
2259 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
2266 The string '``undef``' can be used anywhere a constant is expected, and
2267 indicates that the user of the value may receive an unspecified
2268 bit-pattern. Undefined values may be of any type (other than '``label``'
2269 or '``void``') and be used anywhere a constant is permitted.
2271 Undefined values are useful because they indicate to the compiler that
2272 the program is well defined no matter what value is used. This gives the
2273 compiler more freedom to optimize. Here are some examples of
2274 (potentially surprising) transformations that are valid (in pseudo IR):
2276 .. code-block:: llvm
2286 This is safe because all of the output bits are affected by the undef
2287 bits. Any output bit can have a zero or one depending on the input bits.
2289 .. code-block:: llvm
2300 These logical operations have bits that are not always affected by the
2301 input. For example, if ``%X`` has a zero bit, then the output of the
2302 '``and``' operation will always be a zero for that bit, no matter what
2303 the corresponding bit from the '``undef``' is. As such, it is unsafe to
2304 optimize or assume that the result of the '``and``' is '``undef``'.
2305 However, it is safe to assume that all bits of the '``undef``' could be
2306 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
2307 all the bits of the '``undef``' operand to the '``or``' could be set,
2308 allowing the '``or``' to be folded to -1.
2310 .. code-block:: llvm
2312 %A = select undef, %X, %Y
2313 %B = select undef, 42, %Y
2314 %C = select %X, %Y, undef
2324 This set of examples shows that undefined '``select``' (and conditional
2325 branch) conditions can go *either way*, but they have to come from one
2326 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
2327 both known to have a clear low bit, then ``%A`` would have to have a
2328 cleared low bit. However, in the ``%C`` example, the optimizer is
2329 allowed to assume that the '``undef``' operand could be the same as
2330 ``%Y``, allowing the whole '``select``' to be eliminated.
2332 .. code-block:: llvm
2334 %A = xor undef, undef
2351 This example points out that two '``undef``' operands are not
2352 necessarily the same. This can be surprising to people (and also matches
2353 C semantics) where they assume that "``X^X``" is always zero, even if
2354 ``X`` is undefined. This isn't true for a number of reasons, but the
2355 short answer is that an '``undef``' "variable" can arbitrarily change
2356 its value over its "live range". This is true because the variable
2357 doesn't actually *have a live range*. Instead, the value is logically
2358 read from arbitrary registers that happen to be around when needed, so
2359 the value is not necessarily consistent over time. In fact, ``%A`` and
2360 ``%C`` need to have the same semantics or the core LLVM "replace all
2361 uses with" concept would not hold.
2363 .. code-block:: llvm
2371 These examples show the crucial difference between an *undefined value*
2372 and *undefined behavior*. An undefined value (like '``undef``') is
2373 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2374 operation can be constant folded to '``undef``', because the '``undef``'
2375 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2376 However, in the second example, we can make a more aggressive
2377 assumption: because the ``undef`` is allowed to be an arbitrary value,
2378 we are allowed to assume that it could be zero. Since a divide by zero
2379 has *undefined behavior*, we are allowed to assume that the operation
2380 does not execute at all. This allows us to delete the divide and all
2381 code after it. Because the undefined operation "can't happen", the
2382 optimizer can assume that it occurs in dead code.
2384 .. code-block:: llvm
2386 a: store undef -> %X
2387 b: store %X -> undef
2392 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2393 value can be assumed to not have any effect; we can assume that the
2394 value is overwritten with bits that happen to match what was already
2395 there. However, a store *to* an undefined location could clobber
2396 arbitrary memory, therefore, it has undefined behavior.
2403 Poison values are similar to :ref:`undef values <undefvalues>`, however
2404 they also represent the fact that an instruction or constant expression
2405 which cannot evoke side effects has nevertheless detected a condition
2406 which results in undefined behavior.
2408 There is currently no way of representing a poison value in the IR; they
2409 only exist when produced by operations such as :ref:`add <i_add>` with
2412 Poison value behavior is defined in terms of value *dependence*:
2414 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2415 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2416 their dynamic predecessor basic block.
2417 - Function arguments depend on the corresponding actual argument values
2418 in the dynamic callers of their functions.
2419 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2420 instructions that dynamically transfer control back to them.
2421 - :ref:`Invoke <i_invoke>` instructions depend on the
2422 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2423 call instructions that dynamically transfer control back to them.
2424 - Non-volatile loads and stores depend on the most recent stores to all
2425 of the referenced memory addresses, following the order in the IR
2426 (including loads and stores implied by intrinsics such as
2427 :ref:`@llvm.memcpy <int_memcpy>`.)
2428 - An instruction with externally visible side effects depends on the
2429 most recent preceding instruction with externally visible side
2430 effects, following the order in the IR. (This includes :ref:`volatile
2431 operations <volatile>`.)
2432 - An instruction *control-depends* on a :ref:`terminator
2433 instruction <terminators>` if the terminator instruction has
2434 multiple successors and the instruction is always executed when
2435 control transfers to one of the successors, and may not be executed
2436 when control is transferred to another.
2437 - Additionally, an instruction also *control-depends* on a terminator
2438 instruction if the set of instructions it otherwise depends on would
2439 be different if the terminator had transferred control to a different
2441 - Dependence is transitive.
2443 Poison Values have the same behavior as :ref:`undef values <undefvalues>`,
2444 with the additional affect that any instruction which has a *dependence*
2445 on a poison value has undefined behavior.
2447 Here are some examples:
2449 .. code-block:: llvm
2452 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2453 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2454 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
2455 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2457 store i32 %poison, i32* @g ; Poison value stored to memory.
2458 %poison2 = load i32* @g ; Poison value loaded back from memory.
2460 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2462 %narrowaddr = bitcast i32* @g to i16*
2463 %wideaddr = bitcast i32* @g to i64*
2464 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2465 %poison4 = load i64* %wideaddr ; Returns a poison value.
2467 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2468 br i1 %cmp, label %true, label %end ; Branch to either destination.
2471 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2472 ; it has undefined behavior.
2476 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2477 ; Both edges into this PHI are
2478 ; control-dependent on %cmp, so this
2479 ; always results in a poison value.
2481 store volatile i32 0, i32* @g ; This would depend on the store in %true
2482 ; if %cmp is true, or the store in %entry
2483 ; otherwise, so this is undefined behavior.
2485 br i1 %cmp, label %second_true, label %second_end
2486 ; The same branch again, but this time the
2487 ; true block doesn't have side effects.
2494 store volatile i32 0, i32* @g ; This time, the instruction always depends
2495 ; on the store in %end. Also, it is
2496 ; control-equivalent to %end, so this is
2497 ; well-defined (ignoring earlier undefined
2498 ; behavior in this example).
2502 Addresses of Basic Blocks
2503 -------------------------
2505 ``blockaddress(@function, %block)``
2507 The '``blockaddress``' constant computes the address of the specified
2508 basic block in the specified function, and always has an ``i8*`` type.
2509 Taking the address of the entry block is illegal.
2511 This value only has defined behavior when used as an operand to the
2512 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2513 against null. Pointer equality tests between labels addresses results in
2514 undefined behavior --- though, again, comparison against null is ok, and
2515 no label is equal to the null pointer. This may be passed around as an
2516 opaque pointer sized value as long as the bits are not inspected. This
2517 allows ``ptrtoint`` and arithmetic to be performed on these values so
2518 long as the original value is reconstituted before the ``indirectbr``
2521 Finally, some targets may provide defined semantics when using the value
2522 as the operand to an inline assembly, but that is target specific.
2526 Constant Expressions
2527 --------------------
2529 Constant expressions are used to allow expressions involving other
2530 constants to be used as constants. Constant expressions may be of any
2531 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2532 that does not have side effects (e.g. load and call are not supported).
2533 The following is the syntax for constant expressions:
2535 ``trunc (CST to TYPE)``
2536 Truncate a constant to another type. The bit size of CST must be
2537 larger than the bit size of TYPE. Both types must be integers.
2538 ``zext (CST to TYPE)``
2539 Zero extend a constant to another type. The bit size of CST must be
2540 smaller than the bit size of TYPE. Both types must be integers.
2541 ``sext (CST to TYPE)``
2542 Sign extend a constant to another type. The bit size of CST must be
2543 smaller than the bit size of TYPE. Both types must be integers.
2544 ``fptrunc (CST to TYPE)``
2545 Truncate a floating point constant to another floating point type.
2546 The size of CST must be larger than the size of TYPE. Both types
2547 must be floating point.
2548 ``fpext (CST to TYPE)``
2549 Floating point extend a constant to another type. The size of CST
2550 must be smaller or equal to the size of TYPE. Both types must be
2552 ``fptoui (CST to TYPE)``
2553 Convert a floating point constant to the corresponding unsigned
2554 integer constant. TYPE must be a scalar or vector integer type. CST
2555 must be of scalar or vector floating point type. Both CST and TYPE
2556 must be scalars, or vectors of the same number of elements. If the
2557 value won't fit in the integer type, the results are undefined.
2558 ``fptosi (CST to TYPE)``
2559 Convert a floating point constant to the corresponding signed
2560 integer constant. TYPE must be a scalar or vector integer type. CST
2561 must be of scalar or vector floating point type. Both CST and TYPE
2562 must be scalars, or vectors of the same number of elements. If the
2563 value won't fit in the integer type, the results are undefined.
2564 ``uitofp (CST to TYPE)``
2565 Convert an unsigned integer constant to the corresponding floating
2566 point constant. TYPE must be a scalar or vector floating point type.
2567 CST must be of scalar or vector integer type. Both CST and TYPE must
2568 be scalars, or vectors of the same number of elements. If the value
2569 won't fit in the floating point type, the results are undefined.
2570 ``sitofp (CST to TYPE)``
2571 Convert a signed integer constant to the corresponding floating
2572 point constant. TYPE must be a scalar or vector floating point type.
2573 CST must be of scalar or vector integer type. Both CST and TYPE must
2574 be scalars, or vectors of the same number of elements. If the value
2575 won't fit in the floating point type, the results are undefined.
2576 ``ptrtoint (CST to TYPE)``
2577 Convert a pointer typed constant to the corresponding integer
2578 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2579 pointer type. The ``CST`` value is zero extended, truncated, or
2580 unchanged to make it fit in ``TYPE``.
2581 ``inttoptr (CST to TYPE)``
2582 Convert an integer constant to a pointer constant. TYPE must be a
2583 pointer type. CST must be of integer type. The CST value is zero
2584 extended, truncated, or unchanged to make it fit in a pointer size.
2585 This one is *really* dangerous!
2586 ``bitcast (CST to TYPE)``
2587 Convert a constant, CST, to another TYPE. The constraints of the
2588 operands are the same as those for the :ref:`bitcast
2589 instruction <i_bitcast>`.
2590 ``addrspacecast (CST to TYPE)``
2591 Convert a constant pointer or constant vector of pointer, CST, to another
2592 TYPE in a different address space. The constraints of the operands are the
2593 same as those for the :ref:`addrspacecast instruction <i_addrspacecast>`.
2594 ``getelementptr (CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)``
2595 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2596 constants. As with the :ref:`getelementptr <i_getelementptr>`
2597 instruction, the index list may have zero or more indexes, which are
2598 required to make sense for the type of "CSTPTR".
2599 ``select (COND, VAL1, VAL2)``
2600 Perform the :ref:`select operation <i_select>` on constants.
2601 ``icmp COND (VAL1, VAL2)``
2602 Performs the :ref:`icmp operation <i_icmp>` on constants.
2603 ``fcmp COND (VAL1, VAL2)``
2604 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2605 ``extractelement (VAL, IDX)``
2606 Perform the :ref:`extractelement operation <i_extractelement>` on
2608 ``insertelement (VAL, ELT, IDX)``
2609 Perform the :ref:`insertelement operation <i_insertelement>` on
2611 ``shufflevector (VEC1, VEC2, IDXMASK)``
2612 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2614 ``extractvalue (VAL, IDX0, IDX1, ...)``
2615 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2616 constants. The index list is interpreted in a similar manner as
2617 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2618 least one index value must be specified.
2619 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2620 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2621 The index list is interpreted in a similar manner as indices in a
2622 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2623 value must be specified.
2624 ``OPCODE (LHS, RHS)``
2625 Perform the specified operation of the LHS and RHS constants. OPCODE
2626 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2627 binary <bitwiseops>` operations. The constraints on operands are
2628 the same as those for the corresponding instruction (e.g. no bitwise
2629 operations on floating point values are allowed).
2636 Inline Assembler Expressions
2637 ----------------------------
2639 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2640 Inline Assembly <moduleasm>`) through the use of a special value. This
2641 value represents the inline assembler as a string (containing the
2642 instructions to emit), a list of operand constraints (stored as a
2643 string), a flag that indicates whether or not the inline asm expression
2644 has side effects, and a flag indicating whether the function containing
2645 the asm needs to align its stack conservatively. An example inline
2646 assembler expression is:
2648 .. code-block:: llvm
2650 i32 (i32) asm "bswap $0", "=r,r"
2652 Inline assembler expressions may **only** be used as the callee operand
2653 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2654 Thus, typically we have:
2656 .. code-block:: llvm
2658 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2660 Inline asms with side effects not visible in the constraint list must be
2661 marked as having side effects. This is done through the use of the
2662 '``sideeffect``' keyword, like so:
2664 .. code-block:: llvm
2666 call void asm sideeffect "eieio", ""()
2668 In some cases inline asms will contain code that will not work unless
2669 the stack is aligned in some way, such as calls or SSE instructions on
2670 x86, yet will not contain code that does that alignment within the asm.
2671 The compiler should make conservative assumptions about what the asm
2672 might contain and should generate its usual stack alignment code in the
2673 prologue if the '``alignstack``' keyword is present:
2675 .. code-block:: llvm
2677 call void asm alignstack "eieio", ""()
2679 Inline asms also support using non-standard assembly dialects. The
2680 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2681 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2682 the only supported dialects. An example is:
2684 .. code-block:: llvm
2686 call void asm inteldialect "eieio", ""()
2688 If multiple keywords appear the '``sideeffect``' keyword must come
2689 first, the '``alignstack``' keyword second and the '``inteldialect``'
2695 The call instructions that wrap inline asm nodes may have a
2696 "``!srcloc``" MDNode attached to it that contains a list of constant
2697 integers. If present, the code generator will use the integer as the
2698 location cookie value when report errors through the ``LLVMContext``
2699 error reporting mechanisms. This allows a front-end to correlate backend
2700 errors that occur with inline asm back to the source code that produced
2703 .. code-block:: llvm
2705 call void asm sideeffect "something bad", ""(), !srcloc !42
2707 !42 = !{ i32 1234567 }
2709 It is up to the front-end to make sense of the magic numbers it places
2710 in the IR. If the MDNode contains multiple constants, the code generator
2711 will use the one that corresponds to the line of the asm that the error
2716 Metadata Nodes and Metadata Strings
2717 -----------------------------------
2719 LLVM IR allows metadata to be attached to instructions in the program
2720 that can convey extra information about the code to the optimizers and
2721 code generator. One example application of metadata is source-level
2722 debug information. There are two metadata primitives: strings and nodes.
2723 All metadata has the ``metadata`` type and is identified in syntax by a
2724 preceding exclamation point ('``!``').
2726 A metadata string is a string surrounded by double quotes. It can
2727 contain any character by escaping non-printable characters with
2728 "``\xx``" where "``xx``" is the two digit hex code. For example:
2731 Metadata nodes are represented with notation similar to structure
2732 constants (a comma separated list of elements, surrounded by braces and
2733 preceded by an exclamation point). Metadata nodes can have any values as
2734 their operand. For example:
2736 .. code-block:: llvm
2738 !{ metadata !"test\00", i32 10}
2740 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2741 metadata nodes, which can be looked up in the module symbol table. For
2744 .. code-block:: llvm
2746 !foo = metadata !{!4, !3}
2748 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2749 function is using two metadata arguments:
2751 .. code-block:: llvm
2753 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2755 Metadata can be attached with an instruction. Here metadata ``!21`` is
2756 attached to the ``add`` instruction using the ``!dbg`` identifier:
2758 .. code-block:: llvm
2760 %indvar.next = add i64 %indvar, 1, !dbg !21
2762 More information about specific metadata nodes recognized by the
2763 optimizers and code generator is found below.
2768 In LLVM IR, memory does not have types, so LLVM's own type system is not
2769 suitable for doing TBAA. Instead, metadata is added to the IR to
2770 describe a type system of a higher level language. This can be used to
2771 implement typical C/C++ TBAA, but it can also be used to implement
2772 custom alias analysis behavior for other languages.
2774 The current metadata format is very simple. TBAA metadata nodes have up
2775 to three fields, e.g.:
2777 .. code-block:: llvm
2779 !0 = metadata !{ metadata !"an example type tree" }
2780 !1 = metadata !{ metadata !"int", metadata !0 }
2781 !2 = metadata !{ metadata !"float", metadata !0 }
2782 !3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2784 The first field is an identity field. It can be any value, usually a
2785 metadata string, which uniquely identifies the type. The most important
2786 name in the tree is the name of the root node. Two trees with different
2787 root node names are entirely disjoint, even if they have leaves with
2790 The second field identifies the type's parent node in the tree, or is
2791 null or omitted for a root node. A type is considered to alias all of
2792 its descendants and all of its ancestors in the tree. Also, a type is
2793 considered to alias all types in other trees, so that bitcode produced
2794 from multiple front-ends is handled conservatively.
2796 If the third field is present, it's an integer which if equal to 1
2797 indicates that the type is "constant" (meaning
2798 ``pointsToConstantMemory`` should return true; see `other useful
2799 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
2801 '``tbaa.struct``' Metadata
2802 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2804 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
2805 aggregate assignment operations in C and similar languages, however it
2806 is defined to copy a contiguous region of memory, which is more than
2807 strictly necessary for aggregate types which contain holes due to
2808 padding. Also, it doesn't contain any TBAA information about the fields
2811 ``!tbaa.struct`` metadata can describe which memory subregions in a
2812 memcpy are padding and what the TBAA tags of the struct are.
2814 The current metadata format is very simple. ``!tbaa.struct`` metadata
2815 nodes are a list of operands which are in conceptual groups of three.
2816 For each group of three, the first operand gives the byte offset of a
2817 field in bytes, the second gives its size in bytes, and the third gives
2820 .. code-block:: llvm
2822 !4 = metadata !{ i64 0, i64 4, metadata !1, i64 8, i64 4, metadata !2 }
2824 This describes a struct with two fields. The first is at offset 0 bytes
2825 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
2826 and has size 4 bytes and has tbaa tag !2.
2828 Note that the fields need not be contiguous. In this example, there is a
2829 4 byte gap between the two fields. This gap represents padding which
2830 does not carry useful data and need not be preserved.
2832 '``noalias``' and '``alias.scope``' Metadata
2833 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2835 ``noalias`` and ``alias.scope`` metadata provide the ability to specify generic
2836 noalias memory-access sets. This means that some collection of memory access
2837 instructions (loads, stores, memory-accessing calls, etc.) that carry
2838 ``noalias`` metadata can specifically be specified not to alias with some other
2839 collection of memory access instructions that carry ``alias.scope`` metadata.
2840 Each type of metadata specifies a list of scopes where each scope has an id and
2841 a domain. When evaluating an aliasing query, if for some some domain, the set
2842 of scopes with that domain in one instruction's ``alias.scope`` list is a
2843 subset of (or qual to) the set of scopes for that domain in another
2844 instruction's ``noalias`` list, then the two memory accesses are assumed not to
2847 The metadata identifying each domain is itself a list containing one or two
2848 entries. The first entry is the name of the domain. Note that if the name is a
2849 string then it can be combined accross functions and translation units. A
2850 self-reference can be used to create globally unique domain names. A
2851 descriptive string may optionally be provided as a second list entry.
2853 The metadata identifying each scope is also itself a list containing two or
2854 three entries. The first entry is the name of the scope. Note that if the name
2855 is a string then it can be combined accross functions and translation units. A
2856 self-reference can be used to create globally unique scope names. A metadata
2857 reference to the scope's domain is the second entry. A descriptive string may
2858 optionally be provided as a third list entry.
