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).
133 It also shows a convention that we follow in this document. When
134 demonstrating instructions, we will follow an instruction with a comment
135 that defines the type and name of value produced.
143 LLVM programs are composed of ``Module``'s, each of which is a
144 translation unit of the input programs. Each module consists of
145 functions, global variables, and symbol table entries. Modules may be
146 combined together with the LLVM linker, which merges function (and
147 global variable) definitions, resolves forward declarations, and merges
148 symbol table entries. Here is an example of the "hello world" module:
152 ; Declare the string constant as a global constant.
153 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
155 ; External declaration of the puts function
156 declare i32 @puts(i8* nocapture) nounwind
158 ; Definition of main function
159 define i32 @main() { ; i32()*
160 ; Convert [13 x i8]* to i8 *...
161 %cast210 = getelementptr [13 x i8]* @.str, i64 0, i64 0
163 ; Call puts function to write out the string to stdout.
164 call i32 @puts(i8* %cast210)
169 !1 = metadata !{i32 42}
172 This example is made up of a :ref:`global variable <globalvars>` named
173 "``.str``", an external declaration of the "``puts``" function, a
174 :ref:`function definition <functionstructure>` for "``main``" and
175 :ref:`named metadata <namedmetadatastructure>` "``foo``".
177 In general, a module is made up of a list of global values (where both
178 functions and global variables are global values). Global values are
179 represented by a pointer to a memory location (in this case, a pointer
180 to an array of char, and a pointer to a function), and have one of the
181 following :ref:`linkage types <linkage>`.
188 All Global Variables and Functions have one of the following types of
192 Global values with "``private``" linkage are only directly
193 accessible by objects in the current module. In particular, linking
194 code into a module with an private global value may cause the
195 private to be renamed as necessary to avoid collisions. Because the
196 symbol is private to the module, all references can be updated. This
197 doesn't show up in any symbol table in the object file.
199 Similar to ``private``, but the symbol is passed through the
200 assembler and evaluated by the linker. Unlike normal strong symbols,
201 they are removed by the linker from the final linked image
202 (executable or dynamic library).
203 ``linker_private_weak``
204 Similar to "``linker_private``", but the symbol is weak. Note that
205 ``linker_private_weak`` symbols are subject to coalescing by the
206 linker. The symbols are removed by the linker from the final linked
207 image (executable or dynamic library).
209 Similar to private, but the value shows as a local symbol
210 (``STB_LOCAL`` in the case of ELF) in the object file. This
211 corresponds to the notion of the '``static``' keyword in C.
212 ``available_externally``
213 Globals with "``available_externally``" linkage are never emitted
214 into the object file corresponding to the LLVM module. They exist to
215 allow inlining and other optimizations to take place given knowledge
216 of the definition of the global, which is known to be somewhere
217 outside the module. Globals with ``available_externally`` linkage
218 are allowed to be discarded at will, and are otherwise the same as
219 ``linkonce_odr``. This linkage type is only allowed on definitions,
222 Globals with "``linkonce``" linkage are merged with other globals of
223 the same name when linkage occurs. This can be used to implement
224 some forms of inline functions, templates, or other code which must
225 be generated in each translation unit that uses it, but where the
226 body may be overridden with a more definitive definition later.
227 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
228 that ``linkonce`` linkage does not actually allow the optimizer to
229 inline the body of this function into callers because it doesn't
230 know if this definition of the function is the definitive definition
231 within the program or whether it will be overridden by a stronger
232 definition. To enable inlining and other optimizations, use
233 "``linkonce_odr``" linkage.
235 "``weak``" linkage has the same merging semantics as ``linkonce``
236 linkage, except that unreferenced globals with ``weak`` linkage may
237 not be discarded. This is used for globals that are declared "weak"
240 "``common``" linkage is most similar to "``weak``" linkage, but they
241 are used for tentative definitions in C, such as "``int X;``" at
242 global scope. Symbols with "``common``" linkage are merged in the
243 same way as ``weak symbols``, and they may not be deleted if
244 unreferenced. ``common`` symbols may not have an explicit section,
245 must have a zero initializer, and may not be marked
246 ':ref:`constant <globalvars>`'. Functions and aliases may not have
249 .. _linkage_appending:
252 "``appending``" linkage may only be applied to global variables of
253 pointer to array type. When two global variables with appending
254 linkage are linked together, the two global arrays are appended
255 together. This is the LLVM, typesafe, equivalent of having the
256 system linker append together "sections" with identical names when
259 The semantics of this linkage follow the ELF object file model: the
260 symbol is weak until linked, if not linked, the symbol becomes null
261 instead of being an undefined reference.
262 ``linkonce_odr``, ``weak_odr``
263 Some languages allow differing globals to be merged, such as two
264 functions with different semantics. Other languages, such as
265 ``C++``, ensure that only equivalent globals are ever merged (the
266 "one definition rule" --- "ODR"). Such languages can use the
267 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
268 global will only be merged with equivalent globals. These linkage
269 types are otherwise the same as their non-``odr`` versions.
270 ``linkonce_odr_auto_hide``
271 Similar to "``linkonce_odr``", but nothing in the translation unit
272 takes the address of this definition. For instance, functions that
273 had an inline definition, but the compiler decided not to inline it.
274 ``linkonce_odr_auto_hide`` may have only ``default`` visibility. The
275 symbols are removed by the linker from the final linked image
276 (executable or dynamic library).
278 If none of the above identifiers are used, the global is externally
279 visible, meaning that it participates in linkage and can be used to
280 resolve external symbol references.
282 The next two types of linkage are targeted for Microsoft Windows
283 platform only. They are designed to support importing (exporting)
284 symbols from (to) DLLs (Dynamic Link Libraries).
287 "``dllimport``" linkage causes the compiler to reference a function
288 or variable via a global pointer to a pointer that is set up by the
289 DLL exporting the symbol. On Microsoft Windows targets, the pointer
290 name is formed by combining ``__imp_`` and the function or variable
293 "``dllexport``" linkage causes the compiler to provide a global
294 pointer to a pointer in a DLL, so that it can be referenced with the
295 ``dllimport`` attribute. On Microsoft Windows targets, the pointer
296 name is formed by combining ``__imp_`` and the function or variable
299 For example, since the "``.LC0``" variable is defined to be internal, if
300 another module defined a "``.LC0``" variable and was linked with this
301 one, one of the two would be renamed, preventing a collision. Since
302 "``main``" and "``puts``" are external (i.e., lacking any linkage
303 declarations), they are accessible outside of the current module.
305 It is illegal for a function *declaration* to have any linkage type
306 other than ``external``, ``dllimport`` or ``extern_weak``.
308 Aliases can have only ``external``, ``internal``, ``weak`` or
309 ``weak_odr`` linkages.
316 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
317 :ref:`invokes <i_invoke>` can all have an optional calling convention
318 specified for the call. The calling convention of any pair of dynamic
319 caller/callee must match, or the behavior of the program is undefined.
320 The following calling conventions are supported by LLVM, and more may be
323 "``ccc``" - The C calling convention
324 This calling convention (the default if no other calling convention
325 is specified) matches the target C calling conventions. This calling
326 convention supports varargs function calls and tolerates some
327 mismatch in the declared prototype and implemented declaration of
328 the function (as does normal C).
329 "``fastcc``" - The fast calling convention
330 This calling convention attempts to make calls as fast as possible
331 (e.g. by passing things in registers). This calling convention
332 allows the target to use whatever tricks it wants to produce fast
333 code for the target, without having to conform to an externally
334 specified ABI (Application Binary Interface). `Tail calls can only
335 be optimized when this, the GHC or the HiPE convention is
336 used. <CodeGenerator.html#id80>`_ This calling convention does not
337 support varargs and requires the prototype of all callees to exactly
338 match the prototype of the function definition.
339 "``coldcc``" - The cold calling convention
340 This calling convention attempts to make code in the caller as
341 efficient as possible under the assumption that the call is not
342 commonly executed. As such, these calls often preserve all registers
343 so that the call does not break any live ranges in the caller side.
344 This calling convention does not support varargs and requires the
345 prototype of all callees to exactly match the prototype of the
347 "``cc 10``" - GHC convention
348 This calling convention has been implemented specifically for use by
349 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
350 It passes everything in registers, going to extremes to achieve this
351 by disabling callee save registers. This calling convention should
352 not be used lightly but only for specific situations such as an
353 alternative to the *register pinning* performance technique often
354 used when implementing functional programming languages. At the
355 moment only X86 supports this convention and it has the following
358 - On *X86-32* only supports up to 4 bit type parameters. No
359 floating point types are supported.
360 - On *X86-64* only supports up to 10 bit type parameters and 6
361 floating point parameters.
363 This calling convention supports `tail call
364 optimization <CodeGenerator.html#id80>`_ but requires both the
365 caller and callee are using it.
366 "``cc 11``" - The HiPE calling convention
367 This calling convention has been implemented specifically for use by
368 the `High-Performance Erlang
369 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
370 native code compiler of the `Ericsson's Open Source Erlang/OTP
371 system <http://www.erlang.org/download.shtml>`_. It uses more
372 registers for argument passing than the ordinary C calling
373 convention and defines no callee-saved registers. The calling
374 convention properly supports `tail call
375 optimization <CodeGenerator.html#id80>`_ but requires that both the
376 caller and the callee use it. It uses a *register pinning*
377 mechanism, similar to GHC's convention, for keeping frequently
378 accessed runtime components pinned to specific hardware registers.
379 At the moment only X86 supports this convention (both 32 and 64
381 "``cc <n>``" - Numbered convention
382 Any calling convention may be specified by number, allowing
383 target-specific calling conventions to be used. Target specific
384 calling conventions start at 64.
386 More calling conventions can be added/defined on an as-needed basis, to
387 support Pascal conventions or any other well-known target-independent
393 All Global Variables and Functions have one of the following visibility
396 "``default``" - Default style
397 On targets that use the ELF object file format, default visibility
398 means that the declaration is visible to other modules and, in
399 shared libraries, means that the declared entity may be overridden.
400 On Darwin, default visibility means that the declaration is visible
401 to other modules. Default visibility corresponds to "external
402 linkage" in the language.
403 "``hidden``" - Hidden style
404 Two declarations of an object with hidden visibility refer to the
405 same object if they are in the same shared object. Usually, hidden
406 visibility indicates that the symbol will not be placed into the
407 dynamic symbol table, so no other module (executable or shared
408 library) can reference it directly.
409 "``protected``" - Protected style
410 On ELF, protected visibility indicates that the symbol will be
411 placed in the dynamic symbol table, but that references within the
412 defining module will bind to the local symbol. That is, the symbol
413 cannot be overridden by another module.
418 LLVM IR allows you to specify name aliases for certain types. This can
419 make it easier to read the IR and make the IR more condensed
420 (particularly when recursive types are involved). An example of a name
425 %mytype = type { %mytype*, i32 }
427 You may give a name to any :ref:`type <typesystem>` except
428 ":ref:`void <t_void>`". Type name aliases may be used anywhere a type is
429 expected with the syntax "%mytype".
431 Note that type names are aliases for the structural type that they
432 indicate, and that you can therefore specify multiple names for the same
433 type. This often leads to confusing behavior when dumping out a .ll
434 file. Since LLVM IR uses structural typing, the name is not part of the
435 type. When printing out LLVM IR, the printer will pick *one name* to
436 render all types of a particular shape. This means that if you have code
437 where two different source types end up having the same LLVM type, that
438 the dumper will sometimes print the "wrong" or unexpected type. This is
439 an important design point and isn't going to change.
446 Global variables define regions of memory allocated at compilation time
447 instead of run-time. Global variables may optionally be initialized, may
448 have an explicit section to be placed in, and may have an optional
449 explicit alignment specified.
451 A variable may be defined as ``thread_local``, which means that it will
452 not be shared by threads (each thread will have a separated copy of the
453 variable). Not all targets support thread-local variables. Optionally, a
454 TLS model may be specified:
457 For variables that are only used within the current shared library.
459 For variables in modules that will not be loaded dynamically.
461 For variables defined in the executable and only used within it.
463 The models correspond to the ELF TLS models; see `ELF Handling For
464 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
465 more information on under which circumstances the different models may
466 be used. The target may choose a different TLS model if the specified
467 model is not supported, or if a better choice of model can be made.
469 A variable may be defined as a global ``constant``, which indicates that
470 the contents of the variable will **never** be modified (enabling better
471 optimization, allowing the global data to be placed in the read-only
472 section of an executable, etc). Note that variables that need runtime
473 initialization cannot be marked ``constant`` as there is a store to the
476 LLVM explicitly allows *declarations* of global variables to be marked
477 constant, even if the final definition of the global is not. This
478 capability can be used to enable slightly better optimization of the
479 program, but requires the language definition to guarantee that
480 optimizations based on the 'constantness' are valid for the translation
481 units that do not include the definition.
483 As SSA values, global variables define pointer values that are in scope
484 (i.e. they dominate) all basic blocks in the program. Global variables
485 always define a pointer to their "content" type because they describe a
486 region of memory, and all memory objects in LLVM are accessed through
489 Global variables can be marked with ``unnamed_addr`` which indicates
490 that the address is not significant, only the content. Constants marked
491 like this can be merged with other constants if they have the same
492 initializer. Note that a constant with significant address *can* be
493 merged with a ``unnamed_addr`` constant, the result being a constant
494 whose address is significant.
496 A global variable may be declared to reside in a target-specific
497 numbered address space. For targets that support them, address spaces
498 may affect how optimizations are performed and/or what target
499 instructions are used to access the variable. The default address space
500 is zero. The address space qualifier must precede any other attributes.
502 LLVM allows an explicit section to be specified for globals. If the
503 target supports it, it will emit globals to the section specified.
505 By default, global initializers are optimized by assuming that global
506 variables defined within the module are not modified from their
507 initial values before the start of the global initializer. This is
508 true even for variables potentially accessible from outside the
509 module, including those with external linkage or appearing in
510 ``@llvm.used``. This assumption may be suppressed by marking the
511 variable with ``externally_initialized``.
513 An explicit alignment may be specified for a global, which must be a
514 power of 2. If not present, or if the alignment is set to zero, the
515 alignment of the global is set by the target to whatever it feels
516 convenient. If an explicit alignment is specified, the global is forced
517 to have exactly that alignment. Targets and optimizers are not allowed
518 to over-align the global if the global has an assigned section. In this
519 case, the extra alignment could be observable: for example, code could
520 assume that the globals are densely packed in their section and try to
521 iterate over them as an array, alignment padding would break this
524 For example, the following defines a global in a numbered address space
525 with an initializer, section, and alignment:
529 @G = addrspace(5) constant float 1.0, section "foo", align 4
531 The following example defines a thread-local global with the
532 ``initialexec`` TLS model:
536 @G = thread_local(initialexec) global i32 0, align 4
538 .. _functionstructure:
543 LLVM function definitions consist of the "``define``" keyword, an
544 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
545 style <visibility>`, an optional :ref:`calling convention <callingconv>`,
546 an optional ``unnamed_addr`` attribute, a return type, an optional
547 :ref:`parameter attribute <paramattrs>` for the return type, a function
548 name, a (possibly empty) argument list (each with optional :ref:`parameter
549 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
550 an optional section, an optional alignment, an optional :ref:`garbage
551 collector name <gc>`, an opening curly brace, a list of basic blocks,
552 and a closing curly brace.
554 LLVM function declarations consist of the "``declare``" keyword, an
555 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
556 style <visibility>`, an optional :ref:`calling convention <callingconv>`,
557 an optional ``unnamed_addr`` attribute, a return type, an optional
558 :ref:`parameter attribute <paramattrs>` for the return type, a function
559 name, a possibly empty list of arguments, an optional alignment, and an
560 optional :ref:`garbage collector name <gc>`.
562 A function definition contains a list of basic blocks, forming the CFG
563 (Control Flow Graph) for the function. Each basic block may optionally
564 start with a label (giving the basic block a symbol table entry),
565 contains a list of instructions, and ends with a
566 :ref:`terminator <terminators>` instruction (such as a branch or function
567 return). If explicit label is not provided, a block is assigned an
568 implicit numbered label, using a next value from the same counter as used
569 for unnamed temporaries (:ref:`see above<identifiers>`). For example, if a
570 function entry block does not have explicit label, it will be assigned
571 label "%0", then first unnamed temporary in that block will be "%1", etc.
573 The first basic block in a function is special in two ways: it is
574 immediately executed on entrance to the function, and it is not allowed
575 to have predecessor basic blocks (i.e. there can not be any branches to
576 the entry block of a function). Because the block can have no
577 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
579 LLVM allows an explicit section to be specified for functions. If the
580 target supports it, it will emit functions to the section specified.