2862 .. code-block:: llvm
2864 ; Two scope domains:
2865 !0 = metadata !{metadata !0}
2866 !1 = metadata !{metadata !1}
2868 ; Some scopes in these domains:
2869 !2 = metadata !{metadata !2, metadata !0}
2870 !3 = metadata !{metadata !3, metadata !0}
2871 !4 = metadata !{metadata !4, metadata !1}
2874 !5 = metadata !{metadata !4} ; A list containing only scope !4
2875 !6 = metadata !{metadata !4, metadata !3, metadata !2}
2876 !7 = metadata !{metadata !3}
2878 ; These two instructions don't alias:
2879 %0 = load float* %c, align 4, !alias.scope !5
2880 store float %0, float* %arrayidx.i, align 4, !noalias !5
2882 ; These two instructions also don't alias (for domain !1, the set of scopes
2883 ; in the !alias.scope equals that in the !noalias list):
2884 %2 = load float* %c, align 4, !alias.scope !5
2885 store float %2, float* %arrayidx.i2, align 4, !noalias !6
2887 ; These two instructions don't alias (for domain !0, the set of scopes in
2888 ; the !noalias list is not a superset of, or equal to, the scopes in the
2889 ; !alias.scope list):
2890 %2 = load float* %c, align 4, !alias.scope !6
2891 store float %0, float* %arrayidx.i, align 4, !noalias !7
2893 '``fpmath``' Metadata
2894 ^^^^^^^^^^^^^^^^^^^^^
2896 ``fpmath`` metadata may be attached to any instruction of floating point
2897 type. It can be used to express the maximum acceptable error in the
2898 result of that instruction, in ULPs, thus potentially allowing the
2899 compiler to use a more efficient but less accurate method of computing
2900 it. ULP is defined as follows:
2902 If ``x`` is a real number that lies between two finite consecutive
2903 floating-point numbers ``a`` and ``b``, without being equal to one
2904 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
2905 distance between the two non-equal finite floating-point numbers
2906 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
2908 The metadata node shall consist of a single positive floating point
2909 number representing the maximum relative error, for example:
2911 .. code-block:: llvm
2913 !0 = metadata !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
2915 '``range``' Metadata
2916 ^^^^^^^^^^^^^^^^^^^^
2918 ``range`` metadata may be attached only to ``load``, ``call`` and ``invoke`` of
2919 integer types. It expresses the possible ranges the loaded value or the value
2920 returned by the called function at this call site is in. The ranges are
2921 represented with a flattened list of integers. The loaded value or the value
2922 returned is known to be in the union of the ranges defined by each consecutive
2923 pair. Each pair has the following properties:
2925 - The type must match the type loaded by the instruction.
2926 - The pair ``a,b`` represents the range ``[a,b)``.
2927 - Both ``a`` and ``b`` are constants.
2928 - The range is allowed to wrap.
2929 - The range should not represent the full or empty set. That is,
2932 In addition, the pairs must be in signed order of the lower bound and
2933 they must be non-contiguous.
2937 .. code-block:: llvm
2939 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
2940 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
2941 %c = call i8 @foo(), !range !2 ; Can only be 0, 1, 3, 4 or 5
2942 %d = invoke i8 @bar() to label %cont
2943 unwind label %lpad, !range !3 ; Can only be -2, -1, 3, 4 or 5
2945 !0 = metadata !{ i8 0, i8 2 }
2946 !1 = metadata !{ i8 255, i8 2 }
2947 !2 = metadata !{ i8 0, i8 2, i8 3, i8 6 }
2948 !3 = metadata !{ i8 -2, i8 0, i8 3, i8 6 }
2953 It is sometimes useful to attach information to loop constructs. Currently,
2954 loop metadata is implemented as metadata attached to the branch instruction
2955 in the loop latch block. This type of metadata refer to a metadata node that is
2956 guaranteed to be separate for each loop. The loop identifier metadata is
2957 specified with the name ``llvm.loop``.
2959 The loop identifier metadata is implemented using a metadata that refers to
2960 itself to avoid merging it with any other identifier metadata, e.g.,
2961 during module linkage or function inlining. That is, each loop should refer
2962 to their own identification metadata even if they reside in separate functions.
2963 The following example contains loop identifier metadata for two separate loop
2966 .. code-block:: llvm
2968 !0 = metadata !{ metadata !0 }
2969 !1 = metadata !{ metadata !1 }
2971 The loop identifier metadata can be used to specify additional
2972 per-loop metadata. Any operands after the first operand can be treated
2973 as user-defined metadata. For example the ``llvm.loop.unroll.count``
2974 suggests an unroll factor to the loop unroller:
2976 .. code-block:: llvm
2978 br i1 %exitcond, label %._crit_edge, label %.lr.ph, !llvm.loop !0
2980 !0 = metadata !{ metadata !0, metadata !1 }
2981 !1 = metadata !{ metadata !"llvm.loop.unroll.count", i32 4 }
2983 '``llvm.loop.vectorize``' and '``llvm.loop.interleave``'
2984 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2986 Metadata prefixed with ``llvm.loop.vectorize`` or ``llvm.loop.interleave`` are
2987 used to control per-loop vectorization and interleaving parameters such as
2988 vectorization width and interleave count. These metadata should be used in
2989 conjunction with ``llvm.loop`` loop identification metadata. The
2990 ``llvm.loop.vectorize`` and ``llvm.loop.interleave`` metadata are only
2991 optimization hints and the optimizer will only interleave and vectorize loops if
2992 it believes it is safe to do so. The ``llvm.mem.parallel_loop_access`` metadata
2993 which contains information about loop-carried memory dependencies can be helpful
2994 in determining the safety of these transformations.
2996 '``llvm.loop.interleave.count``' Metadata
2997 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2999 This metadata suggests an interleave count to the loop interleaver.
3000 The first operand is the string ``llvm.loop.interleave.count`` and the
3001 second operand is an integer specifying the interleave count. For
3004 .. code-block:: llvm
3006 !0 = metadata !{ metadata !"llvm.loop.interleave.count", i32 4 }
3008 Note that setting ``llvm.loop.interleave.count`` to 1 disables interleaving
3009 multiple iterations of the loop. If ``llvm.loop.interleave.count`` is set to 0
3010 then the interleave count will be determined automatically.
3012 '``llvm.loop.vectorize.enable``' Metadata
3013 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3015 This metadata selectively enables or disables vectorization for the loop. The
3016 first operand is the string ``llvm.loop.vectorize.enable`` and the second operand
3017 is a bit. If the bit operand value is 1 vectorization is enabled. A value of
3018 0 disables vectorization:
3020 .. code-block:: llvm
3022 !0 = metadata !{ metadata !"llvm.loop.vectorize.enable", i1 0 }
3023 !1 = metadata !{ metadata !"llvm.loop.vectorize.enable", i1 1 }
3025 '``llvm.loop.vectorize.width``' Metadata
3026 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3028 This metadata sets the target width of the vectorizer. The first
3029 operand is the string ``llvm.loop.vectorize.width`` and the second
3030 operand is an integer specifying the width. For example:
3032 .. code-block:: llvm
3034 !0 = metadata !{ metadata !"llvm.loop.vectorize.width", i32 4 }
3036 Note that setting ``llvm.loop.vectorize.width`` to 1 disables
3037 vectorization of the loop. If ``llvm.loop.vectorize.width`` is set to
3038 0 or if the loop does not have this metadata the width will be
3039 determined automatically.
3041 '``llvm.loop.unroll``'
3042 ^^^^^^^^^^^^^^^^^^^^^^
3044 Metadata prefixed with ``llvm.loop.unroll`` are loop unrolling
3045 optimization hints such as the unroll factor. ``llvm.loop.unroll``
3046 metadata should be used in conjunction with ``llvm.loop`` loop
3047 identification metadata. The ``llvm.loop.unroll`` metadata are only
3048 optimization hints and the unrolling will only be performed if the
3049 optimizer believes it is safe to do so.
3051 '``llvm.loop.unroll.count``' Metadata
3052 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3054 This metadata suggests an unroll factor to the loop unroller. The
3055 first operand is the string ``llvm.loop.unroll.count`` and the second
3056 operand is a positive integer specifying the unroll factor. For
3059 .. code-block:: llvm
3061 !0 = metadata !{ metadata !"llvm.loop.unroll.count", i32 4 }
3063 If the trip count of the loop is less than the unroll count the loop
3064 will be partially unrolled.
3066 '``llvm.loop.unroll.disable``' Metadata
3067 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3069 This metadata either disables loop unrolling. The metadata has a single operand
3070 which is the string ``llvm.loop.unroll.disable``. For example:
3072 .. code-block:: llvm
3074 !0 = metadata !{ metadata !"llvm.loop.unroll.disable" }
3076 '``llvm.loop.unroll.full``' Metadata
3077 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3079 This metadata either suggests that the loop should be unrolled fully. The
3080 metadata has a single operand which is the string ``llvm.loop.unroll.disable``.
3083 .. code-block:: llvm
3085 !0 = metadata !{ metadata !"llvm.loop.unroll.full" }
3090 Metadata types used to annotate memory accesses with information helpful
3091 for optimizations are prefixed with ``llvm.mem``.
3093 '``llvm.mem.parallel_loop_access``' Metadata
3094 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3096 The ``llvm.mem.parallel_loop_access`` metadata refers to a loop identifier,
3097 or metadata containing a list of loop identifiers for nested loops.
3098 The metadata is attached to memory accessing instructions and denotes that
3099 no loop carried memory dependence exist between it and other instructions denoted
3100 with the same loop identifier.
3102 Precisely, given two instructions ``m1`` and ``m2`` that both have the
3103 ``llvm.mem.parallel_loop_access`` metadata, with ``L1`` and ``L2`` being the
3104 set of loops associated with that metadata, respectively, then there is no loop
3105 carried dependence between ``m1`` and ``m2`` for loops in both ``L1`` and
3108 As a special case, if all memory accessing instructions in a loop have
3109 ``llvm.mem.parallel_loop_access`` metadata that refers to that loop, then the
3110 loop has no loop carried memory dependences and is considered to be a parallel
3113 Note that if not all memory access instructions have such metadata referring to
3114 the loop, then the loop is considered not being trivially parallel. Additional
3115 memory dependence analysis is required to make that determination. As a fail
3116 safe mechanism, this causes loops that were originally parallel to be considered
3117 sequential (if optimization passes that are unaware of the parallel semantics
3118 insert new memory instructions into the loop body).
3120 Example of a loop that is considered parallel due to its correct use of
3121 both ``llvm.loop`` and ``llvm.mem.parallel_loop_access``
3122 metadata types that refer to the same loop identifier metadata.
3124 .. code-block:: llvm
3128 %val0 = load i32* %arrayidx, !llvm.mem.parallel_loop_access !0
3130 store i32 %val0, i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3132 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
3136 !0 = metadata !{ metadata !0 }
3138 It is also possible to have nested parallel loops. In that case the
3139 memory accesses refer to a list of loop identifier metadata nodes instead of
3140 the loop identifier metadata node directly:
3142 .. code-block:: llvm
3146 %val1 = load i32* %arrayidx3, !llvm.mem.parallel_loop_access !2
3148 br label %inner.for.body
3152 %val0 = load i32* %arrayidx1, !llvm.mem.parallel_loop_access !0
3154 store i32 %val0, i32* %arrayidx2, !llvm.mem.parallel_loop_access !0
3156 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop !1
3160 store i32 %val1, i32* %arrayidx4, !llvm.mem.parallel_loop_access !2
3162 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop !2
3164 outer.for.end: ; preds = %for.body
3166 !0 = metadata !{ metadata !1, metadata !2 } ; a list of loop identifiers
3167 !1 = metadata !{ metadata !1 } ; an identifier for the inner loop
3168 !2 = metadata !{ metadata !2 } ; an identifier for the outer loop
3170 Module Flags Metadata
3171 =====================
3173 Information about the module as a whole is difficult to convey to LLVM's
3174 subsystems. The LLVM IR isn't sufficient to transmit this information.
3175 The ``llvm.module.flags`` named metadata exists in order to facilitate
3176 this. These flags are in the form of key / value pairs --- much like a
3177 dictionary --- making it easy for any subsystem who cares about a flag to
3180 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
3181 Each triplet has the following form:
3183 - The first element is a *behavior* flag, which specifies the behavior
3184 when two (or more) modules are merged together, and it encounters two
3185 (or more) metadata with the same ID. The supported behaviors are
3187 - The second element is a metadata string that is a unique ID for the
3188 metadata. Each module may only have one flag entry for each unique ID (not
3189 including entries with the **Require** behavior).
3190 - The third element is the value of the flag.
3192 When two (or more) modules are merged together, the resulting
3193 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
3194 each unique metadata ID string, there will be exactly one entry in the merged
3195 modules ``llvm.module.flags`` metadata table, and the value for that entry will
3196 be determined by the merge behavior flag, as described below. The only exception
3197 is that entries with the *Require* behavior are always preserved.
3199 The following behaviors are supported:
3210 Emits an error if two values disagree, otherwise the resulting value
3211 is that of the operands.
3215 Emits a warning if two values disagree. The result value will be the
3216 operand for the flag from the first module being linked.
3220 Adds a requirement that another module flag be present and have a
3221 specified value after linking is performed. The value must be a
3222 metadata pair, where the first element of the pair is the ID of the
3223 module flag to be restricted, and the second element of the pair is
3224 the value the module flag should be restricted to. This behavior can
3225 be used to restrict the allowable results (via triggering of an
3226 error) of linking IDs with the **Override** behavior.
3230 Uses the specified value, regardless of the behavior or value of the
3231 other module. If both modules specify **Override**, but the values
3232 differ, an error will be emitted.
3236 Appends the two values, which are required to be metadata nodes.
3240 Appends the two values, which are required to be metadata
3241 nodes. However, duplicate entries in the second list are dropped
3242 during the append operation.
3244 It is an error for a particular unique flag ID to have multiple behaviors,
3245 except in the case of **Require** (which adds restrictions on another metadata
3246 value) or **Override**.
3248 An example of module flags:
3250 .. code-block:: llvm
3252 !0 = metadata !{ i32 1, metadata !"foo", i32 1 }
3253 !1 = metadata !{ i32 4, metadata !"bar", i32 37 }
3254 !2 = metadata !{ i32 2, metadata !"qux", i32 42 }
3255 !3 = metadata !{ i32 3, metadata !"qux",
3257 metadata !"foo", i32 1
3260 !llvm.module.flags = !{ !0, !1, !2, !3 }
3262 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
3263 if two or more ``!"foo"`` flags are seen is to emit an error if their
3264 values are not equal.
3266 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
3267 behavior if two or more ``!"bar"`` flags are seen is to use the value
3270 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
3271 behavior if two or more ``!"qux"`` flags are seen is to emit a
3272 warning if their values are not equal.
3274 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
3278 metadata !{ metadata !"foo", i32 1 }
3280 The behavior is to emit an error if the ``llvm.module.flags`` does not
3281 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
3284 Objective-C Garbage Collection Module Flags Metadata
3285 ----------------------------------------------------
3287 On the Mach-O platform, Objective-C stores metadata about garbage
3288 collection in a special section called "image info". The metadata
3289 consists of a version number and a bitmask specifying what types of
3290 garbage collection are supported (if any) by the file. If two or more
3291 modules are linked together their garbage collection metadata needs to
3292 be merged rather than appended together.
3294 The Objective-C garbage collection module flags metadata consists of the
3295 following key-value pairs:
3304 * - ``Objective-C Version``
3305 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
3307 * - ``Objective-C Image Info Version``
3308 - **[Required]** --- The version of the image info section. Currently
3311 * - ``Objective-C Image Info Section``
3312 - **[Required]** --- The section to place the metadata. Valid values are
3313 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
3314 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
3315 Objective-C ABI version 2.
3317 * - ``Objective-C Garbage Collection``
3318 - **[Required]** --- Specifies whether garbage collection is supported or
3319 not. Valid values are 0, for no garbage collection, and 2, for garbage
3320 collection supported.
3322 * - ``Objective-C GC Only``
3323 - **[Optional]** --- Specifies that only garbage collection is supported.
3324 If present, its value must be 6. This flag requires that the
3325 ``Objective-C Garbage Collection`` flag have the value 2.
3327 Some important flag interactions:
3329 - If a module with ``Objective-C Garbage Collection`` set to 0 is
3330 merged with a module with ``Objective-C Garbage Collection`` set to
3331 2, then the resulting module has the
3332 ``Objective-C Garbage Collection`` flag set to 0.
3333 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
3334 merged with a module with ``Objective-C GC Only`` set to 6.
3336 Automatic Linker Flags Module Flags Metadata
3337 --------------------------------------------
3339 Some targets support embedding flags to the linker inside individual object
3340 files. Typically this is used in conjunction with language extensions which
3341 allow source files to explicitly declare the libraries they depend on, and have
3342 these automatically be transmitted to the linker via object files.
3344 These flags are encoded in the IR using metadata in the module flags section,
3345 using the ``Linker Options`` key. The merge behavior for this flag is required
3346 to be ``AppendUnique``, and the value for the key is expected to be a metadata
3347 node which should be a list of other metadata nodes, each of which should be a
3348 list of metadata strings defining linker options.
3350 For example, the following metadata section specifies two separate sets of
3351 linker options, presumably to link against ``libz`` and the ``Cocoa``
3354 !0 = metadata !{ i32 6, metadata !"Linker Options",
3356 metadata !{ metadata !"-lz" },
3357 metadata !{ metadata !"-framework", metadata !"Cocoa" } } }
3358 !llvm.module.flags = !{ !0 }
3360 The metadata encoding as lists of lists of options, as opposed to a collapsed
3361 list of options, is chosen so that the IR encoding can use multiple option
3362 strings to specify e.g., a single library, while still having that specifier be
3363 preserved as an atomic element that can be recognized by a target specific
3364 assembly writer or object file emitter.
3366 Each individual option is required to be either a valid option for the target's
3367 linker, or an option that is reserved by the target specific assembly writer or
3368 object file emitter. No other aspect of these options is defined by the IR.
3370 C type width Module Flags Metadata
3371 ----------------------------------
3373 The ARM backend emits a section into each generated object file describing the
3374 options that it was compiled with (in a compiler-independent way) to prevent
3375 linking incompatible objects, and to allow automatic library selection. Some
3376 of these options are not visible at the IR level, namely wchar_t width and enum
3379 To pass this information to the backend, these options are encoded in module
3380 flags metadata, using the following key-value pairs:
3390 - * 0 --- sizeof(wchar_t) == 4
3391 * 1 --- sizeof(wchar_t) == 2
3394 - * 0 --- Enums are at least as large as an ``int``.
3395 * 1 --- Enums are stored in the smallest integer type which can
3396 represent all of its values.
3398 For example, the following metadata section specifies that the module was
3399 compiled with a ``wchar_t`` width of 4 bytes, and the underlying type of an
3400 enum is the smallest type which can represent all of its values::
3402 !llvm.module.flags = !{!0, !1}
3403 !0 = metadata !{i32 1, metadata !"short_wchar", i32 1}
3404 !1 = metadata !{i32 1, metadata !"short_enum", i32 0}
3406 .. _intrinsicglobalvariables:
3408 Intrinsic Global Variables
3409 ==========================
3411 LLVM has a number of "magic" global variables that contain data that
3412 affect code generation or other IR semantics. These are documented here.
3413 All globals of this sort should have a section specified as
3414 "``llvm.metadata``". This section and all globals that start with
3415 "``llvm.``" are reserved for use by LLVM.
3419 The '``llvm.used``' Global Variable
3420 -----------------------------------
3422 The ``@llvm.used`` global is an array which has
3423 :ref:`appending linkage <linkage_appending>`. This array contains a list of
3424 pointers to named global variables, functions and aliases which may optionally
3425 have a pointer cast formed of bitcast or getelementptr. For example, a legal
3428 .. code-block:: llvm
3433 @llvm.used = appending global [2 x i8*] [
3435 i8* bitcast (i32* @Y to i8*)
3436 ], section "llvm.metadata"
3438 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
3439 and linker are required to treat the symbol as if there is a reference to the
3440 symbol that it cannot see (which is why they have to be named). For example, if
3441 a variable has internal linkage and no references other than that from the
3442 ``@llvm.used`` list, it cannot be deleted. This is commonly used to represent
3443 references from inline asms and other things the compiler cannot "see", and
3444 corresponds to "``attribute((used))``" in GNU C.
3446 On some targets, the code generator must emit a directive to the
3447 assembler or object file to prevent the assembler and linker from
3448 molesting the symbol.
3450 .. _gv_llvmcompilerused:
3452 The '``llvm.compiler.used``' Global Variable
3453 --------------------------------------------
3455 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
3456 directive, except that it only prevents the compiler from touching the
3457 symbol. On targets that support it, this allows an intelligent linker to
3458 optimize references to the symbol without being impeded as it would be
3461 This is a rare construct that should only be used in rare circumstances,
3462 and should not be exposed to source languages.
3464 .. _gv_llvmglobalctors:
3466 The '``llvm.global_ctors``' Global Variable
3467 -------------------------------------------
3469 .. code-block:: llvm
3471 %0 = type { i32, void ()*, i8* }
3472 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor, i8* @data }]
3474 The ``@llvm.global_ctors`` array contains a list of constructor
3475 functions, priorities, and an optional associated global or function.
3476 The functions referenced by this array will be called in ascending order
3477 of priority (i.e. lowest first) when the module is loaded. The order of
3478 functions with the same priority is not defined.
3480 If the third field is present, non-null, and points to a global variable
3481 or function, the initializer function will only run if the associated
3482 data from the current module is not discarded.
3484 .. _llvmglobaldtors:
3486 The '``llvm.global_dtors``' Global Variable
3487 -------------------------------------------
3489 .. code-block:: llvm
3491 %0 = type { i32, void ()*, i8* }
3492 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor, i8* @data }]
3494 The ``@llvm.global_dtors`` array contains a list of destructor
3495 functions, priorities, and an optional associated global or function.
3496 The functions referenced by this array will be called in descending
3497 order of priority (i.e. highest first) when the module is unloaded. The
3498 order of functions with the same priority is not defined.
3500 If the third field is present, non-null, and points to a global variable
3501 or function, the destructor function will only run if the associated
3502 data from the current module is not discarded.
3504 Instruction Reference
3505 =====================
3507 The LLVM instruction set consists of several different classifications
3508 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
3509 instructions <binaryops>`, :ref:`bitwise binary
3510 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
3511 :ref:`other instructions <otherops>`.
3515 Terminator Instructions
3516 -----------------------
3518 As mentioned :ref:`previously <functionstructure>`, every basic block in a
3519 program ends with a "Terminator" instruction, which indicates which
3520 block should be executed after the current block is finished. These
3521 terminator instructions typically yield a '``void``' value: they produce
3522 control flow, not values (the one exception being the
3523 ':ref:`invoke <i_invoke>`' instruction).
3525 The terminator instructions are: ':ref:`ret <i_ret>`',
3526 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
3527 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
3528 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
3532 '``ret``' Instruction
3533 ^^^^^^^^^^^^^^^^^^^^^
3540 ret <type> <value> ; Return a value from a non-void function
3541 ret void ; Return from void function
3546 The '``ret``' instruction is used to return control flow (and optionally
3547 a value) from a function back to the caller.
3549 There are two forms of the '``ret``' instruction: one that returns a
3550 value and then causes control flow, and one that just causes control
3556 The '``ret``' instruction optionally accepts a single argument, the
3557 return value. The type of the return value must be a ':ref:`first
3558 class <t_firstclass>`' type.
3560 A function is not :ref:`well formed <wellformed>` if it it has a non-void
3561 return type and contains a '``ret``' instruction with no return value or
3562 a return value with a type that does not match its type, or if it has a
3563 void return type and contains a '``ret``' instruction with a return
3569 When the '``ret``' instruction is executed, control flow returns back to
3570 the calling function's context. If the caller is a
3571 ":ref:`call <i_call>`" instruction, execution continues at the
3572 instruction after the call. If the caller was an
3573 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
3574 beginning of the "normal" destination block. If the instruction returns
3575 a value, that value shall set the call or invoke instruction's return
3581 .. code-block:: llvm
3583 ret i32 5 ; Return an integer value of 5
3584 ret void ; Return from a void function
3585 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
3589 '``br``' Instruction
3590 ^^^^^^^^^^^^^^^^^^^^
3597 br i1 <cond>, label <iftrue>, label <iffalse>
3598 br label <dest> ; Unconditional branch
3603 The '``br``' instruction is used to cause control flow to transfer to a
3604 different basic block in the current function. There are two forms of
3605 this instruction, corresponding to a conditional branch and an
3606 unconditional branch.
3611 The conditional branch form of the '``br``' instruction takes a single
3612 '``i1``' value and two '``label``' values. The unconditional form of the
3613 '``br``' instruction takes a single '``label``' value as a target.
3618 Upon execution of a conditional '``br``' instruction, the '``i1``'
3619 argument is evaluated. If the value is ``true``, control flows to the
3620 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
3621 to the '``iffalse``' ``label`` argument.
3626 .. code-block:: llvm
3629 %cond = icmp eq i32 %a, %b
3630 br i1 %cond, label %IfEqual, label %IfUnequal
3638 '``switch``' Instruction
3639 ^^^^^^^^^^^^^^^^^^^^^^^^
3646 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
3651 The '``switch``' instruction is used to transfer control flow to one of
3652 several different places. It is a generalization of the '``br``'
3653 instruction, allowing a branch to occur to one of many possible
3659 The '``switch``' instruction uses three parameters: an integer
3660 comparison value '``value``', a default '``label``' destination, and an
3661 array of pairs of comparison value constants and '``label``'s. The table
3662 is not allowed to contain duplicate constant entries.
3667 The ``switch`` instruction specifies a table of values and destinations.
3668 When the '``switch``' instruction is executed, this table is searched
3669 for the given value. If the value is found, control flow is transferred
3670 to the corresponding destination; otherwise, control flow is transferred
3671 to the default destination.
3676 Depending on properties of the target machine and the particular
3677 ``switch`` instruction, this instruction may be code generated in
3678 different ways. For example, it could be generated as a series of
3679 chained conditional branches or with a lookup table.
3684 .. code-block:: llvm
3686 ; Emulate a conditional br instruction
3687 %Val = zext i1 %value to i32
3688 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
3690 ; Emulate an unconditional br instruction
3691 switch i32 0, label %dest [ ]
3693 ; Implement a jump table:
3694 switch i32 %val, label %otherwise [ i32 0, label %onzero
3696 i32 2, label %ontwo ]
3700 '``indirectbr``' Instruction
3701 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3708 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
3713 The '``indirectbr``' instruction implements an indirect branch to a
3714 label within the current function, whose address is specified by
3715 "``address``". Address must be derived from a
3716 :ref:`blockaddress <blockaddress>` constant.
3721 The '``address``' argument is the address of the label to jump to. The
3722 rest of the arguments indicate the full set of possible destinations
3723 that the address may point to. Blocks are allowed to occur multiple
3724 times in the destination list, though this isn't particularly useful.
3726 This destination list is required so that dataflow analysis has an
3727 accurate understanding of the CFG.
3732 Control transfers to the block specified in the address argument. All
3733 possible destination blocks must be listed in the label list, otherwise
3734 this instruction has undefined behavior. This implies that jumps to
3735 labels defined in other functions have undefined behavior as well.
3740 This is typically implemented with a jump through a register.
3745 .. code-block:: llvm
3747 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
3751 '``invoke``' Instruction
3752 ^^^^^^^^^^^^^^^^^^^^^^^^
3759 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
3760 to label <normal label> unwind label <exception label>
3765 The '``invoke``' instruction causes control to transfer to a specified
3766 function, with the possibility of control flow transfer to either the
3767 '``normal``' label or the '``exception``' label. If the callee function
3768 returns with the "``ret``" instruction, control flow will return to the
3769 "normal" label. If the callee (or any indirect callees) returns via the
3770 ":ref:`resume <i_resume>`" instruction or other exception handling
3771 mechanism, control is interrupted and continued at the dynamically
3772 nearest "exception" label.
3774 The '``exception``' label is a `landing
3775 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
3776 '``exception``' label is required to have the
3777 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
3778 information about the behavior of the program after unwinding happens,
3779 as its first non-PHI instruction. The restrictions on the
3780 "``landingpad``" instruction's tightly couples it to the "``invoke``"
3781 instruction, so that the important information contained within the
3782 "``landingpad``" instruction can't be lost through normal code motion.
3787 This instruction requires several arguments:
3789 #. The optional "cconv" marker indicates which :ref:`calling
3790 convention <callingconv>` the call should use. If none is
3791 specified, the call defaults to using C calling conventions.
3792 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
3793 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
3795 #. '``ptr to function ty``': shall be the signature of the pointer to
3796 function value being invoked. In most cases, this is a direct
3797 function invocation, but indirect ``invoke``'s are just as possible,
3798 branching off an arbitrary pointer to function value.
3799 #. '``function ptr val``': An LLVM value containing a pointer to a
3800 function to be invoked.
3801 #. '``function args``': argument list whose types match the function
3802 signature argument types and parameter attributes. All arguments must
3803 be of :ref:`first class <t_firstclass>` type. If the function signature
3804 indicates the function accepts a variable number of arguments, the
3805 extra arguments can be specified.
3806 #. '``normal label``': the label reached when the called function
3807 executes a '``ret``' instruction.
3808 #. '``exception label``': the label reached when a callee returns via
3809 the :ref:`resume <i_resume>` instruction or other exception handling
3811 #. The optional :ref:`function attributes <fnattrs>` list. Only
3812 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
3813 attributes are valid here.
3818 This instruction is designed to operate as a standard '``call``'
3819 instruction in most regards. The primary difference is that it
3820 establishes an association with a label, which is used by the runtime
3821 library to unwind the stack.
3823 This instruction is used in languages with destructors to ensure that
3824 proper cleanup is performed in the case of either a ``longjmp`` or a
3825 thrown exception. Additionally, this is important for implementation of
3826 '``catch``' clauses in high-level languages that support them.
3828 For the purposes of the SSA form, the definition of the value returned
3829 by the '``invoke``' instruction is deemed to occur on the edge from the
3830 current block to the "normal" label. If the callee unwinds then no
3831 return value is available.
3836 .. code-block:: llvm
3838 %retval = invoke i32 @Test(i32 15) to label %Continue
3839 unwind label %TestCleanup ; i32:retval set
3840 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
3841 unwind label %TestCleanup ; i32:retval set
3845 '``resume``' Instruction
3846 ^^^^^^^^^^^^^^^^^^^^^^^^
3853 resume <type> <value>
3858 The '``resume``' instruction is a terminator instruction that has no
3864 The '``resume``' instruction requires one argument, which must have the
3865 same type as the result of any '``landingpad``' instruction in the same
3871 The '``resume``' instruction resumes propagation of an existing
3872 (in-flight) exception whose unwinding was interrupted with a
3873 :ref:`landingpad <i_landingpad>` instruction.
3878 .. code-block:: llvm
3880 resume { i8*, i32 } %exn
3884 '``unreachable``' Instruction
3885 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3897 The '``unreachable``' instruction has no defined semantics. This
3898 instruction is used to inform the optimizer that a particular portion of
3899 the code is not reachable. This can be used to indicate that the code
3900 after a no-return function cannot be reached, and other facts.
3905 The '``unreachable``' instruction has no defined semantics.
3912 Binary operators are used to do most of the computation in a program.
3913 They require two operands of the same type, execute an operation on
3914 them, and produce a single value. The operands might represent multiple
3915 data, as is the case with the :ref:`vector <t_vector>` data type. The
3916 result value has the same type as its operands.
3918 There are several different binary operators:
3922 '``add``' Instruction
3923 ^^^^^^^^^^^^^^^^^^^^^
3930 <result> = add <ty> <op1>, <op2> ; yields ty:result
3931 <result> = add nuw <ty> <op1>, <op2> ; yields ty:result
3932 <result> = add nsw <ty> <op1>, <op2> ; yields ty:result
3933 <result> = add nuw nsw <ty> <op1>, <op2> ; yields ty:result
3938 The '``add``' instruction returns the sum of its two operands.
3943 The two arguments to the '``add``' instruction must be
3944 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3945 arguments must have identical types.
3950 The value produced is the integer sum of the two operands.
3952 If the sum has unsigned overflow, the result returned is the
3953 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3956 Because LLVM integers use a two's complement representation, this
3957 instruction is appropriate for both signed and unsigned integers.
3959 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3960 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3961 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
3962 unsigned and/or signed overflow, respectively, occurs.
3967 .. code-block:: llvm
3969 <result> = add i32 4, %var ; yields i32:result = 4 + %var
3973 '``fadd``' Instruction
3974 ^^^^^^^^^^^^^^^^^^^^^^
3981 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
3986 The '``fadd``' instruction returns the sum of its two operands.
3991 The two arguments to the '``fadd``' instruction must be :ref:`floating
3992 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3993 Both arguments must have identical types.
3998 The value produced is the floating point sum of the two operands. This
3999 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
4000 which are optimization hints to enable otherwise unsafe floating point
4006 .. code-block:: llvm
4008 <result> = fadd float 4.0, %var ; yields float:result = 4.0 + %var
4010 '``sub``' Instruction
4011 ^^^^^^^^^^^^^^^^^^^^^
4018 <result> = sub <ty> <op1>, <op2> ; yields ty:result
4019 <result> = sub nuw <ty> <op1>, <op2> ; yields ty:result
4020 <result> = sub nsw <ty> <op1>, <op2> ; yields ty:result
4021 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields ty:result
4026 The '``sub``' instruction returns the difference of its two operands.
4028 Note that the '``sub``' instruction is used to represent the '``neg``'
4029 instruction present in most other intermediate representations.
4034 The two arguments to the '``sub``' instruction must be
4035 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4036 arguments must have identical types.
4041 The value produced is the integer difference of the two operands.
4043 If the difference has unsigned overflow, the result returned is the
4044 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
4047 Because LLVM integers use a two's complement representation, this
4048 instruction is appropriate for both signed and unsigned integers.
4050 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4051 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4052 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
4053 unsigned and/or signed overflow, respectively, occurs.
4058 .. code-block:: llvm
4060 <result> = sub i32 4, %var ; yields i32:result = 4 - %var
4061 <result> = sub i32 0, %val ; yields i32:result = -%var
4065 '``fsub``' Instruction
4066 ^^^^^^^^^^^^^^^^^^^^^^
4073 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4078 The '``fsub``' instruction returns the difference of its two operands.
4080 Note that the '``fsub``' instruction is used to represent the '``fneg``'
4081 instruction present in most other intermediate representations.
4086 The two arguments to the '``fsub``' instruction must be :ref:`floating
4087 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4088 Both arguments must have identical types.
4093 The value produced is the floating point difference of the two operands.
4094 This instruction can also take any number of :ref:`fast-math
4095 flags <fastmath>`, which are optimization hints to enable otherwise
4096 unsafe floating point optimizations:
4101 .. code-block:: llvm
4103 <result> = fsub float 4.0, %var ; yields float:result = 4.0 - %var
4104 <result> = fsub float -0.0, %val ; yields float:result = -%var
4106 '``mul``' Instruction
4107 ^^^^^^^^^^^^^^^^^^^^^
4114 <result> = mul <ty> <op1>, <op2> ; yields ty:result
4115 <result> = mul nuw <ty> <op1>, <op2> ; yields ty:result
4116 <result> = mul nsw <ty> <op1>, <op2> ; yields ty:result
4117 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields ty:result
4122 The '``mul``' instruction returns the product of its two operands.
4127 The two arguments to the '``mul``' instruction must be
4128 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4129 arguments must have identical types.
4134 The value produced is the integer product of the two operands.
4136 If the result of the multiplication has unsigned overflow, the result
4137 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
4138 bit width of the result.
4140 Because LLVM integers use a two's complement representation, and the
4141 result is the same width as the operands, this instruction returns the
4142 correct result for both signed and unsigned integers. If a full product
4143 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
4144 sign-extended or zero-extended as appropriate to the width of the full
4147 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
4148 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
4149 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
4150 unsigned and/or signed overflow, respectively, occurs.
4155 .. code-block:: llvm
4157 <result> = mul i32 4, %var ; yields i32:result = 4 * %var
4161 '``fmul``' Instruction
4162 ^^^^^^^^^^^^^^^^^^^^^^
4169 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4174 The '``fmul``' instruction returns the product of its two operands.
4179 The two arguments to the '``fmul``' instruction must be :ref:`floating
4180 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4181 Both arguments must have identical types.
4186 The value produced is the floating point product of the two operands.
4187 This instruction can also take any number of :ref:`fast-math
4188 flags <fastmath>`, which are optimization hints to enable otherwise
4189 unsafe floating point optimizations:
4194 .. code-block:: llvm
4196 <result> = fmul float 4.0, %var ; yields float:result = 4.0 * %var
4198 '``udiv``' Instruction
4199 ^^^^^^^^^^^^^^^^^^^^^^
4206 <result> = udiv <ty> <op1>, <op2> ; yields ty:result
4207 <result> = udiv exact <ty> <op1>, <op2> ; yields ty:result
4212 The '``udiv``' instruction returns the quotient of its two operands.
4217 The two arguments to the '``udiv``' instruction must be
4218 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4219 arguments must have identical types.
4224 The value produced is the unsigned integer quotient of the two operands.
4226 Note that unsigned integer division and signed integer division are
4227 distinct operations; for signed integer division, use '``sdiv``'.
4229 Division by zero leads to undefined behavior.
4231 If the ``exact`` keyword is present, the result value of the ``udiv`` is
4232 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
4233 such, "((a udiv exact b) mul b) == a").
4238 .. code-block:: llvm
4240 <result> = udiv i32 4, %var ; yields i32:result = 4 / %var
4242 '``sdiv``' Instruction
4243 ^^^^^^^^^^^^^^^^^^^^^^
4250 <result> = sdiv <ty> <op1>, <op2> ; yields ty:result
4251 <result> = sdiv exact <ty> <op1>, <op2> ; yields ty:result
4256 The '``sdiv``' instruction returns the quotient of its two operands.
4261 The two arguments to the '``sdiv``' instruction must be
4262 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4263 arguments must have identical types.
4268 The value produced is the signed integer quotient of the two operands
4269 rounded towards zero.
4271 Note that signed integer division and unsigned integer division are
4272 distinct operations; for unsigned integer division, use '``udiv``'.
4274 Division by zero leads to undefined behavior. Overflow also leads to
4275 undefined behavior; this is a rare case, but can occur, for example, by
4276 doing a 32-bit division of -2147483648 by -1.
4278 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
4279 a :ref:`poison value <poisonvalues>` if the result would be rounded.
4284 .. code-block:: llvm
4286 <result> = sdiv i32 4, %var ; yields i32:result = 4 / %var
4290 '``fdiv``' Instruction
4291 ^^^^^^^^^^^^^^^^^^^^^^
4298 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4303 The '``fdiv``' instruction returns the quotient of its two operands.
4308 The two arguments to the '``fdiv``' instruction must be :ref:`floating
4309 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4310 Both arguments must have identical types.
4315 The value produced is the floating point quotient of the two operands.
4316 This instruction can also take any number of :ref:`fast-math
4317 flags <fastmath>`, which are optimization hints to enable otherwise
4318 unsafe floating point optimizations:
4323 .. code-block:: llvm
4325 <result> = fdiv float 4.0, %var ; yields float:result = 4.0 / %var
4327 '``urem``' Instruction
4328 ^^^^^^^^^^^^^^^^^^^^^^
4335 <result> = urem <ty> <op1>, <op2> ; yields ty:result
4340 The '``urem``' instruction returns the remainder from the unsigned
4341 division of its two arguments.
4346 The two arguments to the '``urem``' instruction must be
4347 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4348 arguments must have identical types.
4353 This instruction returns the unsigned integer *remainder* of a division.
4354 This instruction always performs an unsigned division to get the
4357 Note that unsigned integer remainder and signed integer remainder are
4358 distinct operations; for signed integer remainder, use '``srem``'.
4360 Taking the remainder of a division by zero leads to undefined behavior.
4365 .. code-block:: llvm
4367 <result> = urem i32 4, %var ; yields i32:result = 4 % %var
4369 '``srem``' Instruction
4370 ^^^^^^^^^^^^^^^^^^^^^^
4377 <result> = srem <ty> <op1>, <op2> ; yields ty:result
4382 The '``srem``' instruction returns the remainder from the signed
4383 division of its two operands. This instruction can also take
4384 :ref:`vector <t_vector>` versions of the values in which case the elements
4390 The two arguments to the '``srem``' instruction must be
4391 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4392 arguments must have identical types.
4397 This instruction returns the *remainder* of a division (where the result
4398 is either zero or has the same sign as the dividend, ``op1``), not the
4399 *modulo* operator (where the result is either zero or has the same sign
4400 as the divisor, ``op2``) of a value. For more information about the
4401 difference, see `The Math
4402 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
4403 table of how this is implemented in various languages, please see
4405 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
4407 Note that signed integer remainder and unsigned integer remainder are
4408 distinct operations; for unsigned integer remainder, use '``urem``'.
4410 Taking the remainder of a division by zero leads to undefined behavior.
4411 Overflow also leads to undefined behavior; this is a rare case, but can
4412 occur, for example, by taking the remainder of a 32-bit division of
4413 -2147483648 by -1. (The remainder doesn't actually overflow, but this
4414 rule lets srem be implemented using instructions that return both the
4415 result of the division and the remainder.)
4420 .. code-block:: llvm
4422 <result> = srem i32 4, %var ; yields i32:result = 4 % %var
4426 '``frem``' Instruction
4427 ^^^^^^^^^^^^^^^^^^^^^^
4434 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields ty:result
4439 The '``frem``' instruction returns the remainder from the division of
4445 The two arguments to the '``frem``' instruction must be :ref:`floating
4446 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
4447 Both arguments must have identical types.
4452 This instruction returns the *remainder* of a division. The remainder
4453 has the same sign as the dividend. This instruction can also take any
4454 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
4455 to enable otherwise unsafe floating point optimizations:
4460 .. code-block:: llvm
4462 <result> = frem float 4.0, %var ; yields float:result = 4.0 % %var
4466 Bitwise Binary Operations
4467 -------------------------
4469 Bitwise binary operators are used to do various forms of bit-twiddling
4470 in a program. They are generally very efficient instructions and can
4471 commonly be strength reduced from other instructions. They require two
4472 operands of the same type, execute an operation on them, and produce a
4473 single value. The resulting value is the same type as its operands.
4475 '``shl``' Instruction
4476 ^^^^^^^^^^^^^^^^^^^^^
4483 <result> = shl <ty> <op1>, <op2> ; yields ty:result
4484 <result> = shl nuw <ty> <op1>, <op2> ; yields ty:result
4485 <result> = shl nsw <ty> <op1>, <op2> ; yields ty:result
4486 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields ty:result
4491 The '``shl``' instruction returns the first operand shifted to the left
4492 a specified number of bits.
4497 Both arguments to the '``shl``' instruction must be the same
4498 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4499 '``op2``' is treated as an unsigned value.