582 An explicit alignment may be specified for a function. If not present,
583 or if the alignment is set to zero, the alignment of the function is set
584 by the target to whatever it feels convenient. If an explicit alignment
585 is specified, the function is forced to have at least that much
586 alignment. All alignments must be a power of 2.
588 If the ``unnamed_addr`` attribute is given, the address is know to not
589 be significant and two identical functions can be merged.
593 define [linkage] [visibility]
595 <ResultType> @<FunctionName> ([argument list])
596 [fn Attrs] [section "name"] [align N]
602 Aliases act as "second name" for the aliasee value (which can be either
603 function, global variable, another alias or bitcast of global value).
604 Aliases may have an optional :ref:`linkage type <linkage>`, and an optional
605 :ref:`visibility style <visibility>`.
609 @<Name> = alias [Linkage] [Visibility] <AliaseeTy> @<Aliasee>
611 .. _namedmetadatastructure:
616 Named metadata is a collection of metadata. :ref:`Metadata
617 nodes <metadata>` (but not metadata strings) are the only valid
618 operands for a named metadata.
622 ; Some unnamed metadata nodes, which are referenced by the named metadata.
623 !0 = metadata !{metadata !"zero"}
624 !1 = metadata !{metadata !"one"}
625 !2 = metadata !{metadata !"two"}
627 !name = !{!0, !1, !2}
634 The return type and each parameter of a function type may have a set of
635 *parameter attributes* associated with them. Parameter attributes are
636 used to communicate additional information about the result or
637 parameters of a function. Parameter attributes are considered to be part
638 of the function, not of the function type, so functions with different
639 parameter attributes can have the same function type.
641 Parameter attributes are simple keywords that follow the type specified.
642 If multiple parameter attributes are needed, they are space separated.
647 declare i32 @printf(i8* noalias nocapture, ...)
648 declare i32 @atoi(i8 zeroext)
649 declare signext i8 @returns_signed_char()
651 Note that any attributes for the function result (``nounwind``,
652 ``readonly``) come immediately after the argument list.
654 Currently, only the following parameter attributes are defined:
657 This indicates to the code generator that the parameter or return
658 value should be zero-extended to the extent required by the target's
659 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
660 the caller (for a parameter) or the callee (for a return value).
662 This indicates to the code generator that the parameter or return
663 value should be sign-extended to the extent required by the target's
664 ABI (which is usually 32-bits) by the caller (for a parameter) or
665 the callee (for a return value).
667 This indicates that this parameter or return value should be treated
668 in a special target-dependent fashion during while emitting code for
669 a function call or return (usually, by putting it in a register as
670 opposed to memory, though some targets use it to distinguish between
671 two different kinds of registers). Use of this attribute is
674 This indicates that the pointer parameter should really be passed by
675 value to the function. The attribute implies that a hidden copy of
676 the pointee is made between the caller and the callee, so the callee
677 is unable to modify the value in the caller. This attribute is only
678 valid on LLVM pointer arguments. It is generally used to pass
679 structs and arrays by value, but is also valid on pointers to
680 scalars. The copy is considered to belong to the caller not the
681 callee (for example, ``readonly`` functions should not write to
682 ``byval`` parameters). This is not a valid attribute for return
685 The byval attribute also supports specifying an alignment with the
686 align attribute. It indicates the alignment of the stack slot to
687 form and the known alignment of the pointer specified to the call
688 site. If the alignment is not specified, then the code generator
689 makes a target-specific assumption.
692 This indicates that the pointer parameter specifies the address of a
693 structure that is the return value of the function in the source
694 program. This pointer must be guaranteed by the caller to be valid:
695 loads and stores to the structure may be assumed by the callee
696 not to trap and to be properly aligned. This may only be applied to
697 the first parameter. This is not a valid attribute for return
700 This indicates that pointer values `*based* <pointeraliasing>` on
701 the argument or return value do not alias pointer values which are
702 not *based* on it, ignoring certain "irrelevant" dependencies. For a
703 call to the parent function, dependencies between memory references
704 from before or after the call and from those during the call are
705 "irrelevant" to the ``noalias`` keyword for the arguments and return
706 value used in that call. The caller shares the responsibility with
707 the callee for ensuring that these requirements are met. For further
708 details, please see the discussion of the NoAlias response in `alias
709 analysis <AliasAnalysis.html#MustMayNo>`_.
711 Note that this definition of ``noalias`` is intentionally similar
712 to the definition of ``restrict`` in C99 for function arguments,
713 though it is slightly weaker.
715 For function return values, C99's ``restrict`` is not meaningful,
716 while LLVM's ``noalias`` is.
718 This indicates that the callee does not make any copies of the
719 pointer that outlive the callee itself. This is not a valid
720 attribute for return values.
725 This indicates that the pointer parameter can be excised using the
726 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
727 attribute for return values and can only be applied to one parameter.
730 This indicates that the value of the function always returns the value
731 of the parameter as its return value. This is an optimization hint to
732 the code generator when generating the caller, allowing tail call
733 optimization and omission of register saves and restores in some cases;
734 it is not checked or enforced when generating the callee. The parameter
735 and the function return type must be valid operands for the
736 :ref:`bitcast instruction <i_bitcast>`. This is not a valid attribute for
737 return values and can only be applied to one parameter.
741 Garbage Collector Names
742 -----------------------
744 Each function may specify a garbage collector name, which is simply a
749 define void @f() gc "name" { ... }
751 The compiler declares the supported values of *name*. Specifying a
752 collector which will cause the compiler to alter its output in order to
753 support the named garbage collection algorithm.
760 Attribute groups are groups of attributes that are referenced by objects within
761 the IR. They are important for keeping ``.ll`` files readable, because a lot of
762 functions will use the same set of attributes. In the degenerative case of a
763 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
764 group will capture the important command line flags used to build that file.
766 An attribute group is a module-level object. To use an attribute group, an
767 object references the attribute group's ID (e.g. ``#37``). An object may refer
768 to more than one attribute group. In that situation, the attributes from the
769 different groups are merged.
771 Here is an example of attribute groups for a function that should always be
772 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
776 ; Target-independent attributes:
777 attributes #0 = { alwaysinline alignstack=4 }
779 ; Target-dependent attributes:
780 attributes #1 = { "no-sse" }
782 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
783 define void @f() #0 #1 { ... }
790 Function attributes are set to communicate additional information about
791 a function. Function attributes are considered to be part of the
792 function, not of the function type, so functions with different function
793 attributes can have the same function type.
795 Function attributes are simple keywords that follow the type specified.
796 If multiple attributes are needed, they are space separated. For
801 define void @f() noinline { ... }
802 define void @f() alwaysinline { ... }
803 define void @f() alwaysinline optsize { ... }
804 define void @f() optsize { ... }
807 This attribute indicates that, when emitting the prologue and
808 epilogue, the backend should forcibly align the stack pointer.
809 Specify the desired alignment, which must be a power of two, in
812 This attribute indicates that the inliner should attempt to inline
813 this function into callers whenever possible, ignoring any active
814 inlining size threshold for this caller.
816 This attribute indicates that this function is rarely called. When
817 computing edge weights, basic blocks post-dominated by a cold
818 function call are also considered to be cold; and, thus, given low
821 This attribute suppresses lazy symbol binding for the function. This
822 may make calls to the function faster, at the cost of extra program
823 startup time if the function is not called during program startup.
825 This attribute indicates that the source code contained a hint that
826 inlining this function is desirable (such as the "inline" keyword in
827 C/C++). It is just a hint; it imposes no requirements on the
830 This attribute disables prologue / epilogue emission for the
831 function. This can have very system-specific consequences.
833 This indicates that the callee function at a call site is not
834 recognized as a built-in function. LLVM will retain the original call
835 and not replace it with equivalent code based on the semantics of the
836 built-in function. This is only valid at call sites, not on function
837 declarations or definitions.
839 This attribute indicates that calls to the function cannot be
840 duplicated. A call to a ``noduplicate`` function may be moved
841 within its parent function, but may not be duplicated within
844 A function containing a ``noduplicate`` call may still
845 be an inlining candidate, provided that the call is not
846 duplicated by inlining. That implies that the function has
847 internal linkage and only has one call site, so the original
848 call is dead after inlining.
850 This attributes disables implicit floating point instructions.
852 This attribute indicates that the inliner should never inline this
853 function in any situation. This attribute may not be used together
854 with the ``alwaysinline`` attribute.
856 This attribute indicates that the code generator should not use a
857 red zone, even if the target-specific ABI normally permits it.
859 This function attribute indicates that the function never returns
860 normally. This produces undefined behavior at runtime if the
861 function ever does dynamically return.
863 This function attribute indicates that the function never returns
864 with an unwind or exceptional control flow. If the function does
865 unwind, its runtime behavior is undefined.
867 This attribute suggests that optimization passes and code generator
868 passes make choices that keep the code size of this function low,
869 and otherwise do optimizations specifically to reduce code size.
871 This attribute indicates that the function computes its result (or
872 decides to unwind an exception) based strictly on its arguments,
873 without dereferencing any pointer arguments or otherwise accessing
874 any mutable state (e.g. memory, control registers, etc) visible to
875 caller functions. It does not write through any pointer arguments
876 (including ``byval`` arguments) and never changes any state visible
877 to callers. This means that it cannot unwind exceptions by calling
878 the ``C++`` exception throwing methods.
880 This attribute indicates that the function does not write through
881 any pointer arguments (including ``byval`` arguments) or otherwise
882 modify any state (e.g. memory, control registers, etc) visible to
883 caller functions. It may dereference pointer arguments and read
884 state that may be set in the caller. A readonly function always
885 returns the same value (or unwinds an exception identically) when
886 called with the same set of arguments and global state. It cannot
887 unwind an exception by calling the ``C++`` exception throwing
890 This attribute indicates that this function can return twice. The C
891 ``setjmp`` is an example of such a function. The compiler disables
892 some optimizations (like tail calls) in the caller of these
895 This attribute indicates that AddressSanitizer checks
896 (dynamic address safety analysis) are enabled for this function.
898 This attribute indicates that MemorySanitizer checks (dynamic detection
899 of accesses to uninitialized memory) are enabled for this function.
901 This attribute indicates that ThreadSanitizer checks
902 (dynamic thread safety analysis) are enabled for this function.
904 This attribute indicates that the function should emit a stack
905 smashing protector. It is in the form of a "canary" --- a random value
906 placed on the stack before the local variables that's checked upon
907 return from the function to see if it has been overwritten. A
908 heuristic is used to determine if a function needs stack protectors
909 or not. The heuristic used will enable protectors for functions with:
911 - Character arrays larger than ``ssp-buffer-size`` (default 8).
912 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
913 - Calls to alloca() with variable sizes or constant sizes greater than
916 If a function that has an ``ssp`` attribute is inlined into a
917 function that doesn't have an ``ssp`` attribute, then the resulting
918 function will have an ``ssp`` attribute.
920 This attribute indicates that the function should *always* emit a
921 stack smashing protector. This overrides the ``ssp`` function
924 If a function that has an ``sspreq`` attribute is inlined into a
925 function that doesn't have an ``sspreq`` attribute or which has an
926 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
927 an ``sspreq`` attribute.
929 This attribute indicates that the function should emit a stack smashing
930 protector. This attribute causes a strong heuristic to be used when
931 determining if a function needs stack protectors. The strong heuristic
932 will enable protectors for functions with:
934 - Arrays of any size and type
935 - Aggregates containing an array of any size and type.
937 - Local variables that have had their address taken.
939 This overrides the ``ssp`` function attribute.
941 If a function that has an ``sspstrong`` attribute is inlined into a
942 function that doesn't have an ``sspstrong`` attribute, then the
943 resulting function will have an ``sspstrong`` attribute.
945 This attribute indicates that the ABI being targeted requires that
946 an unwind table entry be produce for this function even if we can
947 show that no exceptions passes by it. This is normally the case for
948 the ELF x86-64 abi, but it can be disabled for some compilation
953 Module-Level Inline Assembly
954 ----------------------------
956 Modules may contain "module-level inline asm" blocks, which corresponds
957 to the GCC "file scope inline asm" blocks. These blocks are internally
958 concatenated by LLVM and treated as a single unit, but may be separated
959 in the ``.ll`` file if desired. The syntax is very simple:
963 module asm "inline asm code goes here"
964 module asm "more can go here"
966 The strings can contain any character by escaping non-printable
967 characters. The escape sequence used is simply "\\xx" where "xx" is the
968 two digit hex code for the number.
970 The inline asm code is simply printed to the machine code .s file when
971 assembly code is generated.
976 A module may specify a target specific data layout string that specifies
977 how data is to be laid out in memory. The syntax for the data layout is
982 target datalayout = "layout specification"
984 The *layout specification* consists of a list of specifications
985 separated by the minus sign character ('-'). Each specification starts
986 with a letter and may include other information after the letter to
987 define some aspect of the data layout. The specifications accepted are
991 Specifies that the target lays out data in big-endian form. That is,
992 the bits with the most significance have the lowest address
995 Specifies that the target lays out data in little-endian form. That
996 is, the bits with the least significance have the lowest address
999 Specifies the natural alignment of the stack in bits. Alignment
1000 promotion of stack variables is limited to the natural stack
1001 alignment to avoid dynamic stack realignment. The stack alignment
1002 must be a multiple of 8-bits. If omitted, the natural stack
1003 alignment defaults to "unspecified", which does not prevent any
1004 alignment promotions.
1005 ``p[n]:<size>:<abi>:<pref>``
1006 This specifies the *size* of a pointer and its ``<abi>`` and
1007 ``<pref>``\erred alignments for address space ``n``. All sizes are in
1008 bits. Specifying the ``<pref>`` alignment is optional. If omitted, the
1009 preceding ``:`` should be omitted too. The address space, ``n`` is
1010 optional, and if not specified, denotes the default address space 0.
1011 The value of ``n`` must be in the range [1,2^23).
1012 ``i<size>:<abi>:<pref>``
1013 This specifies the alignment for an integer type of a given bit
1014 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
1015 ``v<size>:<abi>:<pref>``
1016 This specifies the alignment for a vector type of a given bit
1018 ``f<size>:<abi>:<pref>``
1019 This specifies the alignment for a floating point type of a given bit
1020 ``<size>``. Only values of ``<size>`` that are supported by the target
1021 will work. 32 (float) and 64 (double) are supported on all targets; 80
1022 or 128 (different flavors of long double) are also supported on some
1024 ``a<size>:<abi>:<pref>``
1025 This specifies the alignment for an aggregate type of a given bit
1027 ``s<size>:<abi>:<pref>``
1028 This specifies the alignment for a stack object of a given bit
1030 ``n<size1>:<size2>:<size3>...``
1031 This specifies a set of native integer widths for the target CPU in
1032 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1033 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1034 this set are considered to support most general arithmetic operations
1037 When constructing the data layout for a given target, LLVM starts with a
1038 default set of specifications which are then (possibly) overridden by
1039 the specifications in the ``datalayout`` keyword. The default
1040 specifications are given in this list:
1042 - ``E`` - big endian
1043 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment
1044 - ``S0`` - natural stack alignment is unspecified
1045 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1046 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1047 - ``i16:16:16`` - i16 is 16-bit aligned
1048 - ``i32:32:32`` - i32 is 32-bit aligned
1049 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1050 alignment of 64-bits
1051 - ``f16:16:16`` - half is 16-bit aligned
1052 - ``f32:32:32`` - float is 32-bit aligned
1053 - ``f64:64:64`` - double is 64-bit aligned
1054 - ``f128:128:128`` - quad is 128-bit aligned
1055 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1056 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1057 - ``a0:0:64`` - aggregates are 64-bit aligned
1059 When LLVM is determining the alignment for a given type, it uses the
1062 #. If the type sought is an exact match for one of the specifications,
1063 that specification is used.
1064 #. If no match is found, and the type sought is an integer type, then
1065 the smallest integer type that is larger than the bitwidth of the
1066 sought type is used. If none of the specifications are larger than
1067 the bitwidth then the largest integer type is used. For example,
1068 given the default specifications above, the i7 type will use the
1069 alignment of i8 (next largest) while both i65 and i256 will use the
1070 alignment of i64 (largest specified).
1071 #. If no match is found, and the type sought is a vector type, then the
1072 largest vector type that is smaller than the sought vector type will
1073 be used as a fall back. This happens because <128 x double> can be
1074 implemented in terms of 64 <2 x double>, for example.
1076 The function of the data layout string may not be what you expect.
1077 Notably, this is not a specification from the frontend of what alignment
1078 the code generator should use.
1080 Instead, if specified, the target data layout is required to match what
1081 the ultimate *code generator* expects. This string is used by the
1082 mid-level optimizers to improve code, and this only works if it matches
1083 what the ultimate code generator uses. If you would like to generate IR
1084 that does not embed this target-specific detail into the IR, then you
1085 don't have to specify the string. This will disable some optimizations
1086 that require precise layout information, but this also prevents those
1087 optimizations from introducing target specificity into the IR.