4504 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
4505 where ``n`` is the width of the result. If ``op2`` is (statically or
4506 dynamically) negative or equal to or larger than the number of bits in
4507 ``op1``, the result is undefined. If the arguments are vectors, each
4508 vector element of ``op1`` is shifted by the corresponding shift amount
4511 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
4512 value <poisonvalues>` if it shifts out any non-zero bits. If the
4513 ``nsw`` keyword is present, then the shift produces a :ref:`poison
4514 value <poisonvalues>` if it shifts out any bits that disagree with the
4515 resultant sign bit. As such, NUW/NSW have the same semantics as they
4516 would if the shift were expressed as a mul instruction with the same
4517 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
4522 .. code-block:: llvm
4524 <result> = shl i32 4, %var ; yields i32: 4 << %var
4525 <result> = shl i32 4, 2 ; yields i32: 16
4526 <result> = shl i32 1, 10 ; yields i32: 1024
4527 <result> = shl i32 1, 32 ; undefined
4528 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
4530 '``lshr``' Instruction
4531 ^^^^^^^^^^^^^^^^^^^^^^
4538 <result> = lshr <ty> <op1>, <op2> ; yields ty:result
4539 <result> = lshr exact <ty> <op1>, <op2> ; yields ty:result
4544 The '``lshr``' instruction (logical shift right) returns the first
4545 operand shifted to the right a specified number of bits with zero fill.
4550 Both arguments to the '``lshr``' instruction must be the same
4551 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4552 '``op2``' is treated as an unsigned value.
4557 This instruction always performs a logical shift right operation. The
4558 most significant bits of the result will be filled with zero bits after
4559 the shift. If ``op2`` is (statically or dynamically) equal to or larger
4560 than the number of bits in ``op1``, the result is undefined. If the
4561 arguments are vectors, each vector element of ``op1`` is shifted by the
4562 corresponding shift amount in ``op2``.
4564 If the ``exact`` keyword is present, the result value of the ``lshr`` is
4565 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4571 .. code-block:: llvm
4573 <result> = lshr i32 4, 1 ; yields i32:result = 2
4574 <result> = lshr i32 4, 2 ; yields i32:result = 1
4575 <result> = lshr i8 4, 3 ; yields i8:result = 0
4576 <result> = lshr i8 -2, 1 ; yields i8:result = 0x7F
4577 <result> = lshr i32 1, 32 ; undefined
4578 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
4580 '``ashr``' Instruction
4581 ^^^^^^^^^^^^^^^^^^^^^^
4588 <result> = ashr <ty> <op1>, <op2> ; yields ty:result
4589 <result> = ashr exact <ty> <op1>, <op2> ; yields ty:result
4594 The '``ashr``' instruction (arithmetic shift right) returns the first
4595 operand shifted to the right a specified number of bits with sign
4601 Both arguments to the '``ashr``' instruction must be the same
4602 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4603 '``op2``' is treated as an unsigned value.
4608 This instruction always performs an arithmetic shift right operation,
4609 The most significant bits of the result will be filled with the sign bit
4610 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
4611 than the number of bits in ``op1``, the result is undefined. If the
4612 arguments are vectors, each vector element of ``op1`` is shifted by the
4613 corresponding shift amount in ``op2``.
4615 If the ``exact`` keyword is present, the result value of the ``ashr`` is
4616 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4622 .. code-block:: llvm
4624 <result> = ashr i32 4, 1 ; yields i32:result = 2
4625 <result> = ashr i32 4, 2 ; yields i32:result = 1
4626 <result> = ashr i8 4, 3 ; yields i8:result = 0
4627 <result> = ashr i8 -2, 1 ; yields i8:result = -1
4628 <result> = ashr i32 1, 32 ; undefined
4629 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
4631 '``and``' Instruction
4632 ^^^^^^^^^^^^^^^^^^^^^
4639 <result> = and <ty> <op1>, <op2> ; yields ty:result
4644 The '``and``' instruction returns the bitwise logical and of its two
4650 The two arguments to the '``and``' instruction must be
4651 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4652 arguments must have identical types.
4657 The truth table used for the '``and``' instruction is:
4674 .. code-block:: llvm
4676 <result> = and i32 4, %var ; yields i32:result = 4 & %var
4677 <result> = and i32 15, 40 ; yields i32:result = 8
4678 <result> = and i32 4, 8 ; yields i32:result = 0
4680 '``or``' Instruction
4681 ^^^^^^^^^^^^^^^^^^^^
4688 <result> = or <ty> <op1>, <op2> ; yields ty:result
4693 The '``or``' instruction returns the bitwise logical inclusive or of its
4699 The two arguments to the '``or``' instruction must be
4700 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4701 arguments must have identical types.
4706 The truth table used for the '``or``' instruction is:
4725 <result> = or i32 4, %var ; yields i32:result = 4 | %var
4726 <result> = or i32 15, 40 ; yields i32:result = 47
4727 <result> = or i32 4, 8 ; yields i32:result = 12
4729 '``xor``' Instruction
4730 ^^^^^^^^^^^^^^^^^^^^^
4737 <result> = xor <ty> <op1>, <op2> ; yields ty:result
4742 The '``xor``' instruction returns the bitwise logical exclusive or of
4743 its two operands. The ``xor`` is used to implement the "one's
4744 complement" operation, which is the "~" operator in C.
4749 The two arguments to the '``xor``' instruction must be
4750 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4751 arguments must have identical types.
4756 The truth table used for the '``xor``' instruction is:
4773 .. code-block:: llvm
4775 <result> = xor i32 4, %var ; yields i32:result = 4 ^ %var
4776 <result> = xor i32 15, 40 ; yields i32:result = 39
4777 <result> = xor i32 4, 8 ; yields i32:result = 12
4778 <result> = xor i32 %V, -1 ; yields i32:result = ~%V
4783 LLVM supports several instructions to represent vector operations in a
4784 target-independent manner. These instructions cover the element-access
4785 and vector-specific operations needed to process vectors effectively.
4786 While LLVM does directly support these vector operations, many
4787 sophisticated algorithms will want to use target-specific intrinsics to
4788 take full advantage of a specific target.
4790 .. _i_extractelement:
4792 '``extractelement``' Instruction
4793 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4800 <result> = extractelement <n x <ty>> <val>, <ty2> <idx> ; yields <ty>
4805 The '``extractelement``' instruction extracts a single scalar element
4806 from a vector at a specified index.
4811 The first operand of an '``extractelement``' instruction is a value of
4812 :ref:`vector <t_vector>` type. The second operand is an index indicating
4813 the position from which to extract the element. The index may be a
4814 variable of any integer type.
4819 The result is a scalar of the same type as the element type of ``val``.
4820 Its value is the value at position ``idx`` of ``val``. If ``idx``
4821 exceeds the length of ``val``, the results are undefined.
4826 .. code-block:: llvm
4828 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
4830 .. _i_insertelement:
4832 '``insertelement``' Instruction
4833 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4840 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, <ty2> <idx> ; yields <n x <ty>>
4845 The '``insertelement``' instruction inserts a scalar element into a
4846 vector at a specified index.
4851 The first operand of an '``insertelement``' instruction is a value of
4852 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
4853 type must equal the element type of the first operand. The third operand
4854 is an index indicating the position at which to insert the value. The
4855 index may be a variable of any integer type.
4860 The result is a vector of the same type as ``val``. Its element values
4861 are those of ``val`` except at position ``idx``, where it gets the value
4862 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
4868 .. code-block:: llvm
4870 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
4872 .. _i_shufflevector:
4874 '``shufflevector``' Instruction
4875 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4882 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
4887 The '``shufflevector``' instruction constructs a permutation of elements
4888 from two input vectors, returning a vector with the same element type as
4889 the input and length that is the same as the shuffle mask.
4894 The first two operands of a '``shufflevector``' instruction are vectors
4895 with the same type. The third argument is a shuffle mask whose element
4896 type is always 'i32'. The result of the instruction is a vector whose
4897 length is the same as the shuffle mask and whose element type is the
4898 same as the element type of the first two operands.
4900 The shuffle mask operand is required to be a constant vector with either
4901 constant integer or undef values.
4906 The elements of the two input vectors are numbered from left to right
4907 across both of the vectors. The shuffle mask operand specifies, for each
4908 element of the result vector, which element of the two input vectors the
4909 result element gets. The element selector may be undef (meaning "don't
4910 care") and the second operand may be undef if performing a shuffle from
4916 .. code-block:: llvm
4918 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4919 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
4920 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
4921 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
4922 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
4923 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
4924 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4925 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
4927 Aggregate Operations
4928 --------------------
4930 LLVM supports several instructions for working with
4931 :ref:`aggregate <t_aggregate>` values.
4935 '``extractvalue``' Instruction
4936 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4943 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
4948 The '``extractvalue``' instruction extracts the value of a member field
4949 from an :ref:`aggregate <t_aggregate>` value.
4954 The first operand of an '``extractvalue``' instruction is a value of
4955 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
4956 constant indices to specify which value to extract in a similar manner
4957 as indices in a '``getelementptr``' instruction.
4959 The major differences to ``getelementptr`` indexing are:
4961 - Since the value being indexed is not a pointer, the first index is
4962 omitted and assumed to be zero.
4963 - At least one index must be specified.
4964 - Not only struct indices but also array indices must be in bounds.
4969 The result is the value at the position in the aggregate specified by
4975 .. code-block:: llvm
4977 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
4981 '``insertvalue``' Instruction
4982 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4989 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
4994 The '``insertvalue``' instruction inserts a value into a member field in
4995 an :ref:`aggregate <t_aggregate>` value.
5000 The first operand of an '``insertvalue``' instruction is a value of
5001 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
5002 a first-class value to insert. The following operands are constant
5003 indices indicating the position at which to insert the value in a
5004 similar manner as indices in a '``extractvalue``' instruction. The value
5005 to insert must have the same type as the value identified by the
5011 The result is an aggregate of the same type as ``val``. Its value is
5012 that of ``val`` except that the value at the position specified by the
5013 indices is that of ``elt``.
5018 .. code-block:: llvm
5020 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
5021 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
5022 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 ; yields {i32 1, float %val}
5026 Memory Access and Addressing Operations
5027 ---------------------------------------
5029 A key design point of an SSA-based representation is how it represents
5030 memory. In LLVM, no memory locations are in SSA form, which makes things
5031 very simple. This section describes how to read, write, and allocate
5036 '``alloca``' Instruction
5037 ^^^^^^^^^^^^^^^^^^^^^^^^
5044 <result> = alloca [inalloca] <type> [, <ty> <NumElements>] [, align <alignment>] ; yields type*:result
5049 The '``alloca``' instruction allocates memory on the stack frame of the
5050 currently executing function, to be automatically released when this
5051 function returns to its caller. The object is always allocated in the
5052 generic address space (address space zero).
5057 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
5058 bytes of memory on the runtime stack, returning a pointer of the
5059 appropriate type to the program. If "NumElements" is specified, it is
5060 the number of elements allocated, otherwise "NumElements" is defaulted
5061 to be one. If a constant alignment is specified, the value result of the
5062 allocation is guaranteed to be aligned to at least that boundary. The
5063 alignment may not be greater than ``1 << 29``. If not specified, or if
5064 zero, the target can choose to align the allocation on any convenient
5065 boundary compatible with the type.
5067 '``type``' may be any sized type.
5072 Memory is allocated; a pointer is returned. The operation is undefined
5073 if there is insufficient stack space for the allocation. '``alloca``'d
5074 memory is automatically released when the function returns. The
5075 '``alloca``' instruction is commonly used to represent automatic
5076 variables that must have an address available. When the function returns
5077 (either with the ``ret`` or ``resume`` instructions), the memory is
5078 reclaimed. Allocating zero bytes is legal, but the result is undefined.
5079 The order in which memory is allocated (ie., which way the stack grows)
5085 .. code-block:: llvm
5087 %ptr = alloca i32 ; yields i32*:ptr
5088 %ptr = alloca i32, i32 4 ; yields i32*:ptr
5089 %ptr = alloca i32, i32 4, align 1024 ; yields i32*:ptr
5090 %ptr = alloca i32, align 1024 ; yields i32*:ptr
5094 '``load``' Instruction
5095 ^^^^^^^^^^^^^^^^^^^^^^
5102 <result> = load [volatile] <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>]
5103 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
5104 !<index> = !{ i32 1 }
5109 The '``load``' instruction is used to read from memory.
5114 The argument to the ``load`` instruction specifies the memory address
5115 from which to load. The pointer must point to a :ref:`first
5116 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
5117 then the optimizer is not allowed to modify the number or order of
5118 execution of this ``load`` with other :ref:`volatile
5119 operations <volatile>`.
5121 If the ``load`` is marked as ``atomic``, it takes an extra
5122 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5123 ``release`` and ``acq_rel`` orderings are not valid on ``load``
5124 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5125 when they may see multiple atomic stores. The type of the pointee must
5126 be an integer type whose bit width is a power of two greater than or
5127 equal to eight and less than or equal to a target-specific size limit.
5128 ``align`` must be explicitly specified on atomic loads, and the load has
5129 undefined behavior if the alignment is not set to a value which is at
5130 least the size in bytes of the pointee. ``!nontemporal`` does not have
5131 any defined semantics for atomic loads.
5133 The optional constant ``align`` argument specifies the alignment of the
5134 operation (that is, the alignment of the memory address). A value of 0
5135 or an omitted ``align`` argument means that the operation has the ABI
5136 alignment for the target. It is the responsibility of the code emitter
5137 to ensure that the alignment information is correct. Overestimating the
5138 alignment results in undefined behavior. Underestimating the alignment
5139 may produce less efficient code. An alignment of 1 is always safe. The
5140 maximum possible alignment is ``1 << 29``.
5142 The optional ``!nontemporal`` metadata must reference a single
5143 metadata name ``<index>`` corresponding to a metadata node with one
5144 ``i32`` entry of value 1. The existence of the ``!nontemporal``
5145 metadata on the instruction tells the optimizer and code generator
5146 that this load is not expected to be reused in the cache. The code
5147 generator may select special instructions to save cache bandwidth, such
5148 as the ``MOVNT`` instruction on x86.
5150 The optional ``!invariant.load`` metadata must reference a single
5151 metadata name ``<index>`` corresponding to a metadata node with no
5152 entries. The existence of the ``!invariant.load`` metadata on the
5153 instruction tells the optimizer and code generator that this load
5154 address points to memory which does not change value during program
5155 execution. The optimizer may then move this load around, for example, by
5156 hoisting it out of loops using loop invariant code motion.
5161 The location of memory pointed to is loaded. If the value being loaded
5162 is of scalar type then the number of bytes read does not exceed the
5163 minimum number of bytes needed to hold all bits of the type. For
5164 example, loading an ``i24`` reads at most three bytes. When loading a
5165 value of a type like ``i20`` with a size that is not an integral number
5166 of bytes, the result is undefined if the value was not originally
5167 written using a store of the same type.
5172 .. code-block:: llvm
5174 %ptr = alloca i32 ; yields i32*:ptr
5175 store i32 3, i32* %ptr ; yields void
5176 %val = load i32* %ptr ; yields i32:val = i32 3
5180 '``store``' Instruction
5181 ^^^^^^^^^^^^^^^^^^^^^^^
5188 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields void
5189 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields void
5194 The '``store``' instruction is used to write to memory.
5199 There are two arguments to the ``store`` instruction: a value to store
5200 and an address at which to store it. The type of the ``<pointer>``
5201 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
5202 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
5203 then the optimizer is not allowed to modify the number or order of
5204 execution of this ``store`` with other :ref:`volatile
5205 operations <volatile>`.
5207 If the ``store`` is marked as ``atomic``, it takes an extra
5208 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
5209 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
5210 instructions. Atomic loads produce :ref:`defined <memmodel>` results
5211 when they may see multiple atomic stores. The type of the pointee must
5212 be an integer type whose bit width is a power of two greater than or
5213 equal to eight and less than or equal to a target-specific size limit.
5214 ``align`` must be explicitly specified on atomic stores, and the store
5215 has undefined behavior if the alignment is not set to a value which is
5216 at least the size in bytes of the pointee. ``!nontemporal`` does not
5217 have any defined semantics for atomic stores.
5219 The optional constant ``align`` argument specifies the alignment of the
5220 operation (that is, the alignment of the memory address). A value of 0
5221 or an omitted ``align`` argument means that the operation has the ABI
5222 alignment for the target. It is the responsibility of the code emitter
5223 to ensure that the alignment information is correct. Overestimating the
5224 alignment results in undefined behavior. Underestimating the
5225 alignment may produce less efficient code. An alignment of 1 is always
5226 safe. The maximum possible alignment is ``1 << 29``.
5228 The optional ``!nontemporal`` metadata must reference a single metadata
5229 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
5230 value 1. The existence of the ``!nontemporal`` metadata on the instruction
5231 tells the optimizer and code generator that this load is not expected to
5232 be reused in the cache. The code generator may select special
5233 instructions to save cache bandwidth, such as the MOVNT instruction on
5239 The contents of memory are updated to contain ``<value>`` at the
5240 location specified by the ``<pointer>`` operand. If ``<value>`` is
5241 of scalar type then the number of bytes written does not exceed the
5242 minimum number of bytes needed to hold all bits of the type. For
5243 example, storing an ``i24`` writes at most three bytes. When writing a
5244 value of a type like ``i20`` with a size that is not an integral number
5245 of bytes, it is unspecified what happens to the extra bits that do not
5246 belong to the type, but they will typically be overwritten.
5251 .. code-block:: llvm
5253 %ptr = alloca i32 ; yields i32*:ptr
5254 store i32 3, i32* %ptr ; yields void
5255 %val = load i32* %ptr ; yields i32:val = i32 3
5259 '``fence``' Instruction
5260 ^^^^^^^^^^^^^^^^^^^^^^^
5267 fence [singlethread] <ordering> ; yields void
5272 The '``fence``' instruction is used to introduce happens-before edges
5278 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
5279 defines what *synchronizes-with* edges they add. They can only be given
5280 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
5285 A fence A which has (at least) ``release`` ordering semantics
5286 *synchronizes with* a fence B with (at least) ``acquire`` ordering
5287 semantics if and only if there exist atomic operations X and Y, both
5288 operating on some atomic object M, such that A is sequenced before X, X
5289 modifies M (either directly or through some side effect of a sequence
5290 headed by X), Y is sequenced before B, and Y observes M. This provides a
5291 *happens-before* dependency between A and B. Rather than an explicit
5292 ``fence``, one (but not both) of the atomic operations X or Y might
5293 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
5294 still *synchronize-with* the explicit ``fence`` and establish the
5295 *happens-before* edge.
5297 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
5298 ``acquire`` and ``release`` semantics specified above, participates in
5299 the global program order of other ``seq_cst`` operations and/or fences.
5301 The optional ":ref:`singlethread <singlethread>`" argument specifies
5302 that the fence only synchronizes with other fences in the same thread.
5303 (This is useful for interacting with signal handlers.)
5308 .. code-block:: llvm
5310 fence acquire ; yields void
5311 fence singlethread seq_cst ; yields void
5315 '``cmpxchg``' Instruction
5316 ^^^^^^^^^^^^^^^^^^^^^^^^^
5323 cmpxchg [weak] [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <success ordering> <failure ordering> ; yields { ty, i1 }
5328 The '``cmpxchg``' instruction is used to atomically modify memory. It
5329 loads a value in memory and compares it to a given value. If they are
5330 equal, it tries to store a new value into the memory.
5335 There are three arguments to the '``cmpxchg``' instruction: an address
5336 to operate on, a value to compare to the value currently be at that
5337 address, and a new value to place at that address if the compared values
5338 are equal. The type of '<cmp>' must be an integer type whose bit width
5339 is a power of two greater than or equal to eight and less than or equal
5340 to a target-specific size limit. '<cmp>' and '<new>' must have the same
5341 type, and the type of '<pointer>' must be a pointer to that type. If the
5342 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
5343 to modify the number or order of execution of this ``cmpxchg`` with
5344 other :ref:`volatile operations <volatile>`.
5346 The success and failure :ref:`ordering <ordering>` arguments specify how this
5347 ``cmpxchg`` synchronizes with other atomic operations. Both ordering parameters
5348 must be at least ``monotonic``, the ordering constraint on failure must be no
5349 stronger than that on success, and the failure ordering cannot be either
5350 ``release`` or ``acq_rel``.
5352 The optional "``singlethread``" argument declares that the ``cmpxchg``
5353 is only atomic with respect to code (usually signal handlers) running in
5354 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
5355 respect to all other code in the system.
5357 The pointer passed into cmpxchg must have alignment greater than or
5358 equal to the size in memory of the operand.
5363 The contents of memory at the location specified by the '``<pointer>``' operand
5364 is read and compared to '``<cmp>``'; if the read value is the equal, the
5365 '``<new>``' is written. The original value at the location is returned, together
5366 with a flag indicating success (true) or failure (false).
5368 If the cmpxchg operation is marked as ``weak`` then a spurious failure is
5369 permitted: the operation may not write ``<new>`` even if the comparison
5372 If the cmpxchg operation is strong (the default), the i1 value is 1 if and only
5373 if the value loaded equals ``cmp``.
5375 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose of
5376 identifying release sequences. A failed ``cmpxchg`` is equivalent to an atomic
5377 load with an ordering parameter determined the second ordering parameter.