1089 .. _pointeraliasing:
1091 Pointer Aliasing Rules
1092 ----------------------
1094 Any memory access must be done through a pointer value associated with
1095 an address range of the memory access, otherwise the behavior is
1096 undefined. Pointer values are associated with address ranges according
1097 to the following rules:
1099 - A pointer value is associated with the addresses associated with any
1100 value it is *based* on.
1101 - An address of a global variable is associated with the address range
1102 of the variable's storage.
1103 - The result value of an allocation instruction is associated with the
1104 address range of the allocated storage.
1105 - A null pointer in the default address-space is associated with no
1107 - An integer constant other than zero or a pointer value returned from
1108 a function not defined within LLVM may be associated with address
1109 ranges allocated through mechanisms other than those provided by
1110 LLVM. Such ranges shall not overlap with any ranges of addresses
1111 allocated by mechanisms provided by LLVM.
1113 A pointer value is *based* on another pointer value according to the
1116 - A pointer value formed from a ``getelementptr`` operation is *based*
1117 on the first operand of the ``getelementptr``.
1118 - The result value of a ``bitcast`` is *based* on the operand of the
1120 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1121 values that contribute (directly or indirectly) to the computation of
1122 the pointer's value.
1123 - The "*based* on" relationship is transitive.
1125 Note that this definition of *"based"* is intentionally similar to the
1126 definition of *"based"* in C99, though it is slightly weaker.
1128 LLVM IR does not associate types with memory. The result type of a
1129 ``load`` merely indicates the size and alignment of the memory from
1130 which to load, as well as the interpretation of the value. The first
1131 operand type of a ``store`` similarly only indicates the size and
1132 alignment of the store.
1134 Consequently, type-based alias analysis, aka TBAA, aka
1135 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1136 :ref:`Metadata <metadata>` may be used to encode additional information
1137 which specialized optimization passes may use to implement type-based
1142 Volatile Memory Accesses
1143 ------------------------
1145 Certain memory accesses, such as :ref:`load <i_load>`'s,
1146 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1147 marked ``volatile``. The optimizers must not change the number of
1148 volatile operations or change their order of execution relative to other
1149 volatile operations. The optimizers *may* change the order of volatile
1150 operations relative to non-volatile operations. This is not Java's
1151 "volatile" and has no cross-thread synchronization behavior.
1153 IR-level volatile loads and stores cannot safely be optimized into
1154 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1155 flagged volatile. Likewise, the backend should never split or merge
1156 target-legal volatile load/store instructions.
1158 .. admonition:: Rationale
1160 Platforms may rely on volatile loads and stores of natively supported
1161 data width to be executed as single instruction. For example, in C
1162 this holds for an l-value of volatile primitive type with native
1163 hardware support, but not necessarily for aggregate types. The
1164 frontend upholds these expectations, which are intentionally
1165 unspecified in the IR. The rules above ensure that IR transformation
1166 do not violate the frontend's contract with the language.
1170 Memory Model for Concurrent Operations
1171 --------------------------------------
1173 The LLVM IR does not define any way to start parallel threads of
1174 execution or to register signal handlers. Nonetheless, there are
1175 platform-specific ways to create them, and we define LLVM IR's behavior
1176 in their presence. This model is inspired by the C++0x memory model.
1178 For a more informal introduction to this model, see the :doc:`Atomics`.
1180 We define a *happens-before* partial order as the least partial order
1183 - Is a superset of single-thread program order, and
1184 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1185 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1186 techniques, like pthread locks, thread creation, thread joining,
1187 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1188 Constraints <ordering>`).
1190 Note that program order does not introduce *happens-before* edges
1191 between a thread and signals executing inside that thread.
1193 Every (defined) read operation (load instructions, memcpy, atomic
1194 loads/read-modify-writes, etc.) R reads a series of bytes written by
1195 (defined) write operations (store instructions, atomic
1196 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1197 section, initialized globals are considered to have a write of the
1198 initializer which is atomic and happens before any other read or write
1199 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1200 may see any write to the same byte, except:
1202 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1203 write\ :sub:`2` happens before R\ :sub:`byte`, then
1204 R\ :sub:`byte` does not see write\ :sub:`1`.
1205 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1206 R\ :sub:`byte` does not see write\ :sub:`3`.
1208 Given that definition, R\ :sub:`byte` is defined as follows:
1210 - If R is volatile, the result is target-dependent. (Volatile is
1211 supposed to give guarantees which can support ``sig_atomic_t`` in
1212 C/C++, and may be used for accesses to addresses which do not behave
1213 like normal memory. It does not generally provide cross-thread
1215 - Otherwise, if there is no write to the same byte that happens before
1216 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1217 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1218 R\ :sub:`byte` returns the value written by that write.
1219 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1220 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1221 Memory Ordering Constraints <ordering>` section for additional
1222 constraints on how the choice is made.
1223 - Otherwise R\ :sub:`byte` returns ``undef``.
1225 R returns the value composed of the series of bytes it read. This
1226 implies that some bytes within the value may be ``undef`` **without**
1227 the entire value being ``undef``. Note that this only defines the
1228 semantics of the operation; it doesn't mean that targets will emit more
1229 than one instruction to read the series of bytes.
1231 Note that in cases where none of the atomic intrinsics are used, this
1232 model places only one restriction on IR transformations on top of what
1233 is required for single-threaded execution: introducing a store to a byte
1234 which might not otherwise be stored is not allowed in general.
1235 (Specifically, in the case where another thread might write to and read
1236 from an address, introducing a store can change a load that may see
1237 exactly one write into a load that may see multiple writes.)
1241 Atomic Memory Ordering Constraints
1242 ----------------------------------
1244 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1245 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1246 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1247 an ordering parameter that determines which other atomic instructions on
1248 the same address they *synchronize with*. These semantics are borrowed
1249 from Java and C++0x, but are somewhat more colloquial. If these
1250 descriptions aren't precise enough, check those specs (see spec
1251 references in the :doc:`atomics guide <Atomics>`).
1252 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1253 differently since they don't take an address. See that instruction's
1254 documentation for details.
1256 For a simpler introduction to the ordering constraints, see the
1260 The set of values that can be read is governed by the happens-before
1261 partial order. A value cannot be read unless some operation wrote
1262 it. This is intended to provide a guarantee strong enough to model
1263 Java's non-volatile shared variables. This ordering cannot be
1264 specified for read-modify-write operations; it is not strong enough
1265 to make them atomic in any interesting way.
1267 In addition to the guarantees of ``unordered``, there is a single
1268 total order for modifications by ``monotonic`` operations on each
1269 address. All modification orders must be compatible with the
1270 happens-before order. There is no guarantee that the modification
1271 orders can be combined to a global total order for the whole program
1272 (and this often will not be possible). The read in an atomic
1273 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1274 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1275 order immediately before the value it writes. If one atomic read
1276 happens before another atomic read of the same address, the later
1277 read must see the same value or a later value in the address's
1278 modification order. This disallows reordering of ``monotonic`` (or
1279 stronger) operations on the same address. If an address is written
1280 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1281 read that address repeatedly, the other threads must eventually see
1282 the write. This corresponds to the C++0x/C1x
1283 ``memory_order_relaxed``.
1285 In addition to the guarantees of ``monotonic``, a
1286 *synchronizes-with* edge may be formed with a ``release`` operation.
1287 This is intended to model C++'s ``memory_order_acquire``.
1289 In addition to the guarantees of ``monotonic``, if this operation
1290 writes a value which is subsequently read by an ``acquire``
1291 operation, it *synchronizes-with* that operation. (This isn't a
1292 complete description; see the C++0x definition of a release
1293 sequence.) This corresponds to the C++0x/C1x
1294 ``memory_order_release``.
1295 ``acq_rel`` (acquire+release)
1296 Acts as both an ``acquire`` and ``release`` operation on its
1297 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1298 ``seq_cst`` (sequentially consistent)
1299 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1300 operation which only reads, ``release`` for an operation which only
1301 writes), there is a global total order on all
1302 sequentially-consistent operations on all addresses, which is
1303 consistent with the *happens-before* partial order and with the
1304 modification orders of all the affected addresses. Each
1305 sequentially-consistent read sees the last preceding write to the
1306 same address in this global order. This corresponds to the C++0x/C1x
1307 ``memory_order_seq_cst`` and Java volatile.
1311 If an atomic operation is marked ``singlethread``, it only *synchronizes
1312 with* or participates in modification and seq\_cst total orderings with
1313 other operations running in the same thread (for example, in signal
1321 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1322 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1323 :ref:`frem <i_frem>`) have the following flags that can set to enable
1324 otherwise unsafe floating point operations
1327 No NaNs - Allow optimizations to assume the arguments and result are not
1328 NaN. Such optimizations are required to retain defined behavior over
1329 NaNs, but the value of the result is undefined.
1332 No Infs - Allow optimizations to assume the arguments and result are not
1333 +/-Inf. Such optimizations are required to retain defined behavior over
1334 +/-Inf, but the value of the result is undefined.
1337 No Signed Zeros - Allow optimizations to treat the sign of a zero
1338 argument or result as insignificant.
1341 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1342 argument rather than perform division.
1345 Fast - Allow algebraically equivalent transformations that may
1346 dramatically change results in floating point (e.g. reassociate). This
1347 flag implies all the others.
1354 The LLVM type system is one of the most important features of the
1355 intermediate representation. Being typed enables a number of
1356 optimizations to be performed on the intermediate representation
1357 directly, without having to do extra analyses on the side before the
1358 transformation. A strong type system makes it easier to read the
1359 generated code and enables novel analyses and transformations that are
1360 not feasible to perform on normal three address code representations.
1362 Type Classifications
1363 --------------------
1365 The types fall into a few useful classifications:
1374 * - :ref:`integer <t_integer>`
1375 - ``i1``, ``i2``, ``i3``, ... ``i8``, ... ``i16``, ... ``i32``, ...
1378 * - :ref:`floating point <t_floating>`
1379 - ``half``, ``float``, ``double``, ``x86_fp80``, ``fp128``,
1387 - :ref:`integer <t_integer>`, :ref:`floating point <t_floating>`,
1388 :ref:`pointer <t_pointer>`, :ref:`vector <t_vector>`,
1389 :ref:`structure <t_struct>`, :ref:`array <t_array>`,
1390 :ref:`label <t_label>`, :ref:`metadata <t_metadata>`.
1392 * - :ref:`primitive <t_primitive>`
1393 - :ref:`label <t_label>`,
1394 :ref:`void <t_void>`,
1395 :ref:`integer <t_integer>`,
1396 :ref:`floating point <t_floating>`,
1397 :ref:`x86mmx <t_x86mmx>`,
1398 :ref:`metadata <t_metadata>`.
1400 * - :ref:`derived <t_derived>`
1401 - :ref:`array <t_array>`,
1402 :ref:`function <t_function>`,
1403 :ref:`pointer <t_pointer>`,
1404 :ref:`structure <t_struct>`,
1405 :ref:`vector <t_vector>`,
1406 :ref:`opaque <t_opaque>`.
1408 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1409 Values of these types are the only ones which can be produced by
1417 The primitive types are the fundamental building blocks of the LLVM
1428 The integer type is a very simple type that simply specifies an
1429 arbitrary bit width for the integer type desired. Any bit width from 1
1430 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1439 The number of bits the integer will occupy is specified by the ``N``
1445 +----------------+------------------------------------------------+
1446 | ``i1`` | a single-bit integer. |
1447 +----------------+------------------------------------------------+
1448 | ``i32`` | a 32-bit integer. |
1449 +----------------+------------------------------------------------+
1450 | ``i1942652`` | a really big integer of over 1 million bits. |
1451 +----------------+------------------------------------------------+
1455 Floating Point Types
1456 ^^^^^^^^^^^^^^^^^^^^
1465 - 16-bit floating point value
1468 - 32-bit floating point value
1471 - 64-bit floating point value
1474 - 128-bit floating point value (112-bit mantissa)
1477 - 80-bit floating point value (X87)
1480 - 128-bit floating point value (two 64-bits)
1490 The x86mmx type represents a value held in an MMX register on an x86
1491 machine. The operations allowed on it are quite limited: parameters and
1492 return values, load and store, and bitcast. User-specified MMX
1493 instructions are represented as intrinsic or asm calls with arguments
1494 and/or results of this type. There are no arrays, vectors or constants
1512 The void type does not represent any value and has no size.
1529 The label type represents code labels.
1546 The metadata type represents embedded metadata. No derived types may be
1547 created from metadata except for :ref:`function <t_function>` arguments.
1561 The real power in LLVM comes from the derived types in the system. This
1562 is what allows a programmer to represent arrays, functions, pointers,
1563 and other useful types. Each of these types contain one or more element
1564 types which may be a primitive type, or another derived type. For
1565 example, it is possible to have a two dimensional array, using an array
1566 as the element type of another array.
1573 Aggregate Types are a subset of derived types that can contain multiple
1574 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
1575 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
1586 The array type is a very simple derived type that arranges elements
1587 sequentially in memory. The array type requires a size (number of
1588 elements) and an underlying data type.
1595 [<# elements> x <elementtype>]
1597 The number of elements is a constant integer value; ``elementtype`` may
1598 be any type with a size.
1603 +------------------+--------------------------------------+
1604 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
1605 +------------------+--------------------------------------+
1606 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
1607 +------------------+--------------------------------------+
1608 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
1609 +------------------+--------------------------------------+
1611 Here are some examples of multidimensional arrays:
1613 +-----------------------------+----------------------------------------------------------+
1614 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
1615 +-----------------------------+----------------------------------------------------------+
1616 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
1617 +-----------------------------+----------------------------------------------------------+
1618 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
1619 +-----------------------------+----------------------------------------------------------+
1621 There is no restriction on indexing beyond the end of the array implied
1622 by a static type (though there are restrictions on indexing beyond the
1623 bounds of an allocated object in some cases). This means that
1624 single-dimension 'variable sized array' addressing can be implemented in
1625 LLVM with a zero length array type. An implementation of 'pascal style
1626 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
1637 The function type can be thought of as a function signature. It consists
1638 of a return type and a list of formal parameter types. The return type
1639 of a function type is a first class type or a void type.
1646 <returntype> (<parameter list>)
1648 ...where '``<parameter list>``' is a comma-separated list of type
1649 specifiers. Optionally, the parameter list may include a type ``...``,
1650 which indicates that the function takes a variable number of arguments.
1651 Variable argument functions can access their arguments with the
1652 :ref:`variable argument handling intrinsic <int_varargs>` functions.
1653 '``<returntype>``' is any type except :ref:`label <t_label>`.
1658 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1659 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1660 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1661 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1662 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1663 | ``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. |
1664 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1665 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1666 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1676 The structure type is used to represent a collection of data members
1677 together in memory. The elements of a structure may be any type that has
1680 Structures in memory are accessed using '``load``' and '``store``' by
1681 getting a pointer to a field with the '``getelementptr``' instruction.
1682 Structures in registers are accessed using the '``extractvalue``' and
1683 '``insertvalue``' instructions.
1685 Structures may optionally be "packed" structures, which indicate that
1686 the alignment of the struct is one byte, and that there is no padding
1687 between the elements. In non-packed structs, padding between field types
1688 is inserted as defined by the DataLayout string in the module, which is
1689 required to match what the underlying code generator expects.
1691 Structures can either be "literal" or "identified". A literal structure
1692 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
1693 identified types are always defined at the top level with a name.
1694 Literal types are uniqued by their contents and can never be recursive
1695 or opaque since there is no way to write one. Identified types can be
1696 recursive, can be opaqued, and are never uniqued.
1703 %T1 = type { <type list> } ; Identified normal struct type
1704 %T2 = type <{ <type list> }> ; Identified packed struct type
1709 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1710 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
1711 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1712 | ``{ 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``. |
1713 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1714 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
1715 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1719 Opaque Structure Types
1720 ^^^^^^^^^^^^^^^^^^^^^^
1725 Opaque structure types are used to represent named structure types that
1726 do not have a body specified. This corresponds (for example) to the C
1727 notion of a forward declared structure.
1740 +--------------+-------------------+
1741 | ``opaque`` | An opaque type. |
1742 +--------------+-------------------+
1752 The pointer type is used to specify memory locations. Pointers are
1753 commonly used to reference objects in memory.
1755 Pointer types may have an optional address space attribute defining the
1756 numbered address space where the pointed-to object resides. The default
1757 address space is number zero. The semantics of non-zero address spaces
1758 are target-specific.
1760 Note that LLVM does not permit pointers to void (``void*``) nor does it
1761 permit pointers to labels (``label*``). Use ``i8*`` instead.