5382 .. code-block:: llvm
5385 %orig = atomic load i32* %ptr unordered ; yields i32
5389 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
5390 %squared = mul i32 %cmp, %cmp
5391 %val_success = cmpxchg i32* %ptr, i32 %cmp, i32 %squared acq_rel monotonic ; yields { i32, i1 }
5392 %value_loaded = extractvalue { i32, i1 } %val_success, 0
5393 %success = extractvalue { i32, i1 } %val_success, 1
5394 br i1 %success, label %done, label %loop
5401 '``atomicrmw``' Instruction
5402 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
5409 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields ty
5414 The '``atomicrmw``' instruction is used to atomically modify memory.
5419 There are three arguments to the '``atomicrmw``' instruction: an
5420 operation to apply, an address whose value to modify, an argument to the
5421 operation. The operation must be one of the following keywords:
5435 The type of '<value>' must be an integer type whose bit width is a power
5436 of two greater than or equal to eight and less than or equal to a
5437 target-specific size limit. The type of the '``<pointer>``' operand must
5438 be a pointer to that type. If the ``atomicrmw`` is marked as
5439 ``volatile``, then the optimizer is not allowed to modify the number or
5440 order of execution of this ``atomicrmw`` with other :ref:`volatile
5441 operations <volatile>`.
5446 The contents of memory at the location specified by the '``<pointer>``'
5447 operand are atomically read, modified, and written back. The original
5448 value at the location is returned. The modification is specified by the
5451 - xchg: ``*ptr = val``
5452 - add: ``*ptr = *ptr + val``
5453 - sub: ``*ptr = *ptr - val``
5454 - and: ``*ptr = *ptr & val``
5455 - nand: ``*ptr = ~(*ptr & val)``
5456 - or: ``*ptr = *ptr | val``
5457 - xor: ``*ptr = *ptr ^ val``
5458 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
5459 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
5460 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
5462 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
5468 .. code-block:: llvm
5470 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields i32
5472 .. _i_getelementptr:
5474 '``getelementptr``' Instruction
5475 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5482 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
5483 <result> = getelementptr inbounds <pty>* <ptrval>{, <ty> <idx>}*
5484 <result> = getelementptr <ptr vector> ptrval, <vector index type> idx
5489 The '``getelementptr``' instruction is used to get the address of a
5490 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
5491 address calculation only and does not access memory.
5496 The first argument is always a pointer or a vector of pointers, and
5497 forms the basis of the calculation. The remaining arguments are indices
5498 that indicate which of the elements of the aggregate object are indexed.
5499 The interpretation of each index is dependent on the type being indexed
5500 into. The first index always indexes the pointer value given as the
5501 first argument, the second index indexes a value of the type pointed to
5502 (not necessarily the value directly pointed to, since the first index
5503 can be non-zero), etc. The first type indexed into must be a pointer
5504 value, subsequent types can be arrays, vectors, and structs. Note that
5505 subsequent types being indexed into can never be pointers, since that
5506 would require loading the pointer before continuing calculation.
5508 The type of each index argument depends on the type it is indexing into.
5509 When indexing into a (optionally packed) structure, only ``i32`` integer
5510 **constants** are allowed (when using a vector of indices they must all
5511 be the **same** ``i32`` integer constant). When indexing into an array,
5512 pointer or vector, integers of any width are allowed, and they are not
5513 required to be constant. These integers are treated as signed values
5516 For example, let's consider a C code fragment and how it gets compiled
5532 int *foo(struct ST *s) {
5533 return &s[1].Z.B[5][13];
5536 The LLVM code generated by Clang is:
5538 .. code-block:: llvm
5540 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
5541 %struct.ST = type { i32, double, %struct.RT }
5543 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
5545 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5552 In the example above, the first index is indexing into the
5553 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
5554 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
5555 indexes into the third element of the structure, yielding a
5556 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
5557 structure. The third index indexes into the second element of the
5558 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
5559 dimensions of the array are subscripted into, yielding an '``i32``'
5560 type. The '``getelementptr``' instruction returns a pointer to this
5561 element, thus computing a value of '``i32*``' type.
5563 Note that it is perfectly legal to index partially through a structure,
5564 returning a pointer to an inner element. Because of this, the LLVM code
5565 for the given testcase is equivalent to:
5567 .. code-block:: llvm
5569 define i32* @foo(%struct.ST* %s) {
5570 %t1 = getelementptr %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
5571 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
5572 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
5573 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
5574 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
5578 If the ``inbounds`` keyword is present, the result value of the
5579 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
5580 pointer is not an *in bounds* address of an allocated object, or if any
5581 of the addresses that would be formed by successive addition of the
5582 offsets implied by the indices to the base address with infinitely
5583 precise signed arithmetic are not an *in bounds* address of that
5584 allocated object. The *in bounds* addresses for an allocated object are
5585 all the addresses that point into the object, plus the address one byte
5586 past the end. In cases where the base is a vector of pointers the
5587 ``inbounds`` keyword applies to each of the computations element-wise.
5589 If the ``inbounds`` keyword is not present, the offsets are added to the
5590 base address with silently-wrapping two's complement arithmetic. If the
5591 offsets have a different width from the pointer, they are sign-extended
5592 or truncated to the width of the pointer. The result value of the
5593 ``getelementptr`` may be outside the object pointed to by the base
5594 pointer. The result value may not necessarily be used to access memory
5595 though, even if it happens to point into allocated storage. See the
5596 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
5599 The getelementptr instruction is often confusing. For some more insight
5600 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
5605 .. code-block:: llvm
5607 ; yields [12 x i8]*:aptr
5608 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5610 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
5612 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
5614 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
5616 In cases where the pointer argument is a vector of pointers, each index
5617 must be a vector with the same number of elements. For example:
5619 .. code-block:: llvm
5621 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5623 Conversion Operations
5624 ---------------------
5626 The instructions in this category are the conversion instructions
5627 (casting) which all take a single operand and a type. They perform
5628 various bit conversions on the operand.
5630 '``trunc .. to``' Instruction
5631 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5638 <result> = trunc <ty> <value> to <ty2> ; yields ty2
5643 The '``trunc``' instruction truncates its operand to the type ``ty2``.
5648 The '``trunc``' instruction takes a value to trunc, and a type to trunc
5649 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
5650 of the same number of integers. The bit size of the ``value`` must be
5651 larger than the bit size of the destination type, ``ty2``. Equal sized
5652 types are not allowed.
5657 The '``trunc``' instruction truncates the high order bits in ``value``
5658 and converts the remaining bits to ``ty2``. Since the source size must
5659 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
5660 It will always truncate bits.
5665 .. code-block:: llvm
5667 %X = trunc i32 257 to i8 ; yields i8:1
5668 %Y = trunc i32 123 to i1 ; yields i1:true
5669 %Z = trunc i32 122 to i1 ; yields i1:false
5670 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
5672 '``zext .. to``' Instruction
5673 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5680 <result> = zext <ty> <value> to <ty2> ; yields ty2
5685 The '``zext``' instruction zero extends its operand to type ``ty2``.
5690 The '``zext``' instruction takes a value to cast, and a type to cast it
5691 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5692 the same number of integers. The bit size of the ``value`` must be
5693 smaller than the bit size of the destination type, ``ty2``.
5698 The ``zext`` fills the high order bits of the ``value`` with zero bits
5699 until it reaches the size of the destination type, ``ty2``.
5701 When zero extending from i1, the result will always be either 0 or 1.
5706 .. code-block:: llvm
5708 %X = zext i32 257 to i64 ; yields i64:257
5709 %Y = zext i1 true to i32 ; yields i32:1
5710 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5712 '``sext .. to``' Instruction
5713 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5720 <result> = sext <ty> <value> to <ty2> ; yields ty2
5725 The '``sext``' sign extends ``value`` to the type ``ty2``.
5730 The '``sext``' instruction takes a value to cast, and a type to cast it
5731 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5732 the same number of integers. The bit size of the ``value`` must be
5733 smaller than the bit size of the destination type, ``ty2``.
5738 The '``sext``' instruction performs a sign extension by copying the sign
5739 bit (highest order bit) of the ``value`` until it reaches the bit size
5740 of the type ``ty2``.
5742 When sign extending from i1, the extension always results in -1 or 0.
5747 .. code-block:: llvm
5749 %X = sext i8 -1 to i16 ; yields i16 :65535
5750 %Y = sext i1 true to i32 ; yields i32:-1
5751 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5753 '``fptrunc .. to``' Instruction
5754 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5761 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
5766 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
5771 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
5772 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
5773 The size of ``value`` must be larger than the size of ``ty2``. This
5774 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
5779 The '``fptrunc``' instruction truncates a ``value`` from a larger
5780 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
5781 point <t_floating>` type. If the value cannot fit within the
5782 destination type, ``ty2``, then the results are undefined.
5787 .. code-block:: llvm
5789 %X = fptrunc double 123.0 to float ; yields float:123.0
5790 %Y = fptrunc double 1.0E+300 to float ; yields undefined
5792 '``fpext .. to``' Instruction
5793 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5800 <result> = fpext <ty> <value> to <ty2> ; yields ty2
5805 The '``fpext``' extends a floating point ``value`` to a larger floating
5811 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
5812 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
5813 to. The source type must be smaller than the destination type.
5818 The '``fpext``' instruction extends the ``value`` from a smaller
5819 :ref:`floating point <t_floating>` type to a larger :ref:`floating
5820 point <t_floating>` type. The ``fpext`` cannot be used to make a
5821 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
5822 *no-op cast* for a floating point cast.
5827 .. code-block:: llvm
5829 %X = fpext float 3.125 to double ; yields double:3.125000e+00
5830 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
5832 '``fptoui .. to``' Instruction
5833 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5840 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
5845 The '``fptoui``' converts a floating point ``value`` to its unsigned
5846 integer equivalent of type ``ty2``.
5851 The '``fptoui``' instruction takes a value to cast, which must be a
5852 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5853 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5854 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5855 type with the same number of elements as ``ty``
5860 The '``fptoui``' instruction converts its :ref:`floating
5861 point <t_floating>` operand into the nearest (rounding towards zero)
5862 unsigned integer value. If the value cannot fit in ``ty2``, the results
5868 .. code-block:: llvm
5870 %X = fptoui double 123.0 to i32 ; yields i32:123
5871 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
5872 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
5874 '``fptosi .. to``' Instruction
5875 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5882 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
5887 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
5888 ``value`` to type ``ty2``.
5893 The '``fptosi``' instruction takes a value to cast, which must be a
5894 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5895 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5896 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5897 type with the same number of elements as ``ty``
5902 The '``fptosi``' instruction converts its :ref:`floating
5903 point <t_floating>` operand into the nearest (rounding towards zero)
5904 signed integer value. If the value cannot fit in ``ty2``, the results
5910 .. code-block:: llvm
5912 %X = fptosi double -123.0 to i32 ; yields i32:-123
5913 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
5914 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
5916 '``uitofp .. to``' Instruction
5917 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5924 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
5929 The '``uitofp``' instruction regards ``value`` as an unsigned integer
5930 and converts that value to the ``ty2`` type.
5935 The '``uitofp``' instruction takes a value to cast, which must be a
5936 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5937 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5938 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5939 type with the same number of elements as ``ty``
5944 The '``uitofp``' instruction interprets its operand as an unsigned
5945 integer quantity and converts it to the corresponding floating point
5946 value. If the value cannot fit in the floating point value, the results
5952 .. code-block:: llvm
5954 %X = uitofp i32 257 to float ; yields float:257.0
5955 %Y = uitofp i8 -1 to double ; yields double:255.0
5957 '``sitofp .. to``' Instruction
5958 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5965 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
5970 The '``sitofp``' instruction regards ``value`` as a signed integer and
5971 converts that value to the ``ty2`` type.
5976 The '``sitofp``' instruction takes a value to cast, which must be a
5977 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5978 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5979 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5980 type with the same number of elements as ``ty``
5985 The '``sitofp``' instruction interprets its operand as a signed integer
5986 quantity and converts it to the corresponding floating point value. If
5987 the value cannot fit in the floating point value, the results are
5993 .. code-block:: llvm
5995 %X = sitofp i32 257 to float ; yields float:257.0
5996 %Y = sitofp i8 -1 to double ; yields double:-1.0
6000 '``ptrtoint .. to``' Instruction
6001 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6008 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
6013 The '``ptrtoint``' instruction converts the pointer or a vector of
6014 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
6019 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
6020 a a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
6021 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
6022 a vector of integers type.
6027 The '``ptrtoint``' instruction converts ``value`` to integer type
6028 ``ty2`` by interpreting the pointer value as an integer and either
6029 truncating or zero extending that value to the size of the integer type.
6030 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
6031 ``value`` is larger than ``ty2`` then a truncation is done. If they are
6032 the same size, then nothing is done (*no-op cast*) other than a type
6038 .. code-block:: llvm
6040 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
6041 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
6042 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
6046 '``inttoptr .. to``' Instruction
6047 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6054 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
6059 The '``inttoptr``' instruction converts an integer ``value`` to a
6060 pointer type, ``ty2``.
6065 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
6066 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
6072 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
6073 applying either a zero extension or a truncation depending on the size
6074 of the integer ``value``. If ``value`` is larger than the size of a
6075 pointer then a truncation is done. If ``value`` is smaller than the size
6076 of a pointer then a zero extension is done. If they are the same size,
6077 nothing is done (*no-op cast*).
6082 .. code-block:: llvm
6084 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
6085 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
6086 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
6087 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
6091 '``bitcast .. to``' Instruction
6092 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6099 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
6104 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
6110 The '``bitcast``' instruction takes a value to cast, which must be a
6111 non-aggregate first class value, and a type to cast it to, which must
6112 also be a non-aggregate :ref:`first class <t_firstclass>` type. The
6113 bit sizes of ``value`` and the destination type, ``ty2``, must be
6114 identical. If the source type is a pointer, the destination type must
6115 also be a pointer of the same size. This instruction supports bitwise
6116 conversion of vectors to integers and to vectors of other types (as
6117 long as they have the same size).
6122 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It
6123 is always a *no-op cast* because no bits change with this
6124 conversion. The conversion is done as if the ``value`` had been stored
6125 to memory and read back as type ``ty2``. Pointer (or vector of
6126 pointers) types may only be converted to other pointer (or vector of
6127 pointers) types with the same address space through this instruction.
6128 To convert pointers to other types, use the :ref:`inttoptr <i_inttoptr>`
6129 or :ref:`ptrtoint <i_ptrtoint>` instructions first.
6134 .. code-block:: llvm
6136 %X = bitcast i8 255 to i8 ; yields i8 :-1
6137 %Y = bitcast i32* %x to sint* ; yields sint*:%x
6138 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
6139 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
6141 .. _i_addrspacecast:
6143 '``addrspacecast .. to``' Instruction
6144 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6151 <result> = addrspacecast <pty> <ptrval> to <pty2> ; yields pty2
6156 The '``addrspacecast``' instruction converts ``ptrval`` from ``pty`` in
6157 address space ``n`` to type ``pty2`` in address space ``m``.
6162 The '``addrspacecast``' instruction takes a pointer or vector of pointer value
6163 to cast and a pointer type to cast it to, which must have a different
6169 The '``addrspacecast``' instruction converts the pointer value
6170 ``ptrval`` to type ``pty2``. It can be a *no-op cast* or a complex
6171 value modification, depending on the target and the address space
6172 pair. Pointer conversions within the same address space must be
6173 performed with the ``bitcast`` instruction. Note that if the address space
6174 conversion is legal then both result and operand refer to the same memory
6180 .. code-block:: llvm
6182 %X = addrspacecast i32* %x to i32 addrspace(1)* ; yields i32 addrspace(1)*:%x
6183 %Y = addrspacecast i32 addrspace(1)* %y to i64 addrspace(2)* ; yields i64 addrspace(2)*:%y
6184 %Z = addrspacecast <4 x i32*> %z to <4 x float addrspace(3)*> ; yields <4 x float addrspace(3)*>:%z
6191 The instructions in this category are the "miscellaneous" instructions,
6192 which defy better classification.
6196 '``icmp``' Instruction
6197 ^^^^^^^^^^^^^^^^^^^^^^
6204 <result> = icmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6209 The '``icmp``' instruction returns a boolean value or a vector of
6210 boolean values based on comparison of its two integer, integer vector,
6211 pointer, or pointer vector operands.
6216 The '``icmp``' instruction takes three operands. The first operand is
6217 the condition code indicating the kind of comparison to perform. It is
6218 not a value, just a keyword. The possible condition code are:
6221 #. ``ne``: not equal
6222 #. ``ugt``: unsigned greater than
6223 #. ``uge``: unsigned greater or equal
6224 #. ``ult``: unsigned less than
6225 #. ``ule``: unsigned less or equal
6226 #. ``sgt``: signed greater than
6227 #. ``sge``: signed greater or equal
6228 #. ``slt``: signed less than
6229 #. ``sle``: signed less or equal
6231 The remaining two arguments must be :ref:`integer <t_integer>` or
6232 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
6233 must also be identical types.
6238 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
6239 code given as ``cond``. The comparison performed always yields either an
6240 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
6242 #. ``eq``: yields ``true`` if the operands are equal, ``false``
6243 otherwise. No sign interpretation is necessary or performed.
6244 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
6245 otherwise. No sign interpretation is necessary or performed.
6246 #. ``ugt``: interprets the operands as unsigned values and yields
6247 ``true`` if ``op1`` is greater than ``op2``.
6248 #. ``uge``: interprets the operands as unsigned values and yields
6249 ``true`` if ``op1`` is greater than or equal to ``op2``.
6250 #. ``ult``: interprets the operands as unsigned values and yields
6251 ``true`` if ``op1`` is less than ``op2``.
6252 #. ``ule``: interprets the operands as unsigned values and yields
6253 ``true`` if ``op1`` is less than or equal to ``op2``.
6254 #. ``sgt``: interprets the operands as signed values and yields ``true``
6255 if ``op1`` is greater than ``op2``.
6256 #. ``sge``: interprets the operands as signed values and yields ``true``
6257 if ``op1`` is greater than or equal to ``op2``.
6258 #. ``slt``: interprets the operands as signed values and yields ``true``
6259 if ``op1`` is less than ``op2``.
6260 #. ``sle``: interprets the operands as signed values and yields ``true``
6261 if ``op1`` is less than or equal to ``op2``.
6263 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
6264 are compared as if they were integers.
6266 If the operands are integer vectors, then they are compared element by
6267 element. The result is an ``i1`` vector with the same number of elements
6268 as the values being compared. Otherwise, the result is an ``i1``.
6273 .. code-block:: llvm
6275 <result> = icmp eq i32 4, 5 ; yields: result=false
6276 <result> = icmp ne float* %X, %X ; yields: result=false
6277 <result> = icmp ult i16 4, 5 ; yields: result=true
6278 <result> = icmp sgt i16 4, 5 ; yields: result=false
6279 <result> = icmp ule i16 -4, 5 ; yields: result=false
6280 <result> = icmp sge i16 4, 5 ; yields: result=false
6282 Note that the code generator does not yet support vector types with the
6283 ``icmp`` instruction.
6287 '``fcmp``' Instruction
6288 ^^^^^^^^^^^^^^^^^^^^^^
6295 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields i1 or <N x i1>:result
6300 The '``fcmp``' instruction returns a boolean value or vector of boolean
6301 values based on comparison of its operands.
6303 If the operands are floating point scalars, then the result type is a
6304 boolean (:ref:`i1 <t_integer>`).
6306 If the operands are floating point vectors, then the result type is a
6307 vector of boolean with the same number of elements as the operands being
6313 The '``fcmp``' instruction takes three operands. The first operand is
6314 the condition code indicating the kind of comparison to perform. It is
6315 not a value, just a keyword. The possible condition code are:
6317 #. ``false``: no comparison, always returns false
6318 #. ``oeq``: ordered and equal
6319 #. ``ogt``: ordered and greater than
6320 #. ``oge``: ordered and greater than or equal
6321 #. ``olt``: ordered and less than
6322 #. ``ole``: ordered and less than or equal
6323 #. ``one``: ordered and not equal
6324 #. ``ord``: ordered (no nans)
6325 #. ``ueq``: unordered or equal
6326 #. ``ugt``: unordered or greater than
6327 #. ``uge``: unordered or greater than or equal
6328 #. ``ult``: unordered or less than
6329 #. ``ule``: unordered or less than or equal
6330 #. ``une``: unordered or not equal
6331 #. ``uno``: unordered (either nans)
6332 #. ``true``: no comparison, always returns true
6334 *Ordered* means that neither operand is a QNAN while *unordered* means
6335 that either operand may be a QNAN.
6337 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
6338 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
6339 type. They must have identical types.
6344 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
6345 condition code given as ``cond``. If the operands are vectors, then the
6346 vectors are compared element by element. Each comparison performed
6347 always yields an :ref:`i1 <t_integer>` result, as follows:
6349 #. ``false``: always yields ``false``, regardless of operands.
6350 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
6351 is equal to ``op2``.
6352 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
6353 is greater than ``op2``.
6354 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
6355 is greater than or equal to ``op2``.
6356 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
6357 is less than ``op2``.
6358 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
6359 is less than or equal to ``op2``.
6360 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
6361 is not equal to ``op2``.
6362 #. ``ord``: yields ``true`` if both operands are not a QNAN.
6363 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
6365 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
6366 greater than ``op2``.
6367 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
6368 greater than or equal to ``op2``.
6369 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
6371 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
6372 less than or equal to ``op2``.
6373 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
6374 not equal to ``op2``.
6375 #. ``uno``: yields ``true`` if either operand is a QNAN.
6376 #. ``true``: always yields ``true``, regardless of operands.
6381 .. code-block:: llvm
6383 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
6384 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
6385 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
6386 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
6388 Note that the code generator does not yet support vector types with the
6389 ``fcmp`` instruction.
6393 '``phi``' Instruction
6394 ^^^^^^^^^^^^^^^^^^^^^
6401 <result> = phi <ty> [ <val0>, <label0>], ...
6406 The '``phi``' instruction is used to implement the φ node in the SSA
6407 graph representing the function.
6412 The type of the incoming values is specified with the first type field.
6413 After this, the '``phi``' instruction takes a list of pairs as
6414 arguments, with one pair for each predecessor basic block of the current
6415 block. Only values of :ref:`first class <t_firstclass>` type may be used as
6416 the value arguments to the PHI node. Only labels may be used as the
6419 There must be no non-phi instructions between the start of a basic block
6420 and the PHI instructions: i.e. PHI instructions must be first in a basic
6423 For the purposes of the SSA form, the use of each incoming value is
6424 deemed to occur on the edge from the corresponding predecessor block to
6425 the current block (but after any definition of an '``invoke``'
6426 instruction's return value on the same edge).
6431 At runtime, the '``phi``' instruction logically takes on the value
6432 specified by the pair corresponding to the predecessor basic block that
6433 executed just prior to the current block.
6438 .. code-block:: llvm
6440 Loop: ; Infinite loop that counts from 0 on up...
6441 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
6442 %nextindvar = add i32 %indvar, 1
6447 '``select``' Instruction
6448 ^^^^^^^^^^^^^^^^^^^^^^^^
6455 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
6457 selty is either i1 or {<N x i1>}
6462 The '``select``' instruction is used to choose one value based on a
6463 condition, without IR-level branching.
6468 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
6469 values indicating the condition, and two values of the same :ref:`first
6470 class <t_firstclass>` type. If the val1/val2 are vectors and the
6471 condition is a scalar, then entire vectors are selected, not individual
6477 If the condition is an i1 and it evaluates to 1, the instruction returns
6478 the first value argument; otherwise, it returns the second value
6481 If the condition is a vector of i1, then the value arguments must be
6482 vectors of the same size, and the selection is done element by element.
6487 .. code-block:: llvm
6489 %X = select i1 true, i8 17, i8 42 ; yields i8:17
6493 '``call``' Instruction
6494 ^^^^^^^^^^^^^^^^^^^^^^
6501 <result> = [tail | musttail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
6506 The '``call``' instruction represents a simple function call.
6511 This instruction requires several arguments:
6513 #. The optional ``tail`` and ``musttail`` markers indicate that the optimizers
6514 should perform tail call optimization. The ``tail`` marker is a hint that
6515 `can be ignored <CodeGenerator.html#sibcallopt>`_. The ``musttail`` marker
6516 means that the call must be tail call optimized in order for the program to
6517 be correct. The ``musttail`` marker provides these guarantees:
6519 #. The call will not cause unbounded stack growth if it is part of a
6520 recursive cycle in the call graph.
6521 #. Arguments with the :ref:`inalloca <attr_inalloca>` attribute are
6524 Both markers imply that the callee does not access allocas or varargs from
6525 the caller. Calls marked ``musttail`` must obey the following additional
6528 - The call must immediately precede a :ref:`ret <i_ret>` instruction,
6529 or a pointer bitcast followed by a ret instruction.
6530 - The ret instruction must return the (possibly bitcasted) value
6531 produced by the call or void.
6532 - The caller and callee prototypes must match. Pointer types of
6533 parameters or return types may differ in pointee type, but not
6535 - The calling conventions of the caller and callee must match.
6536 - All ABI-impacting function attributes, such as sret, byval, inreg,
6537 returned, and inalloca, must match.
6539 Tail call optimization for calls marked ``tail`` is guaranteed to occur if
6540 the following conditions are met:
6542 - Caller and callee both have the calling convention ``fastcc``.
6543 - The call is in tail position (ret immediately follows call and ret
6544 uses value of call or is void).
6545 - Option ``-tailcallopt`` is enabled, or
6546 ``llvm::GuaranteedTailCallOpt`` is ``true``.
6547 - `Platform-specific constraints are
6548 met. <CodeGenerator.html#tailcallopt>`_
6550 #. The optional "cconv" marker indicates which :ref:`calling
6551 convention <callingconv>` the call should use. If none is
6552 specified, the call defaults to using C calling conventions. The
6553 calling convention of the call must match the calling convention of
6554 the target function, or else the behavior is undefined.
6555 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
6556 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
6558 #. '``ty``': the type of the call instruction itself which is also the
6559 type of the return value. Functions that return no value are marked
6561 #. '``fnty``': shall be the signature of the pointer to function value
6562 being invoked. The argument types must match the types implied by
6563 this signature. This type can be omitted if the function is not
6564 varargs and if the function type does not return a pointer to a
6566 #. '``fnptrval``': An LLVM value containing a pointer to a function to
6567 be invoked. In most cases, this is a direct function invocation, but
6568 indirect ``call``'s are just as possible, calling an arbitrary pointer
6570 #. '``function args``': argument list whose types match the function
6571 signature argument types and parameter attributes. All arguments must
6572 be of :ref:`first class <t_firstclass>` type. If the function signature
6573 indicates the function accepts a variable number of arguments, the
6574 extra arguments can be specified.
6575 #. The optional :ref:`function attributes <fnattrs>` list. Only
6576 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
6577 attributes are valid here.
6582 The '``call``' instruction is used to cause control flow to transfer to
6583 a specified function, with its incoming arguments bound to the specified
6584 values. Upon a '``ret``' instruction in the called function, control
6585 flow continues with the instruction after the function call, and the
6586 return value of the function is bound to the result argument.
6591 .. code-block:: llvm
6593 %retval = call i32 @test(i32 %argc)
6594 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
6595 %X = tail call i32 @foo() ; yields i32
6596 %Y = tail call fastcc i32 @foo() ; yields i32
6597 call void %foo(i8 97 signext)
6599 %struct.A = type { i32, i8 }
6600 %r = call %struct.A @foo() ; yields { i32, i8 }
6601 %gr = extractvalue %struct.A %r, 0 ; yields i32
6602 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
6603 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
6604 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
6606 llvm treats calls to some functions with names and arguments that match
6607 the standard C99 library as being the C99 library functions, and may
6608 perform optimizations or generate code for them under that assumption.
6609 This is something we'd like to change in the future to provide better
6610 support for freestanding environments and non-C-based languages.
6614 '``va_arg``' Instruction
6615 ^^^^^^^^^^^^^^^^^^^^^^^^
6622 <resultval> = va_arg <va_list*> <arglist>, <argty>
6627 The '``va_arg``' instruction is used to access arguments passed through
6628 the "variable argument" area of a function call. It is used to implement
6629 the ``va_arg`` macro in C.
6634 This instruction takes a ``va_list*`` value and the type of the
6635 argument. It returns a value of the specified argument type and
6636 increments the ``va_list`` to point to the next argument. The actual
6637 type of ``va_list`` is target specific.
6642 The '``va_arg``' instruction loads an argument of the specified type
6643 from the specified ``va_list`` and causes the ``va_list`` to point to
6644 the next argument. For more information, see the variable argument
6645 handling :ref:`Intrinsic Functions <int_varargs>`.
6647 It is legal for this instruction to be called in a function which does
6648 not take a variable number of arguments, for example, the ``vfprintf``
6651 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
6652 function <intrinsics>` because it takes a type as an argument.
6657 See the :ref:`variable argument processing <int_varargs>` section.
6659 Note that the code generator does not yet fully support va\_arg on many
6660 targets. Also, it does not currently support va\_arg with aggregate
6661 types on any target.
6665 '``landingpad``' Instruction
6666 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6673 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
6674 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
6676 <clause> := catch <type> <value>
6677 <clause> := filter <array constant type> <array constant>
6682 The '``landingpad``' instruction is used by `LLVM's exception handling
6683 system <ExceptionHandling.html#overview>`_ to specify that a basic block
6684 is a landing pad --- one where the exception lands, and corresponds to the
6685 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
6686 defines values supplied by the personality function (``pers_fn``) upon
6687 re-entry to the function. The ``resultval`` has the type ``resultty``.
6692 This instruction takes a ``pers_fn`` value. This is the personality
6693 function associated with the unwinding mechanism. The optional
6694 ``cleanup`` flag indicates that the landing pad block is a cleanup.
6696 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
6697 contains the global variable representing the "type" that may be caught
6698 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
6699 clause takes an array constant as its argument. Use
6700 "``[0 x i8**] undef``" for a filter which cannot throw. The
6701 '``landingpad``' instruction must contain *at least* one ``clause`` or
6702 the ``cleanup`` flag.
6707 The '``landingpad``' instruction defines the values which are set by the
6708 personality function (``pers_fn``) upon re-entry to the function, and
6709 therefore the "result type" of the ``landingpad`` instruction. As with
6710 calling conventions, how the personality function results are
6711 represented in LLVM IR is target specific.
6713 The clauses are applied in order from top to bottom. If two
6714 ``landingpad`` instructions are merged together through inlining, the
6715 clauses from the calling function are appended to the list of clauses.
6716 When the call stack is being unwound due to an exception being thrown,
6717 the exception is compared against each ``clause`` in turn. If it doesn't
6718 match any of the clauses, and the ``cleanup`` flag is not set, then
6719 unwinding continues further up the call stack.
6721 The ``landingpad`` instruction has several restrictions:
6723 - A landing pad block is a basic block which is the unwind destination
6724 of an '``invoke``' instruction.
6725 - A landing pad block must have a '``landingpad``' instruction as its
6726 first non-PHI instruction.
6727 - There can be only one '``landingpad``' instruction within the landing
6729 - A basic block that is not a landing pad block may not include a
6730 '``landingpad``' instruction.
6731 - All '``landingpad``' instructions in a function must have the same
6732 personality function.
6737 .. code-block:: llvm
6739 ;; A landing pad which can catch an integer.
6740 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6742 ;; A landing pad that is a cleanup.
6743 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6745 ;; A landing pad which can catch an integer and can only throw a double.
6746 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6748 filter [1 x i8**] [@_ZTId]
6755 LLVM supports the notion of an "intrinsic function". These functions
6756 have well known names and semantics and are required to follow certain
6757 restrictions. Overall, these intrinsics represent an extension mechanism
6758 for the LLVM language that does not require changing all of the
6759 transformations in LLVM when adding to the language (or the bitcode
6760 reader/writer, the parser, etc...).
6762 Intrinsic function names must all start with an "``llvm.``" prefix. This
6763 prefix is reserved in LLVM for intrinsic names; thus, function names may
6764 not begin with this prefix. Intrinsic functions must always be external
6765 functions: you cannot define the body of intrinsic functions. Intrinsic
6766 functions may only be used in call or invoke instructions: it is illegal
6767 to take the address of an intrinsic function. Additionally, because
6768 intrinsic functions are part of the LLVM language, it is required if any
6769 are added that they be documented here.
6771 Some intrinsic functions can be overloaded, i.e., the intrinsic
6772 represents a family of functions that perform the same operation but on
6773 different data types. Because LLVM can represent over 8 million
6774 different integer types, overloading is used commonly to allow an
6775 intrinsic function to operate on any integer type. One or more of the
6776 argument types or the result type can be overloaded to accept any
6777 integer type. Argument types may also be defined as exactly matching a
6778 previous argument's type or the result type. This allows an intrinsic
6779 function which accepts multiple arguments, but needs all of them to be
6780 of the same type, to only be overloaded with respect to a single
6781 argument or the result.
6783 Overloaded intrinsics will have the names of its overloaded argument
6784 types encoded into its function name, each preceded by a period. Only
6785 those types which are overloaded result in a name suffix. Arguments
6786 whose type is matched against another type do not. For example, the
6787 ``llvm.ctpop`` function can take an integer of any width and returns an
6788 integer of exactly the same integer width. This leads to a family of
6789 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
6790 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
6791 overloaded, and only one type suffix is required. Because the argument's
6792 type is matched against the return type, it does not require its own
6795 To learn how to add an intrinsic function, please see the `Extending
6796 LLVM Guide <ExtendingLLVM.html>`_.
6800 Variable Argument Handling Intrinsics
6801 -------------------------------------
6803 Variable argument support is defined in LLVM with the
6804 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
6805 functions. These functions are related to the similarly named macros
6806 defined in the ``<stdarg.h>`` header file.
6808 All of these functions operate on arguments that use a target-specific
6809 value type "``va_list``". The LLVM assembly language reference manual
6810 does not define what this type is, so all transformations should be
6811 prepared to handle these functions regardless of the type used.
6813 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
6814 variable argument handling intrinsic functions are used.
6816 .. code-block:: llvm
6818 define i32 @test(i32 %X, ...) {
6819 ; Initialize variable argument processing
6821 %ap2 = bitcast i8** %ap to i8*
6822 call void @llvm.va_start(i8* %ap2)
6824 ; Read a single integer argument
6825 %tmp = va_arg i8** %ap, i32
6827 ; Demonstrate usage of llvm.va_copy and llvm.va_end
6829 %aq2 = bitcast i8** %aq to i8*
6830 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
6831 call void @llvm.va_end(i8* %aq2)
6833 ; Stop processing of arguments.
6834 call void @llvm.va_end(i8* %ap2)
6838 declare void @llvm.va_start(i8*)
6839 declare void @llvm.va_copy(i8*, i8*)
6840 declare void @llvm.va_end(i8*)
6844 '``llvm.va_start``' Intrinsic
6845 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6852 declare void @llvm.va_start(i8* <arglist>)
6857 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
6858 subsequent use by ``va_arg``.
6863 The argument is a pointer to a ``va_list`` element to initialize.
6868 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
6869 available in C. In a target-dependent way, it initializes the
6870 ``va_list`` element to which the argument points, so that the next call
6871 to ``va_arg`` will produce the first variable argument passed to the
6872 function. Unlike the C ``va_start`` macro, this intrinsic does not need
6873 to know the last argument of the function as the compiler can figure
6876 '``llvm.va_end``' Intrinsic
6877 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6884 declare void @llvm.va_end(i8* <arglist>)
6889 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
6890 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
6895 The argument is a pointer to a ``va_list`` to destroy.
6900 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
6901 available in C. In a target-dependent way, it destroys the ``va_list``
6902 element to which the argument points. Calls to
6903 :ref:`llvm.va_start <int_va_start>` and
6904 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
6909 '``llvm.va_copy``' Intrinsic
6910 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6917 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
6922 The '``llvm.va_copy``' intrinsic copies the current argument position
6923 from the source argument list to the destination argument list.
6928 The first argument is a pointer to a ``va_list`` element to initialize.
6929 The second argument is a pointer to a ``va_list`` element to copy from.
6934 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
6935 available in C. In a target-dependent way, it copies the source
6936 ``va_list`` element into the destination ``va_list`` element. This
6937 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
6938 arbitrarily complex and require, for example, memory allocation.
6940 Accurate Garbage Collection Intrinsics
6941 --------------------------------------
6943 LLVM support for `Accurate Garbage Collection <GarbageCollection.html>`_
6944 (GC) requires the implementation and generation of these intrinsics.
6945 These intrinsics allow identification of :ref:`GC roots on the
6946 stack <int_gcroot>`, as well as garbage collector implementations that
6947 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
6948 Front-ends for type-safe garbage collected languages should generate
6949 these intrinsics to make use of the LLVM garbage collectors. For more
6950 details, see `Accurate Garbage Collection with
6951 LLVM <GarbageCollection.html>`_.
6953 The garbage collection intrinsics only operate on objects in the generic
6954 address space (address space zero).
6958 '``llvm.gcroot``' Intrinsic
6959 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6966 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
6971 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
6972 the code generator, and allows some metadata to be associated with it.
6977 The first argument specifies the address of a stack object that contains
6978 the root pointer. The second pointer (which must be either a constant or
6979 a global value address) contains the meta-data to be associated with the
6985 At runtime, a call to this intrinsic stores a null pointer into the
6986 "ptrloc" location. At compile-time, the code generator generates
6987 information to allow the runtime to find the pointer at GC safe points.
6988 The '``llvm.gcroot``' intrinsic may only be used in a function which
6989 :ref:`specifies a GC algorithm <gc>`.
6993 '``llvm.gcread``' Intrinsic
6994 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7001 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
7006 The '``llvm.gcread``' intrinsic identifies reads of references from heap
7007 locations, allowing garbage collector implementations that require read
7013 The second argument is the address to read from, which should be an
7014 address allocated from the garbage collector. The first object is a
7015 pointer to the start of the referenced object, if needed by the language
7016 runtime (otherwise null).
7021 The '``llvm.gcread``' intrinsic has the same semantics as a load
7022 instruction, but may be replaced with substantially more complex code by
7023 the garbage collector runtime, as needed. The '``llvm.gcread``'
7024 intrinsic may only be used in a function which :ref:`specifies a GC
7029 '``llvm.gcwrite``' Intrinsic
7030 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7037 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
7042 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
7043 locations, allowing garbage collector implementations that require write
7044 barriers (such as generational or reference counting collectors).
7049 The first argument is the reference to store, the second is the start of
7050 the object to store it to, and the third is the address of the field of
7051 Obj to store to. If the runtime does not require a pointer to the
7052 object, Obj may be null.
7057 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
7058 instruction, but may be replaced with substantially more complex code by
7059 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
7060 intrinsic may only be used in a function which :ref:`specifies a GC
7063 Code Generator Intrinsics
7064 -------------------------
7066 These intrinsics are provided by LLVM to expose special features that
7067 may only be implemented with code generator support.
7069 '``llvm.returnaddress``' Intrinsic
7070 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7077 declare i8 *@llvm.returnaddress(i32 <level>)
7082 The '``llvm.returnaddress``' intrinsic attempts to compute a
7083 target-specific value indicating the return address of the current
7084 function or one of its callers.
7089 The argument to this intrinsic indicates which function to return the
7090 address for. Zero indicates the calling function, one indicates its
7091 caller, etc. The argument is **required** to be a constant integer
7097 The '``llvm.returnaddress``' intrinsic either returns a pointer
7098 indicating the return address of the specified call frame, or zero if it
7099 cannot be identified. The value returned by this intrinsic is likely to
7100 be incorrect or 0 for arguments other than zero, so it should only be
7101 used for debugging purposes.
7103 Note that calling this intrinsic does not prevent function inlining or
7104 other aggressive transformations, so the value returned may not be that
7105 of the obvious source-language caller.
7107 '``llvm.frameaddress``' Intrinsic
7108 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7115 declare i8* @llvm.frameaddress(i32 <level>)
7120 The '``llvm.frameaddress``' intrinsic attempts to return the
7121 target-specific frame pointer value for the specified stack frame.
7126 The argument to this intrinsic indicates which function to return the
7127 frame pointer for. Zero indicates the calling function, one indicates
7128 its caller, etc. The argument is **required** to be a constant integer
7134 The '``llvm.frameaddress``' intrinsic either returns a pointer
7135 indicating the frame address of the specified call frame, or zero if it
7136 cannot be identified. The value returned by this intrinsic is likely to
7137 be incorrect or 0 for arguments other than zero, so it should only be
7138 used for debugging purposes.
7140 Note that calling this intrinsic does not prevent function inlining or
7141 other aggressive transformations, so the value returned may not be that
7142 of the obvious source-language caller.
7144 .. _int_read_register:
7145 .. _int_write_register:
7147 '``llvm.read_register``' and '``llvm.write_register``' Intrinsics
7148 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7155 declare i32 @llvm.read_register.i32(metadata)
7156 declare i64 @llvm.read_register.i64(metadata)
7157 declare void @llvm.write_register.i32(metadata, i32 @value)
7158 declare void @llvm.write_register.i64(metadata, i64 @value)
7159 !0 = metadata !{metadata !"sp\00"}
7164 The '``llvm.read_register``' and '``llvm.write_register``' intrinsics
7165 provides access to the named register. The register must be valid on
7166 the architecture being compiled to. The type needs to be compatible
7167 with the register being read.
7172 The '``llvm.read_register``' intrinsic returns the current value of the
7173 register, where possible. The '``llvm.write_register``' intrinsic sets
7174 the current value of the register, where possible.
7176 This is useful to implement named register global variables that need
7177 to always be mapped to a specific register, as is common practice on
7178 bare-metal programs including OS kernels.
7180 The compiler doesn't check for register availability or use of the used
7181 register in surrounding code, including inline assembly. Because of that,
7182 allocatable registers are not supported.
7184 Warning: So far it only works with the stack pointer on selected
7185 architectures (ARM, AArch64, PowerPC and x86_64). Significant amount of
7186 work is needed to support other registers and even more so, allocatable
7191 '``llvm.stacksave``' Intrinsic
7192 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7199 declare i8* @llvm.stacksave()
7204 The '``llvm.stacksave``' intrinsic is used to remember the current state
7205 of the function stack, for use with
7206 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
7207 implementing language features like scoped automatic variable sized
7213 This intrinsic returns a opaque pointer value that can be passed to
7214 :ref:`llvm.stackrestore <int_stackrestore>`. When an
7215 ``llvm.stackrestore`` intrinsic is executed with a value saved from
7216 ``llvm.stacksave``, it effectively restores the state of the stack to
7217 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
7218 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
7219 were allocated after the ``llvm.stacksave`` was executed.
7221 .. _int_stackrestore:
7223 '``llvm.stackrestore``' Intrinsic
7224 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7231 declare void @llvm.stackrestore(i8* %ptr)
7236 The '``llvm.stackrestore``' intrinsic is used to restore the state of
7237 the function stack to the state it was in when the corresponding
7238 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
7239 useful for implementing language features like scoped automatic variable
7240 sized arrays in C99.