1773 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1774 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
1775 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1776 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
1777 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1778 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
1779 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1789 A vector type is a simple derived type that represents a vector of
1790 elements. Vector types are used when multiple primitive data are
1791 operated in parallel using a single instruction (SIMD). A vector type
1792 requires a size (number of elements) and an underlying primitive data
1793 type. Vector types are considered :ref:`first class <t_firstclass>`.
1800 < <# elements> x <elementtype> >
1802 The number of elements is a constant integer value larger than 0;
1803 elementtype may be any integer or floating point type, or a pointer to
1804 these types. Vectors of size zero are not allowed.
1809 +-------------------+--------------------------------------------------+
1810 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
1811 +-------------------+--------------------------------------------------+
1812 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
1813 +-------------------+--------------------------------------------------+
1814 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
1815 +-------------------+--------------------------------------------------+
1816 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
1817 +-------------------+--------------------------------------------------+
1822 LLVM has several different basic types of constants. This section
1823 describes them all and their syntax.
1828 **Boolean constants**
1829 The two strings '``true``' and '``false``' are both valid constants
1831 **Integer constants**
1832 Standard integers (such as '4') are constants of the
1833 :ref:`integer <t_integer>` type. Negative numbers may be used with
1835 **Floating point constants**
1836 Floating point constants use standard decimal notation (e.g.
1837 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
1838 hexadecimal notation (see below). The assembler requires the exact
1839 decimal value of a floating-point constant. For example, the
1840 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
1841 decimal in binary. Floating point constants must have a :ref:`floating
1842 point <t_floating>` type.
1843 **Null pointer constants**
1844 The identifier '``null``' is recognized as a null pointer constant
1845 and must be of :ref:`pointer type <t_pointer>`.
1847 The one non-intuitive notation for constants is the hexadecimal form of
1848 floating point constants. For example, the form
1849 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
1850 than) '``double 4.5e+15``'. The only time hexadecimal floating point
1851 constants are required (and the only time that they are generated by the
1852 disassembler) is when a floating point constant must be emitted but it
1853 cannot be represented as a decimal floating point number in a reasonable
1854 number of digits. For example, NaN's, infinities, and other special
1855 values are represented in their IEEE hexadecimal format so that assembly
1856 and disassembly do not cause any bits to change in the constants.
1858 When using the hexadecimal form, constants of types half, float, and
1859 double are represented using the 16-digit form shown above (which
1860 matches the IEEE754 representation for double); half and float values
1861 must, however, be exactly representable as IEEE 754 half and single
1862 precision, respectively. Hexadecimal format is always used for long
1863 double, and there are three forms of long double. The 80-bit format used
1864 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
1865 128-bit format used by PowerPC (two adjacent doubles) is represented by
1866 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
1867 represented by ``0xL`` followed by 32 hexadecimal digits. Long doubles
1868 will only work if they match the long double format on your target.
1869 The IEEE 16-bit format (half precision) is represented by ``0xH``
1870 followed by 4 hexadecimal digits. All hexadecimal formats are big-endian
1871 (sign bit at the left).
1873 There are no constants of type x86mmx.
1878 Complex constants are a (potentially recursive) combination of simple
1879 constants and smaller complex constants.
1881 **Structure constants**
1882 Structure constants are represented with notation similar to
1883 structure type definitions (a comma separated list of elements,
1884 surrounded by braces (``{}``)). For example:
1885 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
1886 "``@G = external global i32``". Structure constants must have
1887 :ref:`structure type <t_struct>`, and the number and types of elements
1888 must match those specified by the type.
1890 Array constants are represented with notation similar to array type
1891 definitions (a comma separated list of elements, surrounded by
1892 square brackets (``[]``)). For example:
1893 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
1894 :ref:`array type <t_array>`, and the number and types of elements must
1895 match those specified by the type.
1896 **Vector constants**
1897 Vector constants are represented with notation similar to vector
1898 type definitions (a comma separated list of elements, surrounded by
1899 less-than/greater-than's (``<>``)). For example:
1900 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
1901 must have :ref:`vector type <t_vector>`, and the number and types of
1902 elements must match those specified by the type.
1903 **Zero initialization**
1904 The string '``zeroinitializer``' can be used to zero initialize a
1905 value to zero of *any* type, including scalar and
1906 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
1907 having to print large zero initializers (e.g. for large arrays) and
1908 is always exactly equivalent to using explicit zero initializers.
1910 A metadata node is a structure-like constant with :ref:`metadata
1911 type <t_metadata>`. For example:
1912 "``metadata !{ i32 0, metadata !"test" }``". Unlike other
1913 constants that are meant to be interpreted as part of the
1914 instruction stream, metadata is a place to attach additional
1915 information such as debug info.
1917 Global Variable and Function Addresses
1918 --------------------------------------
1920 The addresses of :ref:`global variables <globalvars>` and
1921 :ref:`functions <functionstructure>` are always implicitly valid
1922 (link-time) constants. These constants are explicitly referenced when
1923 the :ref:`identifier for the global <identifiers>` is used and always have
1924 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
1927 .. code-block:: llvm
1931 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1938 The string '``undef``' can be used anywhere a constant is expected, and
1939 indicates that the user of the value may receive an unspecified
1940 bit-pattern. Undefined values may be of any type (other than '``label``'
1941 or '``void``') and be used anywhere a constant is permitted.
1943 Undefined values are useful because they indicate to the compiler that
1944 the program is well defined no matter what value is used. This gives the
1945 compiler more freedom to optimize. Here are some examples of
1946 (potentially surprising) transformations that are valid (in pseudo IR):
1948 .. code-block:: llvm
1958 This is safe because all of the output bits are affected by the undef
1959 bits. Any output bit can have a zero or one depending on the input bits.
1961 .. code-block:: llvm
1972 These logical operations have bits that are not always affected by the
1973 input. For example, if ``%X`` has a zero bit, then the output of the
1974 '``and``' operation will always be a zero for that bit, no matter what
1975 the corresponding bit from the '``undef``' is. As such, it is unsafe to
1976 optimize or assume that the result of the '``and``' is '``undef``'.
1977 However, it is safe to assume that all bits of the '``undef``' could be
1978 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
1979 all the bits of the '``undef``' operand to the '``or``' could be set,
1980 allowing the '``or``' to be folded to -1.
1982 .. code-block:: llvm
1984 %A = select undef, %X, %Y
1985 %B = select undef, 42, %Y
1986 %C = select %X, %Y, undef
1996 This set of examples shows that undefined '``select``' (and conditional
1997 branch) conditions can go *either way*, but they have to come from one
1998 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
1999 both known to have a clear low bit, then ``%A`` would have to have a
2000 cleared low bit. However, in the ``%C`` example, the optimizer is
2001 allowed to assume that the '``undef``' operand could be the same as
2002 ``%Y``, allowing the whole '``select``' to be eliminated.
2004 .. code-block:: llvm
2006 %A = xor undef, undef
2023 This example points out that two '``undef``' operands are not
2024 necessarily the same. This can be surprising to people (and also matches
2025 C semantics) where they assume that "``X^X``" is always zero, even if
2026 ``X`` is undefined. This isn't true for a number of reasons, but the
2027 short answer is that an '``undef``' "variable" can arbitrarily change
2028 its value over its "live range". This is true because the variable
2029 doesn't actually *have a live range*. Instead, the value is logically
2030 read from arbitrary registers that happen to be around when needed, so
2031 the value is not necessarily consistent over time. In fact, ``%A`` and
2032 ``%C`` need to have the same semantics or the core LLVM "replace all
2033 uses with" concept would not hold.
2035 .. code-block:: llvm
2043 These examples show the crucial difference between an *undefined value*
2044 and *undefined behavior*. An undefined value (like '``undef``') is
2045 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2046 operation can be constant folded to '``undef``', because the '``undef``'
2047 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2048 However, in the second example, we can make a more aggressive
2049 assumption: because the ``undef`` is allowed to be an arbitrary value,
2050 we are allowed to assume that it could be zero. Since a divide by zero
2051 has *undefined behavior*, we are allowed to assume that the operation
2052 does not execute at all. This allows us to delete the divide and all
2053 code after it. Because the undefined operation "can't happen", the
2054 optimizer can assume that it occurs in dead code.
2056 .. code-block:: llvm
2058 a: store undef -> %X
2059 b: store %X -> undef
2064 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2065 value can be assumed to not have any effect; we can assume that the
2066 value is overwritten with bits that happen to match what was already
2067 there. However, a store *to* an undefined location could clobber
2068 arbitrary memory, therefore, it has undefined behavior.
2075 Poison values are similar to :ref:`undef values <undefvalues>`, however
2076 they also represent the fact that an instruction or constant expression
2077 which cannot evoke side effects has nevertheless detected a condition
2078 which results in undefined behavior.
2080 There is currently no way of representing a poison value in the IR; they
2081 only exist when produced by operations such as :ref:`add <i_add>` with
2084 Poison value behavior is defined in terms of value *dependence*:
2086 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2087 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2088 their dynamic predecessor basic block.
2089 - Function arguments depend on the corresponding actual argument values
2090 in the dynamic callers of their functions.
2091 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2092 instructions that dynamically transfer control back to them.
2093 - :ref:`Invoke <i_invoke>` instructions depend on the
2094 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2095 call instructions that dynamically transfer control back to them.
2096 - Non-volatile loads and stores depend on the most recent stores to all
2097 of the referenced memory addresses, following the order in the IR
2098 (including loads and stores implied by intrinsics such as
2099 :ref:`@llvm.memcpy <int_memcpy>`.)
2100 - An instruction with externally visible side effects depends on the
2101 most recent preceding instruction with externally visible side
2102 effects, following the order in the IR. (This includes :ref:`volatile
2103 operations <volatile>`.)
2104 - An instruction *control-depends* on a :ref:`terminator
2105 instruction <terminators>` if the terminator instruction has
2106 multiple successors and the instruction is always executed when
2107 control transfers to one of the successors, and may not be executed
2108 when control is transferred to another.
2109 - Additionally, an instruction also *control-depends* on a terminator
2110 instruction if the set of instructions it otherwise depends on would
2111 be different if the terminator had transferred control to a different
2113 - Dependence is transitive.
2115 Poison Values have the same behavior as :ref:`undef values <undefvalues>`,
2116 with the additional affect that any instruction which has a *dependence*
2117 on a poison value has undefined behavior.
2119 Here are some examples:
2121 .. code-block:: llvm
2124 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2125 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2126 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
2127 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2129 store i32 %poison, i32* @g ; Poison value stored to memory.
2130 %poison2 = load i32* @g ; Poison value loaded back from memory.
2132 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2134 %narrowaddr = bitcast i32* @g to i16*
2135 %wideaddr = bitcast i32* @g to i64*
2136 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2137 %poison4 = load i64* %wideaddr ; Returns a poison value.
2139 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2140 br i1 %cmp, label %true, label %end ; Branch to either destination.
2143 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2144 ; it has undefined behavior.
2148 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2149 ; Both edges into this PHI are
2150 ; control-dependent on %cmp, so this
2151 ; always results in a poison value.
2153 store volatile i32 0, i32* @g ; This would depend on the store in %true
2154 ; if %cmp is true, or the store in %entry
2155 ; otherwise, so this is undefined behavior.
2157 br i1 %cmp, label %second_true, label %second_end
2158 ; The same branch again, but this time the
2159 ; true block doesn't have side effects.
2166 store volatile i32 0, i32* @g ; This time, the instruction always depends
2167 ; on the store in %end. Also, it is
2168 ; control-equivalent to %end, so this is
2169 ; well-defined (ignoring earlier undefined
2170 ; behavior in this example).
2174 Addresses of Basic Blocks
2175 -------------------------
2177 ``blockaddress(@function, %block)``
2179 The '``blockaddress``' constant computes the address of the specified
2180 basic block in the specified function, and always has an ``i8*`` type.
2181 Taking the address of the entry block is illegal.
2183 This value only has defined behavior when used as an operand to the
2184 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2185 against null. Pointer equality tests between labels addresses results in
2186 undefined behavior --- though, again, comparison against null is ok, and
2187 no label is equal to the null pointer. This may be passed around as an
2188 opaque pointer sized value as long as the bits are not inspected. This
2189 allows ``ptrtoint`` and arithmetic to be performed on these values so
2190 long as the original value is reconstituted before the ``indirectbr``
2193 Finally, some targets may provide defined semantics when using the value
2194 as the operand to an inline assembly, but that is target specific.
2196 Constant Expressions
2197 --------------------
2199 Constant expressions are used to allow expressions involving other
2200 constants to be used as constants. Constant expressions may be of any
2201 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2202 that does not have side effects (e.g. load and call are not supported).
2203 The following is the syntax for constant expressions:
2205 ``trunc (CST to TYPE)``
2206 Truncate a constant to another type. The bit size of CST must be
2207 larger than the bit size of TYPE. Both types must be integers.
2208 ``zext (CST to TYPE)``
2209 Zero extend a constant to another type. The bit size of CST must be
2210 smaller than the bit size of TYPE. Both types must be integers.
2211 ``sext (CST to TYPE)``
2212 Sign extend a constant to another type. The bit size of CST must be
2213 smaller than the bit size of TYPE. Both types must be integers.
2214 ``fptrunc (CST to TYPE)``
2215 Truncate a floating point constant to another floating point type.
2216 The size of CST must be larger than the size of TYPE. Both types
2217 must be floating point.
2218 ``fpext (CST to TYPE)``
2219 Floating point extend a constant to another type. The size of CST
2220 must be smaller or equal to the size of TYPE. Both types must be
2222 ``fptoui (CST to TYPE)``
2223 Convert a floating point constant to the corresponding unsigned
2224 integer constant. TYPE must be a scalar or vector integer type. CST
2225 must be of scalar or vector floating point type. Both CST and TYPE
2226 must be scalars, or vectors of the same number of elements. If the
2227 value won't fit in the integer type, the results are undefined.
2228 ``fptosi (CST to TYPE)``
2229 Convert a floating point constant to the corresponding signed
2230 integer constant. TYPE must be a scalar or vector integer type. CST
2231 must be of scalar or vector floating point type. Both CST and TYPE
2232 must be scalars, or vectors of the same number of elements. If the
2233 value won't fit in the integer type, the results are undefined.
2234 ``uitofp (CST to TYPE)``
2235 Convert an unsigned integer constant to the corresponding floating
2236 point constant. TYPE must be a scalar or vector floating point type.
2237 CST must be of scalar or vector integer type. Both CST and TYPE must
2238 be scalars, or vectors of the same number of elements. If the value
2239 won't fit in the floating point type, the results are undefined.
2240 ``sitofp (CST to TYPE)``
2241 Convert a signed integer constant to the corresponding floating
2242 point constant. TYPE must be a scalar or vector floating point type.
2243 CST must be of scalar or vector integer type. Both CST and TYPE must
2244 be scalars, or vectors of the same number of elements. If the value
2245 won't fit in the floating point type, the results are undefined.
2246 ``ptrtoint (CST to TYPE)``
2247 Convert a pointer typed constant to the corresponding integer
2248 constant. ``TYPE`` must be an integer type. ``CST`` must be of
2249 pointer type. The ``CST`` value is zero extended, truncated, or
2250 unchanged to make it fit in ``TYPE``.
2251 ``inttoptr (CST to TYPE)``
2252 Convert an integer constant to a pointer constant. TYPE must be a
2253 pointer type. CST must be of integer type. The CST value is zero
2254 extended, truncated, or unchanged to make it fit in a pointer size.
2255 This one is *really* dangerous!
2256 ``bitcast (CST to TYPE)``
2257 Convert a constant, CST, to another TYPE. The constraints of the
2258 operands are the same as those for the :ref:`bitcast
2259 instruction <i_bitcast>`.
2260 ``getelementptr (CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)``
2261 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2262 constants. As with the :ref:`getelementptr <i_getelementptr>`
2263 instruction, the index list may have zero or more indexes, which are
2264 required to make sense for the type of "CSTPTR".
2265 ``select (COND, VAL1, VAL2)``
2266 Perform the :ref:`select operation <i_select>` on constants.
2267 ``icmp COND (VAL1, VAL2)``
2268 Performs the :ref:`icmp operation <i_icmp>` on constants.
2269 ``fcmp COND (VAL1, VAL2)``
2270 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2271 ``extractelement (VAL, IDX)``
2272 Perform the :ref:`extractelement operation <i_extractelement>` on
2274 ``insertelement (VAL, ELT, IDX)``
2275 Perform the :ref:`insertelement operation <i_insertelement>` on
2277 ``shufflevector (VEC1, VEC2, IDXMASK)``
2278 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2280 ``extractvalue (VAL, IDX0, IDX1, ...)``
2281 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2282 constants. The index list is interpreted in a similar manner as
2283 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2284 least one index value must be specified.
2285 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2286 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2287 The index list is interpreted in a similar manner as indices in a
2288 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2289 value must be specified.
2290 ``OPCODE (LHS, RHS)``
2291 Perform the specified operation of the LHS and RHS constants. OPCODE
2292 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2293 binary <bitwiseops>` operations. The constraints on operands are
2294 the same as those for the corresponding instruction (e.g. no bitwise
2295 operations on floating point values are allowed).
2300 Inline Assembler Expressions
2301 ----------------------------
2303 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2304 Inline Assembly <moduleasm>`) through the use of a special value. This
2305 value represents the inline assembler as a string (containing the
2306 instructions to emit), a list of operand constraints (stored as a
2307 string), a flag that indicates whether or not the inline asm expression
2308 has side effects, and a flag indicating whether the function containing
2309 the asm needs to align its stack conservatively. An example inline
2310 assembler expression is:
2312 .. code-block:: llvm
2314 i32 (i32) asm "bswap $0", "=r,r"
2316 Inline assembler expressions may **only** be used as the callee operand
2317 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2318 Thus, typically we have:
2320 .. code-block:: llvm
2322 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2324 Inline asms with side effects not visible in the constraint list must be
2325 marked as having side effects. This is done through the use of the
2326 '``sideeffect``' keyword, like so:
2328 .. code-block:: llvm
2330 call void asm sideeffect "eieio", ""()
2332 In some cases inline asms will contain code that will not work unless
2333 the stack is aligned in some way, such as calls or SSE instructions on
2334 x86, yet will not contain code that does that alignment within the asm.
2335 The compiler should make conservative assumptions about what the asm
2336 might contain and should generate its usual stack alignment code in the
2337 prologue if the '``alignstack``' keyword is present:
2339 .. code-block:: llvm
2341 call void asm alignstack "eieio", ""()
2343 Inline asms also support using non-standard assembly dialects. The
2344 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2345 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2346 the only supported dialects. An example is:
2348 .. code-block:: llvm
2350 call void asm inteldialect "eieio", ""()
2352 If multiple keywords appear the '``sideeffect``' keyword must come
2353 first, the '``alignstack``' keyword second and the '``inteldialect``'
2359 The call instructions that wrap inline asm nodes may have a
2360 "``!srcloc``" MDNode attached to it that contains a list of constant
2361 integers. If present, the code generator will use the integer as the
2362 location cookie value when report errors through the ``LLVMContext``
2363 error reporting mechanisms. This allows a front-end to correlate backend
2364 errors that occur with inline asm back to the source code that produced
2367 .. code-block:: llvm
2369 call void asm sideeffect "something bad", ""(), !srcloc !42
2371 !42 = !{ i32 1234567 }
2373 It is up to the front-end to make sense of the magic numbers it places
2374 in the IR. If the MDNode contains multiple constants, the code generator
2375 will use the one that corresponds to the line of the asm that the error
2380 Metadata Nodes and Metadata Strings
2381 -----------------------------------
2383 LLVM IR allows metadata to be attached to instructions in the program
2384 that can convey extra information about the code to the optimizers and
2385 code generator. One example application of metadata is source-level
2386 debug information. There are two metadata primitives: strings and nodes.
2387 All metadata has the ``metadata`` type and is identified in syntax by a
2388 preceding exclamation point ('``!``').
2390 A metadata string is a string surrounded by double quotes. It can
2391 contain any character by escaping non-printable characters with
2392 "``\xx``" where "``xx``" is the two digit hex code. For example:
2395 Metadata nodes are represented with notation similar to structure
2396 constants (a comma separated list of elements, surrounded by braces and
2397 preceded by an exclamation point). Metadata nodes can have any values as
2398 their operand. For example:
2400 .. code-block:: llvm
2402 !{ metadata !"test\00", i32 10}
2404 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2405 metadata nodes, which can be looked up in the module symbol table. For
2408 .. code-block:: llvm
2410 !foo = metadata !{!4, !3}
2412 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2413 function is using two metadata arguments:
2415 .. code-block:: llvm
2417 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2419 Metadata can be attached with an instruction. Here metadata ``!21`` is
2420 attached to the ``add`` instruction using the ``!dbg`` identifier:
2422 .. code-block:: llvm
2424 %indvar.next = add i64 %indvar, 1, !dbg !21
2426 More information about specific metadata nodes recognized by the
2427 optimizers and code generator is found below.
2432 In LLVM IR, memory does not have types, so LLVM's own type system is not
2433 suitable for doing TBAA. Instead, metadata is added to the IR to
2434 describe a type system of a higher level language. This can be used to
2435 implement typical C/C++ TBAA, but it can also be used to implement
2436 custom alias analysis behavior for other languages.
2438 The current metadata format is very simple. TBAA metadata nodes have up
2439 to three fields, e.g.:
2441 .. code-block:: llvm
2443 !0 = metadata !{ metadata !"an example type tree" }
2444 !1 = metadata !{ metadata !"int", metadata !0 }
2445 !2 = metadata !{ metadata !"float", metadata !0 }
2446 !3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2448 The first field is an identity field. It can be any value, usually a
2449 metadata string, which uniquely identifies the type. The most important
2450 name in the tree is the name of the root node. Two trees with different
2451 root node names are entirely disjoint, even if they have leaves with
2454 The second field identifies the type's parent node in the tree, or is
2455 null or omitted for a root node. A type is considered to alias all of
2456 its descendants and all of its ancestors in the tree. Also, a type is
2457 considered to alias all types in other trees, so that bitcode produced
2458 from multiple front-ends is handled conservatively.
2460 If the third field is present, it's an integer which if equal to 1
2461 indicates that the type is "constant" (meaning
2462 ``pointsToConstantMemory`` should return true; see `other useful
2463 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
2465 '``tbaa.struct``' Metadata
2466 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2468 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
2469 aggregate assignment operations in C and similar languages, however it
2470 is defined to copy a contiguous region of memory, which is more than
2471 strictly necessary for aggregate types which contain holes due to
2472 padding. Also, it doesn't contain any TBAA information about the fields
2475 ``!tbaa.struct`` metadata can describe which memory subregions in a
2476 memcpy are padding and what the TBAA tags of the struct are.
2478 The current metadata format is very simple. ``!tbaa.struct`` metadata
2479 nodes are a list of operands which are in conceptual groups of three.
2480 For each group of three, the first operand gives the byte offset of a
2481 field in bytes, the second gives its size in bytes, and the third gives
2484 .. code-block:: llvm
2486 !4 = metadata !{ i64 0, i64 4, metadata !1, i64 8, i64 4, metadata !2 }
2488 This describes a struct with two fields. The first is at offset 0 bytes
2489 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
2490 and has size 4 bytes and has tbaa tag !2.
2492 Note that the fields need not be contiguous. In this example, there is a
2493 4 byte gap between the two fields. This gap represents padding which
2494 does not carry useful data and need not be preserved.
2496 '``fpmath``' Metadata
2497 ^^^^^^^^^^^^^^^^^^^^^
2499 ``fpmath`` metadata may be attached to any instruction of floating point
2500 type. It can be used to express the maximum acceptable error in the
2501 result of that instruction, in ULPs, thus potentially allowing the
2502 compiler to use a more efficient but less accurate method of computing
2503 it. ULP is defined as follows:
2505 If ``x`` is a real number that lies between two finite consecutive
2506 floating-point numbers ``a`` and ``b``, without being equal to one
2507 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
2508 distance between the two non-equal finite floating-point numbers
2509 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
2511 The metadata node shall consist of a single positive floating point
2512 number representing the maximum relative error, for example:
2514 .. code-block:: llvm
2516 !0 = metadata !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
2518 '``range``' Metadata
2519 ^^^^^^^^^^^^^^^^^^^^
2521 ``range`` metadata may be attached only to loads of integer types. It
2522 expresses the possible ranges the loaded value is in. The ranges are
2523 represented with a flattened list of integers. The loaded value is known
2524 to be in the union of the ranges defined by each consecutive pair. Each
2525 pair has the following properties:
2527 - The type must match the type loaded by the instruction.
2528 - The pair ``a,b`` represents the range ``[a,b)``.
2529 - Both ``a`` and ``b`` are constants.
2530 - The range is allowed to wrap.
2531 - The range should not represent the full or empty set. That is,
2534 In addition, the pairs must be in signed order of the lower bound and
2535 they must be non-contiguous.
2539 .. code-block:: llvm
2541 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
2542 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
2543 %c = load i8* %z, align 1, !range !2 ; Can only be 0, 1, 3, 4 or 5
2544 %d = load i8* %z, align 1, !range !3 ; Can only be -2, -1, 3, 4 or 5
2546 !0 = metadata !{ i8 0, i8 2 }
2547 !1 = metadata !{ i8 255, i8 2 }
2548 !2 = metadata !{ i8 0, i8 2, i8 3, i8 6 }
2549 !3 = metadata !{ i8 -2, i8 0, i8 3, i8 6 }
2554 It is sometimes useful to attach information to loop constructs. Currently,
2555 loop metadata is implemented as metadata attached to the branch instruction
2556 in the loop latch block. This type of metadata refer to a metadata node that is
2557 guaranteed to be separate for each loop. The loop-level metadata is prefixed
2560 The loop identifier metadata is implemented using a metadata that refers to
2561 itself to avoid merging it with any other identifier metadata, e.g.,
2562 during module linkage or function inlining. That is, each loop should refer
2563 to their own identification metadata even if they reside in separate functions.
2564 The following example contains loop identifier metadata for two separate loop
2567 .. code-block:: llvm
2569 !0 = metadata !{ metadata !0 }
2570 !1 = metadata !{ metadata !1 }
2573 '``llvm.loop.parallel``' Metadata
2574 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2576 This loop metadata can be used to communicate that a loop should be considered
2577 a parallel loop. The semantics of parallel loops in this case is the one
2578 with the strongest cross-iteration instruction ordering freedom: the
2579 iterations in the loop can be considered completely independent of each
2580 other (also known as embarrassingly parallel loops).
2582 This metadata can originate from a programming language with parallel loop
2583 constructs. In such a case it is completely the programmer's responsibility
2584 to ensure the instructions from the different iterations of the loop can be
2585 executed in an arbitrary order, in parallel, or intertwined. No loop-carried
2586 dependency checking at all must be expected from the compiler.
2588 In order to fulfill the LLVM requirement for metadata to be safely ignored,
2589 it is important to ensure that a parallel loop is converted to
2590 a sequential loop in case an optimization (agnostic of the parallel loop
2591 semantics) converts the loop back to such. This happens when new memory
2592 accesses that do not fulfill the requirement of free ordering across iterations
2593 are added to the loop. Therefore, this metadata is required, but not
2594 sufficient, to consider the loop at hand a parallel loop. For a loop
2595 to be parallel, all its memory accessing instructions need to be
2596 marked with the ``llvm.mem.parallel_loop_access`` metadata that refer
2597 to the same loop identifier metadata that identify the loop at hand.
2602 Metadata types used to annotate memory accesses with information helpful
2603 for optimizations are prefixed with ``llvm.mem``.
2605 '``llvm.mem.parallel_loop_access``' Metadata
2606 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2608 For a loop to be parallel, in addition to using
2609 the ``llvm.loop.parallel`` metadata to mark the loop latch branch instruction,
2610 also all of the memory accessing instructions in the loop body need to be
2611 marked with the ``llvm.mem.parallel_loop_access`` metadata. If there
2612 is at least one memory accessing instruction not marked with the metadata,
2613 the loop, despite it possibly using the ``llvm.loop.parallel`` metadata,
2614 must be considered a sequential loop. This causes parallel loops to be
2615 converted to sequential loops due to optimization passes that are unaware of
2616 the parallel semantics and that insert new memory instructions to the loop
2619 Example of a loop that is considered parallel due to its correct use of
2620 both ``llvm.loop.parallel`` and ``llvm.mem.parallel_loop_access``
2621 metadata types that refer to the same loop identifier metadata.
2623 .. code-block:: llvm
2627 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2629 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2631 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop.parallel !0
2635 !0 = metadata !{ metadata !0 }
2637 It is also possible to have nested parallel loops. In that case the
2638 memory accesses refer to a list of loop identifier metadata nodes instead of
2639 the loop identifier metadata node directly:
2641 .. code-block:: llvm
2648 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2650 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2652 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop.parallel !1
2656 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2658 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2660 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop.parallel !2
2662 outer.for.end: ; preds = %for.body
2664 !0 = metadata !{ metadata !1, metadata !2 } ; a list of parallel loop identifiers
2665 !1 = metadata !{ metadata !1 } ; an identifier for the inner parallel loop
2666 !2 = metadata !{ metadata !2 } ; an identifier for the outer parallel loop
2669 Module Flags Metadata
2670 =====================
2672 Information about the module as a whole is difficult to convey to LLVM's
2673 subsystems. The LLVM IR isn't sufficient to transmit this information.
2674 The ``llvm.module.flags`` named metadata exists in order to facilitate
2675 this. These flags are in the form of key / value pairs --- much like a
2676 dictionary --- making it easy for any subsystem who cares about a flag to
2679 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
2680 Each triplet has the following form:
2682 - The first element is a *behavior* flag, which specifies the behavior
2683 when two (or more) modules are merged together, and it encounters two
2684 (or more) metadata with the same ID. The supported behaviors are
2686 - The second element is a metadata string that is a unique ID for the
2687 metadata. Each module may only have one flag entry for each unique ID (not
2688 including entries with the **Require** behavior).
2689 - The third element is the value of the flag.
2691 When two (or more) modules are merged together, the resulting
2692 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
2693 each unique metadata ID string, there will be exactly one entry in the merged
2694 modules ``llvm.module.flags`` metadata table, and the value for that entry will
2695 be determined by the merge behavior flag, as described below. The only exception
2696 is that entries with the *Require* behavior are always preserved.
2698 The following behaviors are supported:
2709 Emits an error if two values disagree, otherwise the resulting value
2710 is that of the operands.
2714 Emits a warning if two values disagree. The result value will be the
2715 operand for the flag from the first module being linked.
2719 Adds a requirement that another module flag be present and have a
2720 specified value after linking is performed. The value must be a
2721 metadata pair, where the first element of the pair is the ID of the
2722 module flag to be restricted, and the second element of the pair is
2723 the value the module flag should be restricted to. This behavior can
2724 be used to restrict the allowable results (via triggering of an
2725 error) of linking IDs with the **Override** behavior.
2729 Uses the specified value, regardless of the behavior or value of the
2730 other module. If both modules specify **Override**, but the values
2731 differ, an error will be emitted.
2735 Appends the two values, which are required to be metadata nodes.
2739 Appends the two values, which are required to be metadata
2740 nodes. However, duplicate entries in the second list are dropped
2741 during the append operation.
2743 It is an error for a particular unique flag ID to have multiple behaviors,
2744 except in the case of **Require** (which adds restrictions on another metadata
2745 value) or **Override**.
2747 An example of module flags:
2749 .. code-block:: llvm
2751 !0 = metadata !{ i32 1, metadata !"foo", i32 1 }
2752 !1 = metadata !{ i32 4, metadata !"bar", i32 37 }
2753 !2 = metadata !{ i32 2, metadata !"qux", i32 42 }
2754 !3 = metadata !{ i32 3, metadata !"qux",
2756 metadata !"foo", i32 1
2759 !llvm.module.flags = !{ !0, !1, !2, !3 }
2761 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
2762 if two or more ``!"foo"`` flags are seen is to emit an error if their
2763 values are not equal.
2765 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
2766 behavior if two or more ``!"bar"`` flags are seen is to use the value
2769 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
2770 behavior if two or more ``!"qux"`` flags are seen is to emit a
2771 warning if their values are not equal.
2773 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
2777 metadata !{ metadata !"foo", i32 1 }
2779 The behavior is to emit an error if the ``llvm.module.flags`` does not
2780 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
2783 Objective-C Garbage Collection Module Flags Metadata
2784 ----------------------------------------------------
2786 On the Mach-O platform, Objective-C stores metadata about garbage
2787 collection in a special section called "image info". The metadata
2788 consists of a version number and a bitmask specifying what types of
2789 garbage collection are supported (if any) by the file. If two or more
2790 modules are linked together their garbage collection metadata needs to
2791 be merged rather than appended together.
2793 The Objective-C garbage collection module flags metadata consists of the
2794 following key-value pairs:
2803 * - ``Objective-C Version``
2804 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
2806 * - ``Objective-C Image Info Version``
2807 - **[Required]** --- The version of the image info section. Currently
2810 * - ``Objective-C Image Info Section``
2811 - **[Required]** --- The section to place the metadata. Valid values are
2812 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
2813 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
2814 Objective-C ABI version 2.
2816 * - ``Objective-C Garbage Collection``
2817 - **[Required]** --- Specifies whether garbage collection is supported or
2818 not. Valid values are 0, for no garbage collection, and 2, for garbage
2819 collection supported.