7245 See the description for :ref:`llvm.stacksave <int_stacksave>`.
7247 '``llvm.prefetch``' Intrinsic
7248 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7255 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
7260 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
7261 insert a prefetch instruction if supported; otherwise, it is a noop.
7262 Prefetches have no effect on the behavior of the program but can change
7263 its performance characteristics.
7268 ``address`` is the address to be prefetched, ``rw`` is the specifier
7269 determining if the fetch should be for a read (0) or write (1), and
7270 ``locality`` is a temporal locality specifier ranging from (0) - no
7271 locality, to (3) - extremely local keep in cache. The ``cache type``
7272 specifies whether the prefetch is performed on the data (1) or
7273 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
7274 arguments must be constant integers.
7279 This intrinsic does not modify the behavior of the program. In
7280 particular, prefetches cannot trap and do not produce a value. On
7281 targets that support this intrinsic, the prefetch can provide hints to
7282 the processor cache for better performance.
7284 '``llvm.pcmarker``' Intrinsic
7285 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7292 declare void @llvm.pcmarker(i32 <id>)
7297 The '``llvm.pcmarker``' intrinsic is a method to export a Program
7298 Counter (PC) in a region of code to simulators and other tools. The
7299 method is target specific, but it is expected that the marker will use
7300 exported symbols to transmit the PC of the marker. The marker makes no
7301 guarantees that it will remain with any specific instruction after
7302 optimizations. It is possible that the presence of a marker will inhibit
7303 optimizations. The intended use is to be inserted after optimizations to
7304 allow correlations of simulation runs.
7309 ``id`` is a numerical id identifying the marker.
7314 This intrinsic does not modify the behavior of the program. Backends
7315 that do not support this intrinsic may ignore it.
7317 '``llvm.readcyclecounter``' Intrinsic
7318 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7325 declare i64 @llvm.readcyclecounter()
7330 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
7331 counter register (or similar low latency, high accuracy clocks) on those
7332 targets that support it. On X86, it should map to RDTSC. On Alpha, it
7333 should map to RPCC. As the backing counters overflow quickly (on the
7334 order of 9 seconds on alpha), this should only be used for small
7340 When directly supported, reading the cycle counter should not modify any
7341 memory. Implementations are allowed to either return a application
7342 specific value or a system wide value. On backends without support, this
7343 is lowered to a constant 0.
7345 Note that runtime support may be conditional on the privilege-level code is
7346 running at and the host platform.
7348 '``llvm.clear_cache``' Intrinsic
7349 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7356 declare void @llvm.clear_cache(i8*, i8*)
7361 The '``llvm.clear_cache``' intrinsic ensures visibility of modifications
7362 in the specified range to the execution unit of the processor. On
7363 targets with non-unified instruction and data cache, the implementation
7364 flushes the instruction cache.
7369 On platforms with coherent instruction and data caches (e.g. x86), this
7370 intrinsic is a nop. On platforms with non-coherent instruction and data
7371 cache (e.g. ARM, MIPS), the intrinsic is lowered either to appropriate
7372 instructions or a system call, if cache flushing requires special
7375 The default behavior is to emit a call to ``__clear_cache`` from the run
7378 This instrinsic does *not* empty the instruction pipeline. Modifications
7379 of the current function are outside the scope of the intrinsic.
7381 Standard C Library Intrinsics
7382 -----------------------------
7384 LLVM provides intrinsics for a few important standard C library
7385 functions. These intrinsics allow source-language front-ends to pass
7386 information about the alignment of the pointer arguments to the code
7387 generator, providing opportunity for more efficient code generation.
7391 '``llvm.memcpy``' Intrinsic
7392 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7397 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
7398 integer bit width and for different address spaces. Not all targets
7399 support all bit widths however.
7403 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7404 i32 <len>, i32 <align>, i1 <isvolatile>)
7405 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7406 i64 <len>, i32 <align>, i1 <isvolatile>)
7411 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7412 source location to the destination location.
7414 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
7415 intrinsics do not return a value, takes extra alignment/isvolatile
7416 arguments and the pointers can be in specified address spaces.
7421 The first argument is a pointer to the destination, the second is a
7422 pointer to the source. The third argument is an integer argument
7423 specifying the number of bytes to copy, the fourth argument is the
7424 alignment of the source and destination locations, and the fifth is a
7425 boolean indicating a volatile access.
7427 If the call to this intrinsic has an alignment value that is not 0 or 1,
7428 then the caller guarantees that both the source and destination pointers
7429 are aligned to that boundary.
7431 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
7432 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7433 very cleanly specified and it is unwise to depend on it.
7438 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
7439 source location to the destination location, which are not allowed to
7440 overlap. It copies "len" bytes of memory over. If the argument is known
7441 to be aligned to some boundary, this can be specified as the fourth
7442 argument, otherwise it should be set to 0 or 1 (both meaning no alignment).
7444 '``llvm.memmove``' Intrinsic
7445 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7450 This is an overloaded intrinsic. You can use llvm.memmove on any integer
7451 bit width and for different address space. Not all targets support all
7456 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
7457 i32 <len>, i32 <align>, i1 <isvolatile>)
7458 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
7459 i64 <len>, i32 <align>, i1 <isvolatile>)
7464 The '``llvm.memmove.*``' intrinsics move a block of memory from the
7465 source location to the destination location. It is similar to the
7466 '``llvm.memcpy``' intrinsic but allows the two memory locations to
7469 Note that, unlike the standard libc function, the ``llvm.memmove.*``
7470 intrinsics do not return a value, takes extra alignment/isvolatile
7471 arguments and the pointers can be in specified address spaces.
7476 The first argument is a pointer to the destination, the second is a
7477 pointer to the source. The third argument is an integer argument
7478 specifying the number of bytes to copy, the fourth argument is the
7479 alignment of the source and destination locations, and the fifth is a
7480 boolean indicating a volatile access.
7482 If the call to this intrinsic has an alignment value that is not 0 or 1,
7483 then the caller guarantees that the source and destination pointers are
7484 aligned to that boundary.
7486 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
7487 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
7488 not very cleanly specified and it is unwise to depend on it.
7493 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
7494 source location to the destination location, which may overlap. It
7495 copies "len" bytes of memory over. If the argument is known to be
7496 aligned to some boundary, this can be specified as the fourth argument,
7497 otherwise it should be set to 0 or 1 (both meaning no alignment).
7499 '``llvm.memset.*``' Intrinsics
7500 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7505 This is an overloaded intrinsic. You can use llvm.memset on any integer
7506 bit width and for different address spaces. However, not all targets
7507 support all bit widths.
7511 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
7512 i32 <len>, i32 <align>, i1 <isvolatile>)
7513 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
7514 i64 <len>, i32 <align>, i1 <isvolatile>)
7519 The '``llvm.memset.*``' intrinsics fill a block of memory with a
7520 particular byte value.
7522 Note that, unlike the standard libc function, the ``llvm.memset``
7523 intrinsic does not return a value and takes extra alignment/volatile
7524 arguments. Also, the destination can be in an arbitrary address space.
7529 The first argument is a pointer to the destination to fill, the second
7530 is the byte value with which to fill it, the third argument is an
7531 integer argument specifying the number of bytes to fill, and the fourth
7532 argument is the known alignment of the destination location.
7534 If the call to this intrinsic has an alignment value that is not 0 or 1,
7535 then the caller guarantees that the destination pointer is aligned to
7538 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
7539 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
7540 very cleanly specified and it is unwise to depend on it.
7545 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
7546 at the destination location. If the argument is known to be aligned to
7547 some boundary, this can be specified as the fourth argument, otherwise
7548 it should be set to 0 or 1 (both meaning no alignment).
7550 '``llvm.sqrt.*``' Intrinsic
7551 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7556 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
7557 floating point or vector of floating point type. Not all targets support
7562 declare float @llvm.sqrt.f32(float %Val)
7563 declare double @llvm.sqrt.f64(double %Val)
7564 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
7565 declare fp128 @llvm.sqrt.f128(fp128 %Val)
7566 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
7571 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
7572 returning the same value as the libm '``sqrt``' functions would. Unlike
7573 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
7574 negative numbers other than -0.0 (which allows for better optimization,
7575 because there is no need to worry about errno being set).
7576 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
7581 The argument and return value are floating point numbers of the same
7587 This function returns the sqrt of the specified operand if it is a
7588 nonnegative floating point number.
7590 '``llvm.powi.*``' Intrinsic
7591 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7596 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
7597 floating point or vector of floating point type. Not all targets support
7602 declare float @llvm.powi.f32(float %Val, i32 %power)
7603 declare double @llvm.powi.f64(double %Val, i32 %power)
7604 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7605 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7606 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
7611 The '``llvm.powi.*``' intrinsics return the first operand raised to the
7612 specified (positive or negative) power. The order of evaluation of
7613 multiplications is not defined. When a vector of floating point type is
7614 used, the second argument remains a scalar integer value.
7619 The second argument is an integer power, and the first is a value to
7620 raise to that power.
7625 This function returns the first value raised to the second power with an
7626 unspecified sequence of rounding operations.
7628 '``llvm.sin.*``' Intrinsic
7629 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7634 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
7635 floating point or vector of floating point type. Not all targets support
7640 declare float @llvm.sin.f32(float %Val)
7641 declare double @llvm.sin.f64(double %Val)
7642 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7643 declare fp128 @llvm.sin.f128(fp128 %Val)
7644 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7649 The '``llvm.sin.*``' intrinsics return the sine of the operand.
7654 The argument and return value are floating point numbers of the same
7660 This function returns the sine of the specified operand, returning the
7661 same values as the libm ``sin`` functions would, and handles error
7662 conditions in the same way.
7664 '``llvm.cos.*``' Intrinsic
7665 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7670 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
7671 floating point or vector of floating point type. Not all targets support
7676 declare float @llvm.cos.f32(float %Val)
7677 declare double @llvm.cos.f64(double %Val)
7678 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7679 declare fp128 @llvm.cos.f128(fp128 %Val)
7680 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7685 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
7690 The argument and return value are floating point numbers of the same
7696 This function returns the cosine of the specified operand, returning the
7697 same values as the libm ``cos`` functions would, and handles error
7698 conditions in the same way.
7700 '``llvm.pow.*``' Intrinsic
7701 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7706 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
7707 floating point or vector of floating point type. Not all targets support
7712 declare float @llvm.pow.f32(float %Val, float %Power)
7713 declare double @llvm.pow.f64(double %Val, double %Power)
7714 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7715 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7716 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7721 The '``llvm.pow.*``' intrinsics return the first operand raised to the
7722 specified (positive or negative) power.
7727 The second argument is a floating point power, and the first is a value
7728 to raise to that power.
7733 This function returns the first value raised to the second power,
7734 returning the same values as the libm ``pow`` functions would, and
7735 handles error conditions in the same way.
7737 '``llvm.exp.*``' Intrinsic
7738 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7743 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
7744 floating point or vector of floating point type. Not all targets support
7749 declare float @llvm.exp.f32(float %Val)
7750 declare double @llvm.exp.f64(double %Val)
7751 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7752 declare fp128 @llvm.exp.f128(fp128 %Val)
7753 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7758 The '``llvm.exp.*``' intrinsics perform the exp function.
7763 The argument and return value are floating point numbers of the same
7769 This function returns the same values as the libm ``exp`` functions
7770 would, and handles error conditions in the same way.
7772 '``llvm.exp2.*``' Intrinsic
7773 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7778 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
7779 floating point or vector of floating point type. Not all targets support
7784 declare float @llvm.exp2.f32(float %Val)
7785 declare double @llvm.exp2.f64(double %Val)
7786 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
7787 declare fp128 @llvm.exp2.f128(fp128 %Val)
7788 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
7793 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
7798 The argument and return value are floating point numbers of the same
7804 This function returns the same values as the libm ``exp2`` functions
7805 would, and handles error conditions in the same way.
7807 '``llvm.log.*``' Intrinsic
7808 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7813 This is an overloaded intrinsic. You can use ``llvm.log`` on any
7814 floating point or vector of floating point type. Not all targets support
7819 declare float @llvm.log.f32(float %Val)
7820 declare double @llvm.log.f64(double %Val)
7821 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7822 declare fp128 @llvm.log.f128(fp128 %Val)
7823 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7828 The '``llvm.log.*``' intrinsics perform the log function.
7833 The argument and return value are floating point numbers of the same
7839 This function returns the same values as the libm ``log`` functions
7840 would, and handles error conditions in the same way.
7842 '``llvm.log10.*``' Intrinsic
7843 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7848 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
7849 floating point or vector of floating point type. Not all targets support
7854 declare float @llvm.log10.f32(float %Val)
7855 declare double @llvm.log10.f64(double %Val)
7856 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
7857 declare fp128 @llvm.log10.f128(fp128 %Val)
7858 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
7863 The '``llvm.log10.*``' intrinsics perform the log10 function.
7868 The argument and return value are floating point numbers of the same
7874 This function returns the same values as the libm ``log10`` functions
7875 would, and handles error conditions in the same way.
7877 '``llvm.log2.*``' Intrinsic
7878 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7883 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
7884 floating point or vector of floating point type. Not all targets support
7889 declare float @llvm.log2.f32(float %Val)
7890 declare double @llvm.log2.f64(double %Val)
7891 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
7892 declare fp128 @llvm.log2.f128(fp128 %Val)
7893 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
7898 The '``llvm.log2.*``' intrinsics perform the log2 function.
7903 The argument and return value are floating point numbers of the same
7909 This function returns the same values as the libm ``log2`` functions
7910 would, and handles error conditions in the same way.
7912 '``llvm.fma.*``' Intrinsic
7913 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7918 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
7919 floating point or vector of floating point type. Not all targets support
7924 declare float @llvm.fma.f32(float %a, float %b, float %c)
7925 declare double @llvm.fma.f64(double %a, double %b, double %c)
7926 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7927 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7928 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7933 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
7939 The argument and return value are floating point numbers of the same
7945 This function returns the same values as the libm ``fma`` functions
7946 would, and does not set errno.
7948 '``llvm.fabs.*``' Intrinsic
7949 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7954 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
7955 floating point or vector of floating point type. Not all targets support
7960 declare float @llvm.fabs.f32(float %Val)
7961 declare double @llvm.fabs.f64(double %Val)
7962 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
7963 declare fp128 @llvm.fabs.f128(fp128 %Val)
7964 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
7969 The '``llvm.fabs.*``' intrinsics return the absolute value of the
7975 The argument and return value are floating point numbers of the same
7981 This function returns the same values as the libm ``fabs`` functions
7982 would, and handles error conditions in the same way.
7984 '``llvm.copysign.*``' Intrinsic
7985 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7990 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
7991 floating point or vector of floating point type. Not all targets support
7996 declare float @llvm.copysign.f32(float %Mag, float %Sgn)
7997 declare double @llvm.copysign.f64(double %Mag, double %Sgn)
7998 declare x86_fp80 @llvm.copysign.f80(x86_fp80 %Mag, x86_fp80 %Sgn)
7999 declare fp128 @llvm.copysign.f128(fp128 %Mag, fp128 %Sgn)
8000 declare ppc_fp128 @llvm.copysign.ppcf128(ppc_fp128 %Mag, ppc_fp128 %Sgn)
8005 The '``llvm.copysign.*``' intrinsics return a value with the magnitude of the
8006 first operand and the sign of the second operand.
8011 The arguments and return value are floating point numbers of the same
8017 This function returns the same values as the libm ``copysign``
8018 functions would, and handles error conditions in the same way.
8020 '``llvm.floor.*``' Intrinsic
8021 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8026 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
8027 floating point or vector of floating point type. Not all targets support
8032 declare float @llvm.floor.f32(float %Val)
8033 declare double @llvm.floor.f64(double %Val)
8034 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
8035 declare fp128 @llvm.floor.f128(fp128 %Val)
8036 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
8041 The '``llvm.floor.*``' intrinsics return the floor of the operand.
8046 The argument and return value are floating point numbers of the same
8052 This function returns the same values as the libm ``floor`` functions
8053 would, and handles error conditions in the same way.
8055 '``llvm.ceil.*``' Intrinsic
8056 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8061 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
8062 floating point or vector of floating point type. Not all targets support
8067 declare float @llvm.ceil.f32(float %Val)
8068 declare double @llvm.ceil.f64(double %Val)
8069 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
8070 declare fp128 @llvm.ceil.f128(fp128 %Val)
8071 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
8076 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
8081 The argument and return value are floating point numbers of the same
8087 This function returns the same values as the libm ``ceil`` functions
8088 would, and handles error conditions in the same way.
8090 '``llvm.trunc.*``' Intrinsic
8091 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8096 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
8097 floating point or vector of floating point type. Not all targets support
8102 declare float @llvm.trunc.f32(float %Val)
8103 declare double @llvm.trunc.f64(double %Val)
8104 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
8105 declare fp128 @llvm.trunc.f128(fp128 %Val)
8106 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
8111 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
8112 nearest integer not larger in magnitude than the operand.
8117 The argument and return value are floating point numbers of the same
8123 This function returns the same values as the libm ``trunc`` functions
8124 would, and handles error conditions in the same way.
8126 '``llvm.rint.*``' Intrinsic
8127 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8132 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
8133 floating point or vector of floating point type. Not all targets support
8138 declare float @llvm.rint.f32(float %Val)
8139 declare double @llvm.rint.f64(double %Val)
8140 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
8141 declare fp128 @llvm.rint.f128(fp128 %Val)
8142 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
8147 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
8148 nearest integer. It may raise an inexact floating-point exception if the
8149 operand isn't an integer.
8154 The argument and return value are floating point numbers of the same
8160 This function returns the same values as the libm ``rint`` functions
8161 would, and handles error conditions in the same way.
8163 '``llvm.nearbyint.*``' Intrinsic
8164 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8169 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
8170 floating point or vector of floating point type. Not all targets support
8175 declare float @llvm.nearbyint.f32(float %Val)
8176 declare double @llvm.nearbyint.f64(double %Val)
8177 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
8178 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
8179 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
8184 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
8190 The argument and return value are floating point numbers of the same
8196 This function returns the same values as the libm ``nearbyint``
8197 functions would, and handles error conditions in the same way.
8199 '``llvm.round.*``' Intrinsic
8200 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8205 This is an overloaded intrinsic. You can use ``llvm.round`` on any
8206 floating point or vector of floating point type. Not all targets support
8211 declare float @llvm.round.f32(float %Val)
8212 declare double @llvm.round.f64(double %Val)
8213 declare x86_fp80 @llvm.round.f80(x86_fp80 %Val)
8214 declare fp128 @llvm.round.f128(fp128 %Val)
8215 declare ppc_fp128 @llvm.round.ppcf128(ppc_fp128 %Val)
8220 The '``llvm.round.*``' intrinsics returns the operand rounded to the
8226 The argument and return value are floating point numbers of the same
8232 This function returns the same values as the libm ``round``
8233 functions would, and handles error conditions in the same way.
8235 Bit Manipulation Intrinsics
8236 ---------------------------
8238 LLVM provides intrinsics for a few important bit manipulation
8239 operations. These allow efficient code generation for some algorithms.
8241 '``llvm.bswap.*``' Intrinsics
8242 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8247 This is an overloaded intrinsic function. You can use bswap on any
8248 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
8252 declare i16 @llvm.bswap.i16(i16 <id>)
8253 declare i32 @llvm.bswap.i32(i32 <id>)
8254 declare i64 @llvm.bswap.i64(i64 <id>)
8259 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
8260 values with an even number of bytes (positive multiple of 16 bits).
8261 These are useful for performing operations on data that is not in the
8262 target's native byte order.
8267 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
8268 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
8269 intrinsic returns an i32 value that has the four bytes of the input i32
8270 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
8271 returned i32 will have its bytes in 3, 2, 1, 0 order. The
8272 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
8273 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
8276 '``llvm.ctpop.*``' Intrinsic
8277 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8282 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
8283 bit width, or on any vector with integer elements. Not all targets
8284 support all bit widths or vector types, however.
8288 declare i8 @llvm.ctpop.i8(i8 <src>)
8289 declare i16 @llvm.ctpop.i16(i16 <src>)
8290 declare i32 @llvm.ctpop.i32(i32 <src>)
8291 declare i64 @llvm.ctpop.i64(i64 <src>)
8292 declare i256 @llvm.ctpop.i256(i256 <src>)
8293 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
8298 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
8304 The only argument is the value to be counted. The argument may be of any
8305 integer type, or a vector with integer elements. The return type must
8306 match the argument type.
8311 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
8312 each element of a vector.
8314 '``llvm.ctlz.*``' Intrinsic
8315 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8320 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
8321 integer bit width, or any vector whose elements are integers. Not all
8322 targets support all bit widths or vector types, however.
8326 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
8327 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
8328 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
8329 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
8330 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
8331 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
8336 The '``llvm.ctlz``' family of intrinsic functions counts the number of
8337 leading zeros in a variable.
8342 The first argument is the value to be counted. This argument may be of
8343 any integer type, or a vectory with integer element type. The return
8344 type must match the first argument type.
8346 The second argument must be a constant and is a flag to indicate whether
8347 the intrinsic should ensure that a zero as the first argument produces a
8348 defined result. Historically some architectures did not provide a
8349 defined result for zero values as efficiently, and many algorithms are
8350 now predicated on avoiding zero-value inputs.