2821 * - ``Objective-C GC Only``
2822 - **[Optional]** --- Specifies that only garbage collection is supported.
2823 If present, its value must be 6. This flag requires that the
2824 ``Objective-C Garbage Collection`` flag have the value 2.
2826 Some important flag interactions:
2828 - If a module with ``Objective-C Garbage Collection`` set to 0 is
2829 merged with a module with ``Objective-C Garbage Collection`` set to
2830 2, then the resulting module has the
2831 ``Objective-C Garbage Collection`` flag set to 0.
2832 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
2833 merged with a module with ``Objective-C GC Only`` set to 6.
2835 Automatic Linker Flags Module Flags Metadata
2836 --------------------------------------------
2838 Some targets support embedding flags to the linker inside individual object
2839 files. Typically this is used in conjunction with language extensions which
2840 allow source files to explicitly declare the libraries they depend on, and have
2841 these automatically be transmitted to the linker via object files.
2843 These flags are encoded in the IR using metadata in the module flags section,
2844 using the ``Linker Options`` key. The merge behavior for this flag is required
2845 to be ``AppendUnique``, and the value for the key is expected to be a metadata
2846 node which should be a list of other metadata nodes, each of which should be a
2847 list of metadata strings defining linker options.
2849 For example, the following metadata section specifies two separate sets of
2850 linker options, presumably to link against ``libz`` and the ``Cocoa``
2853 !0 = metadata !{ i32 6, metadata !"Linker Options",
2855 metadata !{ metadata !"-lz" },
2856 metadata !{ metadata !"-framework", metadata !"Cocoa" } } }
2857 !llvm.module.flags = !{ !0 }
2859 The metadata encoding as lists of lists of options, as opposed to a collapsed
2860 list of options, is chosen so that the IR encoding can use multiple option
2861 strings to specify e.g., a single library, while still having that specifier be
2862 preserved as an atomic element that can be recognized by a target specific
2863 assembly writer or object file emitter.
2865 Each individual option is required to be either a valid option for the target's
2866 linker, or an option that is reserved by the target specific assembly writer or
2867 object file emitter. No other aspect of these options is defined by the IR.
2869 Intrinsic Global Variables
2870 ==========================
2872 LLVM has a number of "magic" global variables that contain data that
2873 affect code generation or other IR semantics. These are documented here.
2874 All globals of this sort should have a section specified as
2875 "``llvm.metadata``". This section and all globals that start with
2876 "``llvm.``" are reserved for use by LLVM.
2878 The '``llvm.used``' Global Variable
2879 -----------------------------------
2881 The ``@llvm.used`` global is an array which has
2882 :ref:`appending linkage <linkage_appending>`. This array contains a list of
2883 pointers to global variables, functions and aliases which may optionally have a
2884 pointer cast formed of bitcast or getelementptr. For example, a legal
2887 .. code-block:: llvm
2892 @llvm.used = appending global [2 x i8*] [
2894 i8* bitcast (i32* @Y to i8*)
2895 ], section "llvm.metadata"
2897 If a symbol appears in the ``@llvm.used`` list, then the compiler, assembler,
2898 and linker are required to treat the symbol as if there is a reference to the
2899 symbol that it cannot see. For example, if a variable has internal linkage and
2900 no references other than that from the ``@llvm.used`` list, it cannot be
2901 deleted. This is commonly used to represent references from inline asms and
2902 other things the compiler cannot "see", and corresponds to
2903 "``attribute((used))``" in GNU C.
2905 On some targets, the code generator must emit a directive to the
2906 assembler or object file to prevent the assembler and linker from
2907 molesting the symbol.
2909 The '``llvm.compiler.used``' Global Variable
2910 --------------------------------------------
2912 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
2913 directive, except that it only prevents the compiler from touching the
2914 symbol. On targets that support it, this allows an intelligent linker to
2915 optimize references to the symbol without being impeded as it would be
2918 This is a rare construct that should only be used in rare circumstances,
2919 and should not be exposed to source languages.
2921 The '``llvm.global_ctors``' Global Variable
2922 -------------------------------------------
2924 .. code-block:: llvm
2926 %0 = type { i32, void ()* }
2927 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
2929 The ``@llvm.global_ctors`` array contains a list of constructor
2930 functions and associated priorities. The functions referenced by this
2931 array will be called in ascending order of priority (i.e. lowest first)
2932 when the module is loaded. The order of functions with the same priority
2935 The '``llvm.global_dtors``' Global Variable
2936 -------------------------------------------
2938 .. code-block:: llvm
2940 %0 = type { i32, void ()* }
2941 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
2943 The ``@llvm.global_dtors`` array contains a list of destructor functions
2944 and associated priorities. The functions referenced by this array will
2945 be called in descending order of priority (i.e. highest first) when the
2946 module is loaded. The order of functions with the same priority is not
2949 Instruction Reference
2950 =====================
2952 The LLVM instruction set consists of several different classifications
2953 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
2954 instructions <binaryops>`, :ref:`bitwise binary
2955 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
2956 :ref:`other instructions <otherops>`.
2960 Terminator Instructions
2961 -----------------------
2963 As mentioned :ref:`previously <functionstructure>`, every basic block in a
2964 program ends with a "Terminator" instruction, which indicates which
2965 block should be executed after the current block is finished. These
2966 terminator instructions typically yield a '``void``' value: they produce
2967 control flow, not values (the one exception being the
2968 ':ref:`invoke <i_invoke>`' instruction).
2970 The terminator instructions are: ':ref:`ret <i_ret>`',
2971 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
2972 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
2973 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
2977 '``ret``' Instruction
2978 ^^^^^^^^^^^^^^^^^^^^^
2985 ret <type> <value> ; Return a value from a non-void function
2986 ret void ; Return from void function
2991 The '``ret``' instruction is used to return control flow (and optionally
2992 a value) from a function back to the caller.
2994 There are two forms of the '``ret``' instruction: one that returns a
2995 value and then causes control flow, and one that just causes control
3001 The '``ret``' instruction optionally accepts a single argument, the
3002 return value. The type of the return value must be a ':ref:`first
3003 class <t_firstclass>`' type.
3005 A function is not :ref:`well formed <wellformed>` if it it has a non-void
3006 return type and contains a '``ret``' instruction with no return value or
3007 a return value with a type that does not match its type, or if it has a
3008 void return type and contains a '``ret``' instruction with a return
3014 When the '``ret``' instruction is executed, control flow returns back to
3015 the calling function's context. If the caller is a
3016 ":ref:`call <i_call>`" instruction, execution continues at the
3017 instruction after the call. If the caller was an
3018 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
3019 beginning of the "normal" destination block. If the instruction returns
3020 a value, that value shall set the call or invoke instruction's return
3026 .. code-block:: llvm
3028 ret i32 5 ; Return an integer value of 5
3029 ret void ; Return from a void function
3030 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
3034 '``br``' Instruction
3035 ^^^^^^^^^^^^^^^^^^^^
3042 br i1 <cond>, label <iftrue>, label <iffalse>
3043 br label <dest> ; Unconditional branch
3048 The '``br``' instruction is used to cause control flow to transfer to a
3049 different basic block in the current function. There are two forms of
3050 this instruction, corresponding to a conditional branch and an
3051 unconditional branch.
3056 The conditional branch form of the '``br``' instruction takes a single
3057 '``i1``' value and two '``label``' values. The unconditional form of the
3058 '``br``' instruction takes a single '``label``' value as a target.
3063 Upon execution of a conditional '``br``' instruction, the '``i1``'
3064 argument is evaluated. If the value is ``true``, control flows to the
3065 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
3066 to the '``iffalse``' ``label`` argument.
3071 .. code-block:: llvm
3074 %cond = icmp eq i32 %a, %b
3075 br i1 %cond, label %IfEqual, label %IfUnequal
3083 '``switch``' Instruction
3084 ^^^^^^^^^^^^^^^^^^^^^^^^
3091 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
3096 The '``switch``' instruction is used to transfer control flow to one of
3097 several different places. It is a generalization of the '``br``'
3098 instruction, allowing a branch to occur to one of many possible
3104 The '``switch``' instruction uses three parameters: an integer
3105 comparison value '``value``', a default '``label``' destination, and an
3106 array of pairs of comparison value constants and '``label``'s. The table
3107 is not allowed to contain duplicate constant entries.
3112 The ``switch`` instruction specifies a table of values and destinations.
3113 When the '``switch``' instruction is executed, this table is searched
3114 for the given value. If the value is found, control flow is transferred
3115 to the corresponding destination; otherwise, control flow is transferred
3116 to the default destination.
3121 Depending on properties of the target machine and the particular
3122 ``switch`` instruction, this instruction may be code generated in
3123 different ways. For example, it could be generated as a series of
3124 chained conditional branches or with a lookup table.
3129 .. code-block:: llvm
3131 ; Emulate a conditional br instruction
3132 %Val = zext i1 %value to i32
3133 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
3135 ; Emulate an unconditional br instruction
3136 switch i32 0, label %dest [ ]
3138 ; Implement a jump table:
3139 switch i32 %val, label %otherwise [ i32 0, label %onzero
3141 i32 2, label %ontwo ]
3145 '``indirectbr``' Instruction
3146 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3153 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
3158 The '``indirectbr``' instruction implements an indirect branch to a
3159 label within the current function, whose address is specified by
3160 "``address``". Address must be derived from a
3161 :ref:`blockaddress <blockaddress>` constant.
3166 The '``address``' argument is the address of the label to jump to. The
3167 rest of the arguments indicate the full set of possible destinations
3168 that the address may point to. Blocks are allowed to occur multiple
3169 times in the destination list, though this isn't particularly useful.
3171 This destination list is required so that dataflow analysis has an
3172 accurate understanding of the CFG.
3177 Control transfers to the block specified in the address argument. All
3178 possible destination blocks must be listed in the label list, otherwise
3179 this instruction has undefined behavior. This implies that jumps to
3180 labels defined in other functions have undefined behavior as well.
3185 This is typically implemented with a jump through a register.
3190 .. code-block:: llvm
3192 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
3196 '``invoke``' Instruction
3197 ^^^^^^^^^^^^^^^^^^^^^^^^
3204 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
3205 to label <normal label> unwind label <exception label>
3210 The '``invoke``' instruction causes control to transfer to a specified
3211 function, with the possibility of control flow transfer to either the
3212 '``normal``' label or the '``exception``' label. If the callee function
3213 returns with the "``ret``" instruction, control flow will return to the
3214 "normal" label. If the callee (or any indirect callees) returns via the
3215 ":ref:`resume <i_resume>`" instruction or other exception handling
3216 mechanism, control is interrupted and continued at the dynamically
3217 nearest "exception" label.
3219 The '``exception``' label is a `landing
3220 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
3221 '``exception``' label is required to have the
3222 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
3223 information about the behavior of the program after unwinding happens,
3224 as its first non-PHI instruction. The restrictions on the
3225 "``landingpad``" instruction's tightly couples it to the "``invoke``"
3226 instruction, so that the important information contained within the
3227 "``landingpad``" instruction can't be lost through normal code motion.
3232 This instruction requires several arguments:
3234 #. The optional "cconv" marker indicates which :ref:`calling
3235 convention <callingconv>` the call should use. If none is
3236 specified, the call defaults to using C calling conventions.
3237 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
3238 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
3240 #. '``ptr to function ty``': shall be the signature of the pointer to
3241 function value being invoked. In most cases, this is a direct
3242 function invocation, but indirect ``invoke``'s are just as possible,
3243 branching off an arbitrary pointer to function value.
3244 #. '``function ptr val``': An LLVM value containing a pointer to a
3245 function to be invoked.
3246 #. '``function args``': argument list whose types match the function
3247 signature argument types and parameter attributes. All arguments must
3248 be of :ref:`first class <t_firstclass>` type. If the function signature
3249 indicates the function accepts a variable number of arguments, the
3250 extra arguments can be specified.
3251 #. '``normal label``': the label reached when the called function
3252 executes a '``ret``' instruction.
3253 #. '``exception label``': the label reached when a callee returns via
3254 the :ref:`resume <i_resume>` instruction or other exception handling
3256 #. The optional :ref:`function attributes <fnattrs>` list. Only
3257 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
3258 attributes are valid here.
3263 This instruction is designed to operate as a standard '``call``'
3264 instruction in most regards. The primary difference is that it
3265 establishes an association with a label, which is used by the runtime
3266 library to unwind the stack.
3268 This instruction is used in languages with destructors to ensure that
3269 proper cleanup is performed in the case of either a ``longjmp`` or a
3270 thrown exception. Additionally, this is important for implementation of
3271 '``catch``' clauses in high-level languages that support them.
3273 For the purposes of the SSA form, the definition of the value returned
3274 by the '``invoke``' instruction is deemed to occur on the edge from the
3275 current block to the "normal" label. If the callee unwinds then no
3276 return value is available.
3281 .. code-block:: llvm
3283 %retval = invoke i32 @Test(i32 15) to label %Continue
3284 unwind label %TestCleanup ; {i32}:retval set
3285 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
3286 unwind label %TestCleanup ; {i32}:retval set
3290 '``resume``' Instruction
3291 ^^^^^^^^^^^^^^^^^^^^^^^^
3298 resume <type> <value>
3303 The '``resume``' instruction is a terminator instruction that has no
3309 The '``resume``' instruction requires one argument, which must have the
3310 same type as the result of any '``landingpad``' instruction in the same
3316 The '``resume``' instruction resumes propagation of an existing
3317 (in-flight) exception whose unwinding was interrupted with a
3318 :ref:`landingpad <i_landingpad>` instruction.
3323 .. code-block:: llvm
3325 resume { i8*, i32 } %exn
3329 '``unreachable``' Instruction
3330 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3342 The '``unreachable``' instruction has no defined semantics. This
3343 instruction is used to inform the optimizer that a particular portion of
3344 the code is not reachable. This can be used to indicate that the code
3345 after a no-return function cannot be reached, and other facts.
3350 The '``unreachable``' instruction has no defined semantics.
3357 Binary operators are used to do most of the computation in a program.
3358 They require two operands of the same type, execute an operation on
3359 them, and produce a single value. The operands might represent multiple
3360 data, as is the case with the :ref:`vector <t_vector>` data type. The
3361 result value has the same type as its operands.
3363 There are several different binary operators:
3367 '``add``' Instruction
3368 ^^^^^^^^^^^^^^^^^^^^^
3375 <result> = add <ty> <op1>, <op2> ; yields {ty}:result
3376 <result> = add nuw <ty> <op1>, <op2> ; yields {ty}:result
3377 <result> = add nsw <ty> <op1>, <op2> ; yields {ty}:result
3378 <result> = add nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3383 The '``add``' instruction returns the sum of its two operands.
3388 The two arguments to the '``add``' instruction must be
3389 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3390 arguments must have identical types.
3395 The value produced is the integer sum of the two operands.
3397 If the sum has unsigned overflow, the result returned is the
3398 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3401 Because LLVM integers use a two's complement representation, this
3402 instruction is appropriate for both signed and unsigned integers.
3404 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3405 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3406 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
3407 unsigned and/or signed overflow, respectively, occurs.
3412 .. code-block:: llvm
3414 <result> = add i32 4, %var ; yields {i32}:result = 4 + %var
3418 '``fadd``' Instruction
3419 ^^^^^^^^^^^^^^^^^^^^^^
3426 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3431 The '``fadd``' instruction returns the sum of its two operands.
3436 The two arguments to the '``fadd``' instruction must be :ref:`floating
3437 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3438 Both arguments must have identical types.
3443 The value produced is the floating point sum of the two operands. This
3444 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
3445 which are optimization hints to enable otherwise unsafe floating point
3451 .. code-block:: llvm
3453 <result> = fadd float 4.0, %var ; yields {float}:result = 4.0 + %var
3455 '``sub``' Instruction
3456 ^^^^^^^^^^^^^^^^^^^^^
3463 <result> = sub <ty> <op1>, <op2> ; yields {ty}:result
3464 <result> = sub nuw <ty> <op1>, <op2> ; yields {ty}:result
3465 <result> = sub nsw <ty> <op1>, <op2> ; yields {ty}:result
3466 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3471 The '``sub``' instruction returns the difference of its two operands.
3473 Note that the '``sub``' instruction is used to represent the '``neg``'
3474 instruction present in most other intermediate representations.
3479 The two arguments to the '``sub``' instruction must be
3480 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3481 arguments must have identical types.
3486 The value produced is the integer difference of the two operands.
3488 If the difference has unsigned overflow, the result returned is the
3489 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3492 Because LLVM integers use a two's complement representation, this
3493 instruction is appropriate for both signed and unsigned integers.
3495 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3496 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3497 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
3498 unsigned and/or signed overflow, respectively, occurs.
3503 .. code-block:: llvm
3505 <result> = sub i32 4, %var ; yields {i32}:result = 4 - %var
3506 <result> = sub i32 0, %val ; yields {i32}:result = -%var
3510 '``fsub``' Instruction
3511 ^^^^^^^^^^^^^^^^^^^^^^
3518 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3523 The '``fsub``' instruction returns the difference of its two operands.
3525 Note that the '``fsub``' instruction is used to represent the '``fneg``'
3526 instruction present in most other intermediate representations.
3531 The two arguments to the '``fsub``' instruction must be :ref:`floating
3532 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3533 Both arguments must have identical types.
3538 The value produced is the floating point difference of the two operands.
3539 This instruction can also take any number of :ref:`fast-math
3540 flags <fastmath>`, which are optimization hints to enable otherwise
3541 unsafe floating point optimizations:
3546 .. code-block:: llvm
3548 <result> = fsub float 4.0, %var ; yields {float}:result = 4.0 - %var
3549 <result> = fsub float -0.0, %val ; yields {float}:result = -%var
3551 '``mul``' Instruction
3552 ^^^^^^^^^^^^^^^^^^^^^
3559 <result> = mul <ty> <op1>, <op2> ; yields {ty}:result
3560 <result> = mul nuw <ty> <op1>, <op2> ; yields {ty}:result
3561 <result> = mul nsw <ty> <op1>, <op2> ; yields {ty}:result
3562 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3567 The '``mul``' instruction returns the product of its two operands.
3572 The two arguments to the '``mul``' instruction must be
3573 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3574 arguments must have identical types.
3579 The value produced is the integer product of the two operands.
3581 If the result of the multiplication has unsigned overflow, the result
3582 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
3583 bit width of the result.
3585 Because LLVM integers use a two's complement representation, and the
3586 result is the same width as the operands, this instruction returns the
3587 correct result for both signed and unsigned integers. If a full product
3588 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
3589 sign-extended or zero-extended as appropriate to the width of the full
3592 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3593 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3594 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
3595 unsigned and/or signed overflow, respectively, occurs.
3600 .. code-block:: llvm
3602 <result> = mul i32 4, %var ; yields {i32}:result = 4 * %var
3606 '``fmul``' Instruction
3607 ^^^^^^^^^^^^^^^^^^^^^^
3614 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3619 The '``fmul``' instruction returns the product of its two operands.
3624 The two arguments to the '``fmul``' instruction must be :ref:`floating
3625 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3626 Both arguments must have identical types.
3631 The value produced is the floating point product of the two operands.
3632 This instruction can also take any number of :ref:`fast-math
3633 flags <fastmath>`, which are optimization hints to enable otherwise
3634 unsafe floating point optimizations:
3639 .. code-block:: llvm
3641 <result> = fmul float 4.0, %var ; yields {float}:result = 4.0 * %var
3643 '``udiv``' Instruction
3644 ^^^^^^^^^^^^^^^^^^^^^^
3651 <result> = udiv <ty> <op1>, <op2> ; yields {ty}:result
3652 <result> = udiv exact <ty> <op1>, <op2> ; yields {ty}:result
3657 The '``udiv``' instruction returns the quotient of its two operands.
3662 The two arguments to the '``udiv``' instruction must be
3663 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3664 arguments must have identical types.
3669 The value produced is the unsigned integer quotient of the two operands.
3671 Note that unsigned integer division and signed integer division are
3672 distinct operations; for signed integer division, use '``sdiv``'.
3674 Division by zero leads to undefined behavior.
3676 If the ``exact`` keyword is present, the result value of the ``udiv`` is
3677 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
3678 such, "((a udiv exact b) mul b) == a").
3683 .. code-block:: llvm
3685 <result> = udiv i32 4, %var ; yields {i32}:result = 4 / %var
3687 '``sdiv``' Instruction
3688 ^^^^^^^^^^^^^^^^^^^^^^
3695 <result> = sdiv <ty> <op1>, <op2> ; yields {ty}:result
3696 <result> = sdiv exact <ty> <op1>, <op2> ; yields {ty}:result
3701 The '``sdiv``' instruction returns the quotient of its two operands.
3706 The two arguments to the '``sdiv``' instruction must be
3707 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3708 arguments must have identical types.
3713 The value produced is the signed integer quotient of the two operands
3714 rounded towards zero.
3716 Note that signed integer division and unsigned integer division are
3717 distinct operations; for unsigned integer division, use '``udiv``'.
3719 Division by zero leads to undefined behavior. Overflow also leads to
3720 undefined behavior; this is a rare case, but can occur, for example, by
3721 doing a 32-bit division of -2147483648 by -1.
3723 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
3724 a :ref:`poison value <poisonvalues>` if the result would be rounded.
3729 .. code-block:: llvm
3731 <result> = sdiv i32 4, %var ; yields {i32}:result = 4 / %var
3735 '``fdiv``' Instruction
3736 ^^^^^^^^^^^^^^^^^^^^^^
3743 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3748 The '``fdiv``' instruction returns the quotient of its two operands.
3753 The two arguments to the '``fdiv``' instruction must be :ref:`floating
3754 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3755 Both arguments must have identical types.
3760 The value produced is the floating point quotient of the two operands.
3761 This instruction can also take any number of :ref:`fast-math
3762 flags <fastmath>`, which are optimization hints to enable otherwise
3763 unsafe floating point optimizations:
3768 .. code-block:: llvm
3770 <result> = fdiv float 4.0, %var ; yields {float}:result = 4.0 / %var
3772 '``urem``' Instruction
3773 ^^^^^^^^^^^^^^^^^^^^^^
3780 <result> = urem <ty> <op1>, <op2> ; yields {ty}:result
3785 The '``urem``' instruction returns the remainder from the unsigned
3786 division of its two arguments.
3791 The two arguments to the '``urem``' instruction must be
3792 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3793 arguments must have identical types.
3798 This instruction returns the unsigned integer *remainder* of a division.
3799 This instruction always performs an unsigned division to get the
3802 Note that unsigned integer remainder and signed integer remainder are
3803 distinct operations; for signed integer remainder, use '``srem``'.
3805 Taking the remainder of a division by zero leads to undefined behavior.
3810 .. code-block:: llvm
3812 <result> = urem i32 4, %var ; yields {i32}:result = 4 % %var
3814 '``srem``' Instruction
3815 ^^^^^^^^^^^^^^^^^^^^^^
3822 <result> = srem <ty> <op1>, <op2> ; yields {ty}:result
3827 The '``srem``' instruction returns the remainder from the signed
3828 division of its two operands. This instruction can also take
3829 :ref:`vector <t_vector>` versions of the values in which case the elements
3835 The two arguments to the '``srem``' instruction must be
3836 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3837 arguments must have identical types.
3842 This instruction returns the *remainder* of a division (where the result
3843 is either zero or has the same sign as the dividend, ``op1``), not the
3844 *modulo* operator (where the result is either zero or has the same sign
3845 as the divisor, ``op2``) of a value. For more information about the
3846 difference, see `The Math
3847 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
3848 table of how this is implemented in various languages, please see
3850 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
3852 Note that signed integer remainder and unsigned integer remainder are
3853 distinct operations; for unsigned integer remainder, use '``urem``'.
3855 Taking the remainder of a division by zero leads to undefined behavior.
3856 Overflow also leads to undefined behavior; this is a rare case, but can
3857 occur, for example, by taking the remainder of a 32-bit division of
3858 -2147483648 by -1. (The remainder doesn't actually overflow, but this
3859 rule lets srem be implemented using instructions that return both the
3860 result of the division and the remainder.)
3865 .. code-block:: llvm
3867 <result> = srem i32 4, %var ; yields {i32}:result = 4 % %var
3871 '``frem``' Instruction
3872 ^^^^^^^^^^^^^^^^^^^^^^
3879 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3884 The '``frem``' instruction returns the remainder from the division of
3890 The two arguments to the '``frem``' instruction must be :ref:`floating
3891 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3892 Both arguments must have identical types.
3897 This instruction returns the *remainder* of a division. The remainder
3898 has the same sign as the dividend. This instruction can also take any
3899 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
3900 to enable otherwise unsafe floating point optimizations:
3905 .. code-block:: llvm
3907 <result> = frem float 4.0, %var ; yields {float}:result = 4.0 % %var
3911 Bitwise Binary Operations
3912 -------------------------
3914 Bitwise binary operators are used to do various forms of bit-twiddling
3915 in a program. They are generally very efficient instructions and can
3916 commonly be strength reduced from other instructions. They require two
3917 operands of the same type, execute an operation on them, and produce a
3918 single value. The resulting value is the same type as its operands.
3920 '``shl``' Instruction
3921 ^^^^^^^^^^^^^^^^^^^^^
3928 <result> = shl <ty> <op1>, <op2> ; yields {ty}:result
3929 <result> = shl nuw <ty> <op1>, <op2> ; yields {ty}:result
3930 <result> = shl nsw <ty> <op1>, <op2> ; yields {ty}:result
3931 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3936 The '``shl``' instruction returns the first operand shifted to the left
3937 a specified number of bits.
3942 Both arguments to the '``shl``' instruction must be the same
3943 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
3944 '``op2``' is treated as an unsigned value.
3949 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
3950 where ``n`` is the width of the result. If ``op2`` is (statically or
3951 dynamically) negative or equal to or larger than the number of bits in
3952 ``op1``, the result is undefined. If the arguments are vectors, each
3953 vector element of ``op1`` is shifted by the corresponding shift amount
3956 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
3957 value <poisonvalues>` if it shifts out any non-zero bits. If the
3958 ``nsw`` keyword is present, then the shift produces a :ref:`poison
3959 value <poisonvalues>` if it shifts out any bits that disagree with the
3960 resultant sign bit. As such, NUW/NSW have the same semantics as they
3961 would if the shift were expressed as a mul instruction with the same
3962 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
3967 .. code-block:: llvm
3969 <result> = shl i32 4, %var ; yields {i32}: 4 << %var
3970 <result> = shl i32 4, 2 ; yields {i32}: 16
3971 <result> = shl i32 1, 10 ; yields {i32}: 1024
3972 <result> = shl i32 1, 32 ; undefined
3973 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
3975 '``lshr``' Instruction
3976 ^^^^^^^^^^^^^^^^^^^^^^
3983 <result> = lshr <ty> <op1>, <op2> ; yields {ty}:result
3984 <result> = lshr exact <ty> <op1>, <op2> ; yields {ty}:result
3989 The '``lshr``' instruction (logical shift right) returns the first
3990 operand shifted to the right a specified number of bits with zero fill.
3995 Both arguments to the '``lshr``' instruction must be the same
3996 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
3997 '``op2``' is treated as an unsigned value.
4002 This instruction always performs a logical shift right operation. The
4003 most significant bits of the result will be filled with zero bits after
4004 the shift. If ``op2`` is (statically or dynamically) equal to or larger
4005 than the number of bits in ``op1``, the result is undefined. If the
4006 arguments are vectors, each vector element of ``op1`` is shifted by the
4007 corresponding shift amount in ``op2``.
4009 If the ``exact`` keyword is present, the result value of the ``lshr`` is
4010 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4016 .. code-block:: llvm
4018 <result> = lshr i32 4, 1 ; yields {i32}:result = 2
4019 <result> = lshr i32 4, 2 ; yields {i32}:result = 1
4020 <result> = lshr i8 4, 3 ; yields {i8}:result = 0
4021 <result> = lshr i8 -2, 1 ; yields {i8}:result = 0x7F
4022 <result> = lshr i32 1, 32 ; undefined
4023 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
4025 '``ashr``' Instruction
4026 ^^^^^^^^^^^^^^^^^^^^^^
4033 <result> = ashr <ty> <op1>, <op2> ; yields {ty}:result
4034 <result> = ashr exact <ty> <op1>, <op2> ; yields {ty}:result
4039 The '``ashr``' instruction (arithmetic shift right) returns the first
4040 operand shifted to the right a specified number of bits with sign
4046 Both arguments to the '``ashr``' instruction must be the same
4047 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4048 '``op2``' is treated as an unsigned value.
4053 This instruction always performs an arithmetic shift right operation,
4054 The most significant bits of the result will be filled with the sign bit
4055 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
4056 than the number of bits in ``op1``, the result is undefined. If the
4057 arguments are vectors, each vector element of ``op1`` is shifted by the
4058 corresponding shift amount in ``op2``.
4060 If the ``exact`` keyword is present, the result value of the ``ashr`` is
4061 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4067 .. code-block:: llvm
4069 <result> = ashr i32 4, 1 ; yields {i32}:result = 2
4070 <result> = ashr i32 4, 2 ; yields {i32}:result = 1
4071 <result> = ashr i8 4, 3 ; yields {i8}:result = 0
4072 <result> = ashr i8 -2, 1 ; yields {i8}:result = -1
4073 <result> = ashr i32 1, 32 ; undefined
4074 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
4076 '``and``' Instruction
4077 ^^^^^^^^^^^^^^^^^^^^^
4084 <result> = and <ty> <op1>, <op2> ; yields {ty}:result
4089 The '``and``' instruction returns the bitwise logical and of its two
4095 The two arguments to the '``and``' instruction must be
4096 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4097 arguments must have identical types.
4102 The truth table used for the '``and``' instruction is:
4119 .. code-block:: llvm
4121 <result> = and i32 4, %var ; yields {i32}:result = 4 & %var
4122 <result> = and i32 15, 40 ; yields {i32}:result = 8
4123 <result> = and i32 4, 8 ; yields {i32}:result = 0
4125 '``or``' Instruction
4126 ^^^^^^^^^^^^^^^^^^^^
4133 <result> = or <ty> <op1>, <op2> ; yields {ty}:result
4138 The '``or``' instruction returns the bitwise logical inclusive or of its
4144 The two arguments to the '``or``' instruction must be
4145 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4146 arguments must have identical types.
4151 The truth table used for the '``or``' instruction is:
4170 <result> = or i32 4, %var ; yields {i32}:result = 4 | %var
4171 <result> = or i32 15, 40 ; yields {i32}:result = 47
4172 <result> = or i32 4, 8 ; yields {i32}:result = 12
4174 '``xor``' Instruction
4175 ^^^^^^^^^^^^^^^^^^^^^
4182 <result> = xor <ty> <op1>, <op2> ; yields {ty}:result
4187 The '``xor``' instruction returns the bitwise logical exclusive or of
4188 its two operands. The ``xor`` is used to implement the "one's
4189 complement" operation, which is the "~" operator in C.
4194 The two arguments to the '``xor``' instruction must be
4195 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4196 arguments must have identical types.
4201 The truth table used for the '``xor``' instruction is:
4218 .. code-block:: llvm
4220 <result> = xor i32 4, %var ; yields {i32}:result = 4 ^ %var
4221 <result> = xor i32 15, 40 ; yields {i32}:result = 39
4222 <result> = xor i32 4, 8 ; yields {i32}:result = 12
4223 <result> = xor i32 %V, -1 ; yields {i32}:result = ~%V
4228 LLVM supports several instructions to represent vector operations in a
4229 target-independent manner. These instructions cover the element-access
4230 and vector-specific operations needed to process vectors effectively.
4231 While LLVM does directly support these vector operations, many
4232 sophisticated algorithms will want to use target-specific intrinsics to
4233 take full advantage of a specific target.
4235 .. _i_extractelement:
4237 '``extractelement``' Instruction
4238 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4245 <result> = extractelement <n x <ty>> <val>, i32 <idx> ; yields <ty>
4250 The '``extractelement``' instruction extracts a single scalar element
4251 from a vector at a specified index.
4256 The first operand of an '``extractelement``' instruction is a value of
4257 :ref:`vector <t_vector>` type. The second operand is an index indicating
4258 the position from which to extract the element. The index may be a
4264 The result is a scalar of the same type as the element type of ``val``.
4265 Its value is the value at position ``idx`` of ``val``. If ``idx``
4266 exceeds the length of ``val``, the results are undefined.
4271 .. code-block:: llvm
4273 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
4275 .. _i_insertelement:
4277 '``insertelement``' Instruction
4278 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4285 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, i32 <idx> ; yields <n x <ty>>
4290 The '``insertelement``' instruction inserts a scalar element into a
4291 vector at a specified index.
4296 The first operand of an '``insertelement``' instruction is a value of
4297 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
4298 type must equal the element type of the first operand. The third operand
4299 is an index indicating the position at which to insert the value. The
4300 index may be a variable.
4305 The result is a vector of the same type as ``val``. Its element values
4306 are those of ``val`` except at position ``idx``, where it gets the value
4307 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
4313 .. code-block:: llvm
4315 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
4317 .. _i_shufflevector:
4319 '``shufflevector``' Instruction
4320 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4327 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
4332 The '``shufflevector``' instruction constructs a permutation of elements
4333 from two input vectors, returning a vector with the same element type as
4334 the input and length that is the same as the shuffle mask.