8355 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
8356 zeros in a variable, or within each element of the vector. If
8357 ``src == 0`` then the result is the size in bits of the type of ``src``
8358 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
8359 ``llvm.ctlz(i32 2) = 30``.
8361 '``llvm.cttz.*``' Intrinsic
8362 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8367 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
8368 integer bit width, or any vector of integer elements. Not all targets
8369 support all bit widths or vector types, however.
8373 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
8374 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
8375 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
8376 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
8377 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
8378 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
8383 The '``llvm.cttz``' family of intrinsic functions counts the number of
8389 The first argument is the value to be counted. This argument may be of
8390 any integer type, or a vectory with integer element type. The return
8391 type must match the first argument type.
8393 The second argument must be a constant and is a flag to indicate whether
8394 the intrinsic should ensure that a zero as the first argument produces a
8395 defined result. Historically some architectures did not provide a
8396 defined result for zero values as efficiently, and many algorithms are
8397 now predicated on avoiding zero-value inputs.
8402 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
8403 zeros in a variable, or within each element of a vector. If ``src == 0``
8404 then the result is the size in bits of the type of ``src`` if
8405 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
8406 ``llvm.cttz(2) = 1``.
8408 Arithmetic with Overflow Intrinsics
8409 -----------------------------------
8411 LLVM provides intrinsics for some arithmetic with overflow operations.
8413 '``llvm.sadd.with.overflow.*``' Intrinsics
8414 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8419 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
8420 on any integer bit width.
8424 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
8425 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8426 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
8431 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8432 a signed addition of the two arguments, and indicate whether an overflow
8433 occurred during the signed summation.
8438 The arguments (%a and %b) and the first element of the result structure
8439 may be of integer types of any bit width, but they must have the same
8440 bit width. The second element of the result structure must be of type
8441 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8447 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
8448 a signed addition of the two variables. They return a structure --- the
8449 first element of which is the signed summation, and the second element
8450 of which is a bit specifying if the signed summation resulted in an
8456 .. code-block:: llvm
8458 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
8459 %sum = extractvalue {i32, i1} %res, 0
8460 %obit = extractvalue {i32, i1} %res, 1
8461 br i1 %obit, label %overflow, label %normal
8463 '``llvm.uadd.with.overflow.*``' Intrinsics
8464 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8469 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
8470 on any integer bit width.
8474 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
8475 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8476 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
8481 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8482 an unsigned addition of the two arguments, and indicate whether a carry
8483 occurred during the unsigned summation.
8488 The arguments (%a and %b) and the first element of the result structure
8489 may be of integer types of any bit width, but they must have the same
8490 bit width. The second element of the result structure must be of type
8491 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8497 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
8498 an unsigned addition of the two arguments. They return a structure --- the
8499 first element of which is the sum, and the second element of which is a
8500 bit specifying if the unsigned summation resulted in a carry.
8505 .. code-block:: llvm
8507 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
8508 %sum = extractvalue {i32, i1} %res, 0
8509 %obit = extractvalue {i32, i1} %res, 1
8510 br i1 %obit, label %carry, label %normal
8512 '``llvm.ssub.with.overflow.*``' Intrinsics
8513 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8518 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
8519 on any integer bit width.
8523 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
8524 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8525 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
8530 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8531 a signed subtraction of the two arguments, and indicate whether an
8532 overflow occurred during the signed subtraction.
8537 The arguments (%a and %b) and the first element of the result structure
8538 may be of integer types of any bit width, but they must have the same
8539 bit width. The second element of the result structure must be of type
8540 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8546 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
8547 a signed subtraction of the two arguments. They return a structure --- the
8548 first element of which is the subtraction, and the second element of
8549 which is a bit specifying if the signed subtraction resulted in an
8555 .. code-block:: llvm
8557 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
8558 %sum = extractvalue {i32, i1} %res, 0
8559 %obit = extractvalue {i32, i1} %res, 1
8560 br i1 %obit, label %overflow, label %normal
8562 '``llvm.usub.with.overflow.*``' Intrinsics
8563 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8568 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
8569 on any integer bit width.
8573 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
8574 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8575 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
8580 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8581 an unsigned subtraction of the two arguments, and indicate whether an
8582 overflow occurred during the unsigned subtraction.
8587 The arguments (%a and %b) and the first element of the result structure
8588 may be of integer types of any bit width, but they must have the same
8589 bit width. The second element of the result structure must be of type
8590 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8596 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
8597 an unsigned subtraction of the two arguments. They return a structure ---
8598 the first element of which is the subtraction, and the second element of
8599 which is a bit specifying if the unsigned subtraction resulted in an
8605 .. code-block:: llvm
8607 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
8608 %sum = extractvalue {i32, i1} %res, 0
8609 %obit = extractvalue {i32, i1} %res, 1
8610 br i1 %obit, label %overflow, label %normal
8612 '``llvm.smul.with.overflow.*``' Intrinsics
8613 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8618 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
8619 on any integer bit width.
8623 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
8624 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8625 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
8630 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8631 a signed multiplication of the two arguments, and indicate whether an
8632 overflow occurred during the signed multiplication.
8637 The arguments (%a and %b) and the first element of the result structure
8638 may be of integer types of any bit width, but they must have the same
8639 bit width. The second element of the result structure must be of type
8640 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
8646 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
8647 a signed multiplication of the two arguments. They return a structure ---
8648 the first element of which is the multiplication, and the second element
8649 of which is a bit specifying if the signed multiplication resulted in an
8655 .. code-block:: llvm
8657 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8658 %sum = extractvalue {i32, i1} %res, 0
8659 %obit = extractvalue {i32, i1} %res, 1
8660 br i1 %obit, label %overflow, label %normal
8662 '``llvm.umul.with.overflow.*``' Intrinsics
8663 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8668 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
8669 on any integer bit width.
8673 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
8674 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8675 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
8680 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
8681 a unsigned multiplication of the two arguments, and indicate whether an
8682 overflow occurred during the unsigned multiplication.
8687 The arguments (%a and %b) and the first element of the result structure
8688 may be of integer types of any bit width, but they must have the same
8689 bit width. The second element of the result structure must be of type
8690 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
8696 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
8697 an unsigned multiplication of the two arguments. They return a structure ---
8698 the first element of which is the multiplication, and the second
8699 element of which is a bit specifying if the unsigned multiplication
8700 resulted in an overflow.
8705 .. code-block:: llvm
8707 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8708 %sum = extractvalue {i32, i1} %res, 0
8709 %obit = extractvalue {i32, i1} %res, 1
8710 br i1 %obit, label %overflow, label %normal
8712 Specialised Arithmetic Intrinsics
8713 ---------------------------------
8715 '``llvm.fmuladd.*``' Intrinsic
8716 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8723 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
8724 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
8729 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
8730 expressions that can be fused if the code generator determines that (a) the
8731 target instruction set has support for a fused operation, and (b) that the
8732 fused operation is more efficient than the equivalent, separate pair of mul
8733 and add instructions.
8738 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
8739 multiplicands, a and b, and an addend c.
8748 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
8750 is equivalent to the expression a \* b + c, except that rounding will
8751 not be performed between the multiplication and addition steps if the
8752 code generator fuses the operations. Fusion is not guaranteed, even if
8753 the target platform supports it. If a fused multiply-add is required the
8754 corresponding llvm.fma.\* intrinsic function should be used
8755 instead. This never sets errno, just as '``llvm.fma.*``'.
8760 .. code-block:: llvm
8762 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields float:r2 = (a * b) + c
8764 Half Precision Floating Point Intrinsics
8765 ----------------------------------------
8767 For most target platforms, half precision floating point is a
8768 storage-only format. This means that it is a dense encoding (in memory)
8769 but does not support computation in the format.
8771 This means that code must first load the half-precision floating point
8772 value as an i16, then convert it to float with
8773 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
8774 then be performed on the float value (including extending to double
8775 etc). To store the value back to memory, it is first converted to float
8776 if needed, then converted to i16 with
8777 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
8780 .. _int_convert_to_fp16:
8782 '``llvm.convert.to.fp16``' Intrinsic
8783 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8790 declare i16 @llvm.convert.to.fp16.f32(float %a)
8791 declare i16 @llvm.convert.to.fp16.f64(double %a)
8796 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
8797 conventional floating point type to half precision floating point format.
8802 The intrinsic function contains single argument - the value to be
8808 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion from a
8809 conventional floating point format to half precision floating point format. The
8810 return value is an ``i16`` which contains the converted number.
8815 .. code-block:: llvm
8817 %res = call i16 @llvm.convert.to.fp16.f32(float %a)
8818 store i16 %res, i16* @x, align 2
8820 .. _int_convert_from_fp16:
8822 '``llvm.convert.from.fp16``' Intrinsic
8823 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8830 declare float @llvm.convert.from.fp16.f32(i16 %a)
8831 declare double @llvm.convert.from.fp16.f64(i16 %a)
8836 The '``llvm.convert.from.fp16``' intrinsic function performs a
8837 conversion from half precision floating point format to single precision
8838 floating point format.
8843 The intrinsic function contains single argument - the value to be
8849 The '``llvm.convert.from.fp16``' intrinsic function performs a
8850 conversion from half single precision floating point format to single
8851 precision floating point format. The input half-float value is
8852 represented by an ``i16`` value.
8857 .. code-block:: llvm
8859 %a = load i16* @x, align 2
8860 %res = call float @llvm.convert.from.fp16(i16 %a)
8865 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
8866 prefix), are described in the `LLVM Source Level
8867 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
8870 Exception Handling Intrinsics
8871 -----------------------------
8873 The LLVM exception handling intrinsics (which all start with
8874 ``llvm.eh.`` prefix), are described in the `LLVM Exception
8875 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
8879 Trampoline Intrinsics
8880 ---------------------
8882 These intrinsics make it possible to excise one parameter, marked with
8883 the :ref:`nest <nest>` attribute, from a function. The result is a
8884 callable function pointer lacking the nest parameter - the caller does
8885 not need to provide a value for it. Instead, the value to use is stored
8886 in advance in a "trampoline", a block of memory usually allocated on the
8887 stack, which also contains code to splice the nest value into the
8888 argument list. This is used to implement the GCC nested function address
8891 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
8892 then the resulting function pointer has signature ``i32 (i32, i32)*``.
8893 It can be created as follows:
8895 .. code-block:: llvm
8897 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
8898 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
8899 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
8900 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
8901 %fp = bitcast i8* %p to i32 (i32, i32)*
8903 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
8904 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
8908 '``llvm.init.trampoline``' Intrinsic
8909 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8916 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
8921 This fills the memory pointed to by ``tramp`` with executable code,
8922 turning it into a trampoline.
8927 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
8928 pointers. The ``tramp`` argument must point to a sufficiently large and
8929 sufficiently aligned block of memory; this memory is written to by the
8930 intrinsic. Note that the size and the alignment are target-specific -
8931 LLVM currently provides no portable way of determining them, so a
8932 front-end that generates this intrinsic needs to have some
8933 target-specific knowledge. The ``func`` argument must hold a function
8934 bitcast to an ``i8*``.
8939 The block of memory pointed to by ``tramp`` is filled with target
8940 dependent code, turning it into a function. Then ``tramp`` needs to be
8941 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
8942 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
8943 function's signature is the same as that of ``func`` with any arguments
8944 marked with the ``nest`` attribute removed. At most one such ``nest``
8945 argument is allowed, and it must be of pointer type. Calling the new
8946 function is equivalent to calling ``func`` with the same argument list,
8947 but with ``nval`` used for the missing ``nest`` argument. If, after
8948 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
8949 modified, then the effect of any later call to the returned function
8950 pointer is undefined.
8954 '``llvm.adjust.trampoline``' Intrinsic
8955 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8962 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
8967 This performs any required machine-specific adjustment to the address of
8968 a trampoline (passed as ``tramp``).
8973 ``tramp`` must point to a block of memory which already has trampoline
8974 code filled in by a previous call to
8975 :ref:`llvm.init.trampoline <int_it>`.
8980 On some architectures the address of the code to be executed needs to be
8981 different than the address where the trampoline is actually stored. This
8982 intrinsic returns the executable address corresponding to ``tramp``
8983 after performing the required machine specific adjustments. The pointer
8984 returned can then be :ref:`bitcast and executed <int_trampoline>`.
8989 This class of intrinsics provides information about the lifetime of
8990 memory objects and ranges where variables are immutable.
8994 '``llvm.lifetime.start``' Intrinsic
8995 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9002 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
9007 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
9013 The first argument is a constant integer representing the size of the
9014 object, or -1 if it is variable sized. The second argument is a pointer
9020 This intrinsic indicates that before this point in the code, the value
9021 of the memory pointed to by ``ptr`` is dead. This means that it is known
9022 to never be used and has an undefined value. A load from the pointer
9023 that precedes this intrinsic can be replaced with ``'undef'``.
9027 '``llvm.lifetime.end``' Intrinsic
9028 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9035 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
9040 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
9046 The first argument is a constant integer representing the size of the
9047 object, or -1 if it is variable sized. The second argument is a pointer
9053 This intrinsic indicates that after this point in the code, the value of
9054 the memory pointed to by ``ptr`` is dead. This means that it is known to
9055 never be used and has an undefined value. Any stores into the memory
9056 object following this intrinsic may be removed as dead.
9058 '``llvm.invariant.start``' Intrinsic
9059 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9066 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
9071 The '``llvm.invariant.start``' intrinsic specifies that the contents of
9072 a memory object will not change.
9077 The first argument is a constant integer representing the size of the
9078 object, or -1 if it is variable sized. The second argument is a pointer
9084 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
9085 the return value, the referenced memory location is constant and
9088 '``llvm.invariant.end``' Intrinsic
9089 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9096 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
9101 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
9102 memory object are mutable.
9107 The first argument is the matching ``llvm.invariant.start`` intrinsic.
9108 The second argument is a constant integer representing the size of the
9109 object, or -1 if it is variable sized and the third argument is a
9110 pointer to the object.
9115 This intrinsic indicates that the memory is mutable again.
9120 This class of intrinsics is designed to be generic and has no specific
9123 '``llvm.var.annotation``' Intrinsic
9124 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9131 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9136 The '``llvm.var.annotation``' intrinsic.
9141 The first argument is a pointer to a value, the second is a pointer to a
9142 global string, the third is a pointer to a global string which is the
9143 source file name, and the last argument is the line number.
9148 This intrinsic allows annotation of local variables with arbitrary
9149 strings. This can be useful for special purpose optimizations that want
9150 to look for these annotations. These have no other defined use; they are
9151 ignored by code generation and optimization.
9153 '``llvm.ptr.annotation.*``' Intrinsic
9154 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9159 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
9160 pointer to an integer of any width. *NOTE* you must specify an address space for
9161 the pointer. The identifier for the default address space is the integer
9166 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
9167 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
9168 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
9169 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
9170 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
9175 The '``llvm.ptr.annotation``' intrinsic.
9180 The first argument is a pointer to an integer value of arbitrary bitwidth
9181 (result of some expression), the second is a pointer to a global string, the
9182 third is a pointer to a global string which is the source file name, and the
9183 last argument is the line number. It returns the value of the first argument.
9188 This intrinsic allows annotation of a pointer to an integer with arbitrary
9189 strings. This can be useful for special purpose optimizations that want to look
9190 for these annotations. These have no other defined use; they are ignored by code
9191 generation and optimization.
9193 '``llvm.annotation.*``' Intrinsic
9194 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9199 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
9200 any integer bit width.
9204 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
9205 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
9206 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
9207 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
9208 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
9213 The '``llvm.annotation``' intrinsic.
9218 The first argument is an integer value (result of some expression), the
9219 second is a pointer to a global string, the third is a pointer to a
9220 global string which is the source file name, and the last argument is
9221 the line number. It returns the value of the first argument.
9226 This intrinsic allows annotations to be put on arbitrary expressions
9227 with arbitrary strings. This can be useful for special purpose
9228 optimizations that want to look for these annotations. These have no
9229 other defined use; they are ignored by code generation and optimization.
9231 '``llvm.trap``' Intrinsic
9232 ^^^^^^^^^^^^^^^^^^^^^^^^^
9239 declare void @llvm.trap() noreturn nounwind
9244 The '``llvm.trap``' intrinsic.
9254 This intrinsic is lowered to the target dependent trap instruction. If
9255 the target does not have a trap instruction, this intrinsic will be
9256 lowered to a call of the ``abort()`` function.
9258 '``llvm.debugtrap``' Intrinsic
9259 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9266 declare void @llvm.debugtrap() nounwind
9271 The '``llvm.debugtrap``' intrinsic.
9281 This intrinsic is lowered to code which is intended to cause an
9282 execution trap with the intention of requesting the attention of a
9285 '``llvm.stackprotector``' Intrinsic
9286 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9293 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
9298 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
9299 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
9300 is placed on the stack before local variables.
9305 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
9306 The first argument is the value loaded from the stack guard
9307 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
9308 enough space to hold the value of the guard.
9313 This intrinsic causes the prologue/epilogue inserter to force the position of
9314 the ``AllocaInst`` stack slot to be before local variables on the stack. This is
9315 to ensure that if a local variable on the stack is overwritten, it will destroy
9316 the value of the guard. When the function exits, the guard on the stack is
9317 checked against the original guard by ``llvm.stackprotectorcheck``. If they are
9318 different, then ``llvm.stackprotectorcheck`` causes the program to abort by
9319 calling the ``__stack_chk_fail()`` function.
9321 '``llvm.stackprotectorcheck``' Intrinsic
9322 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9329 declare void @llvm.stackprotectorcheck(i8** <guard>)
9334 The ``llvm.stackprotectorcheck`` intrinsic compares ``guard`` against an already
9335 created stack protector and if they are not equal calls the
9336 ``__stack_chk_fail()`` function.
9341 The ``llvm.stackprotectorcheck`` intrinsic requires one pointer argument, the
9342 the variable ``@__stack_chk_guard``.
9347 This intrinsic is provided to perform the stack protector check by comparing
9348 ``guard`` with the stack slot created by ``llvm.stackprotector`` and if the
9349 values do not match call the ``__stack_chk_fail()`` function.
9351 The reason to provide this as an IR level intrinsic instead of implementing it
9352 via other IR operations is that in order to perform this operation at the IR
9353 level without an intrinsic, one would need to create additional basic blocks to
9354 handle the success/failure cases. This makes it difficult to stop the stack
9355 protector check from disrupting sibling tail calls in Codegen. With this
9356 intrinsic, we are able to generate the stack protector basic blocks late in
9357 codegen after the tail call decision has occurred.
9359 '``llvm.objectsize``' Intrinsic
9360 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9367 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
9368 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
9373 The ``llvm.objectsize`` intrinsic is designed to provide information to
9374 the optimizers to determine at compile time whether a) an operation
9375 (like memcpy) will overflow a buffer that corresponds to an object, or
9376 b) that a runtime check for overflow isn't necessary. An object in this
9377 context means an allocation of a specific class, structure, array, or
9383 The ``llvm.objectsize`` intrinsic takes two arguments. The first
9384 argument is a pointer to or into the ``object``. The second argument is
9385 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
9386 or -1 (if false) when the object size is unknown. The second argument
9387 only accepts constants.
9392 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
9393 the size of the object concerned. If the size cannot be determined at
9394 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
9395 on the ``min`` argument).
9397 '``llvm.expect``' Intrinsic
9398 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
9403 This is an overloaded intrinsic. You can use ``llvm.expect`` on any
9408 declare i1 @llvm.expect.i1(i1 <val>, i1 <expected_val>)
9409 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
9410 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
9415 The ``llvm.expect`` intrinsic provides information about expected (the
9416 most probable) value of ``val``, which can be used by optimizers.
9421 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
9422 a value. The second argument is an expected value, this needs to be a
9423 constant value, variables are not allowed.
9428 This intrinsic is lowered to the ``val``.
9430 '``llvm.assume``' Intrinsic
9431 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9438 declare void @llvm.assume(i1 %cond)
9443 The ``llvm.assume`` allows the optimizer to assume that the provided
9444 condition is true. This information can then be used in simplifying other parts
9450 The condition which the optimizer may assume is always true.
9455 The intrinsic allows the optimizer to assume that the provided condition is
9456 always true whenever the control flow reaches the intrinsic call. No code is
9457 generated for this intrinsic, and instructions that contribute only to the
9458 provided condition are not used for code generation. If the condition is
9459 violated during execution, the behavior is undefined.
9461 Please note that optimizer might limit the transformations performed on values
9462 used by the ``llvm.assume`` intrinsic in order to preserve the instructions
9463 only used to form the intrinsic's input argument. This might prove undesirable
9464 if the extra information provided by the ``llvm.assume`` intrinsic does cause
9465 sufficient overall improvement in code quality. For this reason,
9466 ``llvm.assume`` should not be used to document basic mathematical invariants
9467 that the optimizer can otherwise deduce or facts that are of little use to the
9470 '``llvm.donothing``' Intrinsic
9471 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
9478 declare void @llvm.donothing() nounwind readnone
9483 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's the
9484 only intrinsic that can be called with an invoke instruction.
9494 This intrinsic does nothing, and it's removed by optimizers and ignored
9497 Stack Map Intrinsics
9498 --------------------
9500 LLVM provides experimental intrinsics to support runtime patching
9501 mechanisms commonly desired in dynamic language JITs. These intrinsics
9502 are described in :doc:`StackMaps`.