4339 The first two operands of a '``shufflevector``' instruction are vectors
4340 with the same type. The third argument is a shuffle mask whose element
4341 type is always 'i32'. The result of the instruction is a vector whose
4342 length is the same as the shuffle mask and whose element type is the
4343 same as the element type of the first two operands.
4345 The shuffle mask operand is required to be a constant vector with either
4346 constant integer or undef values.
4351 The elements of the two input vectors are numbered from left to right
4352 across both of the vectors. The shuffle mask operand specifies, for each
4353 element of the result vector, which element of the two input vectors the
4354 result element gets. The element selector may be undef (meaning "don't
4355 care") and the second operand may be undef if performing a shuffle from
4361 .. code-block:: llvm
4363 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4364 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
4365 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
4366 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
4367 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
4368 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
4369 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4370 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
4372 Aggregate Operations
4373 --------------------
4375 LLVM supports several instructions for working with
4376 :ref:`aggregate <t_aggregate>` values.
4380 '``extractvalue``' Instruction
4381 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4388 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
4393 The '``extractvalue``' instruction extracts the value of a member field
4394 from an :ref:`aggregate <t_aggregate>` value.
4399 The first operand of an '``extractvalue``' instruction is a value of
4400 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
4401 constant indices to specify which value to extract in a similar manner
4402 as indices in a '``getelementptr``' instruction.
4404 The major differences to ``getelementptr`` indexing are:
4406 - Since the value being indexed is not a pointer, the first index is
4407 omitted and assumed to be zero.
4408 - At least one index must be specified.
4409 - Not only struct indices but also array indices must be in bounds.
4414 The result is the value at the position in the aggregate specified by
4420 .. code-block:: llvm
4422 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
4426 '``insertvalue``' Instruction
4427 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4434 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
4439 The '``insertvalue``' instruction inserts a value into a member field in
4440 an :ref:`aggregate <t_aggregate>` value.
4445 The first operand of an '``insertvalue``' instruction is a value of
4446 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
4447 a first-class value to insert. The following operands are constant
4448 indices indicating the position at which to insert the value in a
4449 similar manner as indices in a '``extractvalue``' instruction. The value
4450 to insert must have the same type as the value identified by the
4456 The result is an aggregate of the same type as ``val``. Its value is
4457 that of ``val`` except that the value at the position specified by the
4458 indices is that of ``elt``.
4463 .. code-block:: llvm
4465 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
4466 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
4467 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 ; yields {i32 1, float %val}
4471 Memory Access and Addressing Operations
4472 ---------------------------------------
4474 A key design point of an SSA-based representation is how it represents
4475 memory. In LLVM, no memory locations are in SSA form, which makes things
4476 very simple. This section describes how to read, write, and allocate
4481 '``alloca``' Instruction
4482 ^^^^^^^^^^^^^^^^^^^^^^^^
4489 <result> = alloca <type>[, <ty> <NumElements>][, align <alignment>] ; yields {type*}:result
4494 The '``alloca``' instruction allocates memory on the stack frame of the
4495 currently executing function, to be automatically released when this
4496 function returns to its caller. The object is always allocated in the
4497 generic address space (address space zero).
4502 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
4503 bytes of memory on the runtime stack, returning a pointer of the
4504 appropriate type to the program. If "NumElements" is specified, it is
4505 the number of elements allocated, otherwise "NumElements" is defaulted
4506 to be one. If a constant alignment is specified, the value result of the
4507 allocation is guaranteed to be aligned to at least that boundary. If not
4508 specified, or if zero, the target can choose to align the allocation on
4509 any convenient boundary compatible with the type.
4511 '``type``' may be any sized type.
4516 Memory is allocated; a pointer is returned. The operation is undefined
4517 if there is insufficient stack space for the allocation. '``alloca``'d
4518 memory is automatically released when the function returns. The
4519 '``alloca``' instruction is commonly used to represent automatic
4520 variables that must have an address available. When the function returns
4521 (either with the ``ret`` or ``resume`` instructions), the memory is
4522 reclaimed. Allocating zero bytes is legal, but the result is undefined.
4523 The order in which memory is allocated (ie., which way the stack grows)
4529 .. code-block:: llvm
4531 %ptr = alloca i32 ; yields {i32*}:ptr
4532 %ptr = alloca i32, i32 4 ; yields {i32*}:ptr
4533 %ptr = alloca i32, i32 4, align 1024 ; yields {i32*}:ptr
4534 %ptr = alloca i32, align 1024 ; yields {i32*}:ptr
4538 '``load``' Instruction
4539 ^^^^^^^^^^^^^^^^^^^^^^
4546 <result> = load [volatile] <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>]
4547 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
4548 !<index> = !{ i32 1 }
4553 The '``load``' instruction is used to read from memory.
4558 The argument to the ``load`` instruction specifies the memory address
4559 from which to load. The pointer must point to a :ref:`first
4560 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
4561 then the optimizer is not allowed to modify the number or order of
4562 execution of this ``load`` with other :ref:`volatile
4563 operations <volatile>`.
4565 If the ``load`` is marked as ``atomic``, it takes an extra
4566 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
4567 ``release`` and ``acq_rel`` orderings are not valid on ``load``
4568 instructions. Atomic loads produce :ref:`defined <memmodel>` results
4569 when they may see multiple atomic stores. The type of the pointee must
4570 be an integer type whose bit width is a power of two greater than or
4571 equal to eight and less than or equal to a target-specific size limit.
4572 ``align`` must be explicitly specified on atomic loads, and the load has
4573 undefined behavior if the alignment is not set to a value which is at
4574 least the size in bytes of the pointee. ``!nontemporal`` does not have
4575 any defined semantics for atomic loads.
4577 The optional constant ``align`` argument specifies the alignment of the
4578 operation (that is, the alignment of the memory address). A value of 0
4579 or an omitted ``align`` argument means that the operation has the ABI
4580 alignment for the target. It is the responsibility of the code emitter
4581 to ensure that the alignment information is correct. Overestimating the
4582 alignment results in undefined behavior. Underestimating the alignment
4583 may produce less efficient code. An alignment of 1 is always safe.
4585 The optional ``!nontemporal`` metadata must reference a single
4586 metatadata name ``<index>`` corresponding to a metadata node with one
4587 ``i32`` entry of value 1. The existence of the ``!nontemporal``
4588 metatadata on the instruction tells the optimizer and code generator
4589 that this load is not expected to be reused in the cache. The code
4590 generator may select special instructions to save cache bandwidth, such
4591 as the ``MOVNT`` instruction on x86.
4593 The optional ``!invariant.load`` metadata must reference a single
4594 metatadata name ``<index>`` corresponding to a metadata node with no
4595 entries. The existence of the ``!invariant.load`` metatadata on the
4596 instruction tells the optimizer and code generator that this load
4597 address points to memory which does not change value during program
4598 execution. The optimizer may then move this load around, for example, by
4599 hoisting it out of loops using loop invariant code motion.
4604 The location of memory pointed to is loaded. If the value being loaded
4605 is of scalar type then the number of bytes read does not exceed the
4606 minimum number of bytes needed to hold all bits of the type. For
4607 example, loading an ``i24`` reads at most three bytes. When loading a
4608 value of a type like ``i20`` with a size that is not an integral number
4609 of bytes, the result is undefined if the value was not originally
4610 written using a store of the same type.
4615 .. code-block:: llvm
4617 %ptr = alloca i32 ; yields {i32*}:ptr
4618 store i32 3, i32* %ptr ; yields {void}
4619 %val = load i32* %ptr ; yields {i32}:val = i32 3
4623 '``store``' Instruction
4624 ^^^^^^^^^^^^^^^^^^^^^^^
4631 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields {void}
4632 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields {void}
4637 The '``store``' instruction is used to write to memory.
4642 There are two arguments to the ``store`` instruction: a value to store
4643 and an address at which to store it. The type of the ``<pointer>``
4644 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
4645 the ``<value>`` operand. If the ``store`` is marked as ``volatile``,
4646 then the optimizer is not allowed to modify the number or order of
4647 execution of this ``store`` with other :ref:`volatile
4648 operations <volatile>`.
4650 If the ``store`` is marked as ``atomic``, it takes an extra
4651 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
4652 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
4653 instructions. Atomic loads produce :ref:`defined <memmodel>` results
4654 when they may see multiple atomic stores. The type of the pointee must
4655 be an integer type whose bit width is a power of two greater than or
4656 equal to eight and less than or equal to a target-specific size limit.
4657 ``align`` must be explicitly specified on atomic stores, and the store
4658 has undefined behavior if the alignment is not set to a value which is
4659 at least the size in bytes of the pointee. ``!nontemporal`` does not
4660 have any defined semantics for atomic stores.
4662 The optional constant ``align`` argument specifies the alignment of the
4663 operation (that is, the alignment of the memory address). A value of 0
4664 or an omitted ``align`` argument means that the operation has the ABI
4665 alignment for the target. It is the responsibility of the code emitter
4666 to ensure that the alignment information is correct. Overestimating the
4667 alignment results in undefined behavior. Underestimating the
4668 alignment may produce less efficient code. An alignment of 1 is always
4671 The optional ``!nontemporal`` metadata must reference a single metatadata
4672 name ``<index>`` corresponding to a metadata node with one ``i32`` entry of
4673 value 1. The existence of the ``!nontemporal`` metatadata on the instruction
4674 tells the optimizer and code generator that this load is not expected to
4675 be reused in the cache. The code generator may select special
4676 instructions to save cache bandwidth, such as the MOVNT instruction on
4682 The contents of memory are updated to contain ``<value>`` at the
4683 location specified by the ``<pointer>`` operand. If ``<value>`` is
4684 of scalar type then the number of bytes written does not exceed the
4685 minimum number of bytes needed to hold all bits of the type. For
4686 example, storing an ``i24`` writes at most three bytes. When writing a
4687 value of a type like ``i20`` with a size that is not an integral number
4688 of bytes, it is unspecified what happens to the extra bits that do not
4689 belong to the type, but they will typically be overwritten.
4694 .. code-block:: llvm
4696 %ptr = alloca i32 ; yields {i32*}:ptr
4697 store i32 3, i32* %ptr ; yields {void}
4698 %val = load i32* %ptr ; yields {i32}:val = i32 3
4702 '``fence``' Instruction
4703 ^^^^^^^^^^^^^^^^^^^^^^^
4710 fence [singlethread] <ordering> ; yields {void}
4715 The '``fence``' instruction is used to introduce happens-before edges
4721 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
4722 defines what *synchronizes-with* edges they add. They can only be given
4723 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
4728 A fence A which has (at least) ``release`` ordering semantics
4729 *synchronizes with* a fence B with (at least) ``acquire`` ordering
4730 semantics if and only if there exist atomic operations X and Y, both
4731 operating on some atomic object M, such that A is sequenced before X, X
4732 modifies M (either directly or through some side effect of a sequence
4733 headed by X), Y is sequenced before B, and Y observes M. This provides a
4734 *happens-before* dependency between A and B. Rather than an explicit
4735 ``fence``, one (but not both) of the atomic operations X or Y might
4736 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
4737 still *synchronize-with* the explicit ``fence`` and establish the
4738 *happens-before* edge.
4740 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
4741 ``acquire`` and ``release`` semantics specified above, participates in
4742 the global program order of other ``seq_cst`` operations and/or fences.
4744 The optional ":ref:`singlethread <singlethread>`" argument specifies
4745 that the fence only synchronizes with other fences in the same thread.
4746 (This is useful for interacting with signal handlers.)
4751 .. code-block:: llvm
4753 fence acquire ; yields {void}
4754 fence singlethread seq_cst ; yields {void}
4758 '``cmpxchg``' Instruction
4759 ^^^^^^^^^^^^^^^^^^^^^^^^^
4766 cmpxchg [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <ordering> ; yields {ty}
4771 The '``cmpxchg``' instruction is used to atomically modify memory. It
4772 loads a value in memory and compares it to a given value. If they are
4773 equal, it stores a new value into the memory.
4778 There are three arguments to the '``cmpxchg``' instruction: an address
4779 to operate on, a value to compare to the value currently be at that
4780 address, and a new value to place at that address if the compared values
4781 are equal. The type of '<cmp>' must be an integer type whose bit width
4782 is a power of two greater than or equal to eight and less than or equal
4783 to a target-specific size limit. '<cmp>' and '<new>' must have the same
4784 type, and the type of '<pointer>' must be a pointer to that type. If the
4785 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
4786 to modify the number or order of execution of this ``cmpxchg`` with
4787 other :ref:`volatile operations <volatile>`.
4789 The :ref:`ordering <ordering>` argument specifies how this ``cmpxchg``
4790 synchronizes with other atomic operations.
4792 The optional "``singlethread``" argument declares that the ``cmpxchg``
4793 is only atomic with respect to code (usually signal handlers) running in
4794 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
4795 respect to all other code in the system.
4797 The pointer passed into cmpxchg must have alignment greater than or
4798 equal to the size in memory of the operand.
4803 The contents of memory at the location specified by the '``<pointer>``'
4804 operand is read and compared to '``<cmp>``'; if the read value is the
4805 equal, '``<new>``' is written. The original value at the location is
4808 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose
4809 of identifying release sequences. A failed ``cmpxchg`` is equivalent to an
4810 atomic load with an ordering parameter determined by dropping any
4811 ``release`` part of the ``cmpxchg``'s ordering.
4816 .. code-block:: llvm
4819 %orig = atomic load i32* %ptr unordered ; yields {i32}
4823 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
4824 %squared = mul i32 %cmp, %cmp
4825 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared ; yields {i32}
4826 %success = icmp eq i32 %cmp, %old
4827 br i1 %success, label %done, label %loop
4834 '``atomicrmw``' Instruction
4835 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
4842 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields {ty}
4847 The '``atomicrmw``' instruction is used to atomically modify memory.
4852 There are three arguments to the '``atomicrmw``' instruction: an
4853 operation to apply, an address whose value to modify, an argument to the
4854 operation. The operation must be one of the following keywords:
4868 The type of '<value>' must be an integer type whose bit width is a power
4869 of two greater than or equal to eight and less than or equal to a
4870 target-specific size limit. The type of the '``<pointer>``' operand must
4871 be a pointer to that type. If the ``atomicrmw`` is marked as
4872 ``volatile``, then the optimizer is not allowed to modify the number or
4873 order of execution of this ``atomicrmw`` with other :ref:`volatile
4874 operations <volatile>`.
4879 The contents of memory at the location specified by the '``<pointer>``'
4880 operand are atomically read, modified, and written back. The original
4881 value at the location is returned. The modification is specified by the
4884 - xchg: ``*ptr = val``
4885 - add: ``*ptr = *ptr + val``
4886 - sub: ``*ptr = *ptr - val``
4887 - and: ``*ptr = *ptr & val``
4888 - nand: ``*ptr = ~(*ptr & val)``
4889 - or: ``*ptr = *ptr | val``
4890 - xor: ``*ptr = *ptr ^ val``
4891 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
4892 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
4893 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
4895 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
4901 .. code-block:: llvm
4903 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields {i32}
4905 .. _i_getelementptr:
4907 '``getelementptr``' Instruction
4908 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4915 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
4916 <result> = getelementptr inbounds <pty>* <ptrval>{, <ty> <idx>}*
4917 <result> = getelementptr <ptr vector> ptrval, <vector index type> idx
4922 The '``getelementptr``' instruction is used to get the address of a
4923 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
4924 address calculation only and does not access memory.
4929 The first argument is always a pointer or a vector of pointers, and
4930 forms the basis of the calculation. The remaining arguments are indices
4931 that indicate which of the elements of the aggregate object are indexed.
4932 The interpretation of each index is dependent on the type being indexed
4933 into. The first index always indexes the pointer value given as the
4934 first argument, the second index indexes a value of the type pointed to
4935 (not necessarily the value directly pointed to, since the first index
4936 can be non-zero), etc. The first type indexed into must be a pointer
4937 value, subsequent types can be arrays, vectors, and structs. Note that
4938 subsequent types being indexed into can never be pointers, since that
4939 would require loading the pointer before continuing calculation.
4941 The type of each index argument depends on the type it is indexing into.
4942 When indexing into a (optionally packed) structure, only ``i32`` integer
4943 **constants** are allowed (when using a vector of indices they must all
4944 be the **same** ``i32`` integer constant). When indexing into an array,
4945 pointer or vector, integers of any width are allowed, and they are not
4946 required to be constant. These integers are treated as signed values
4949 For example, let's consider a C code fragment and how it gets compiled
4965 int *foo(struct ST *s) {
4966 return &s[1].Z.B[5][13];
4969 The LLVM code generated by Clang is:
4971 .. code-block:: llvm
4973 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
4974 %struct.ST = type { i32, double, %struct.RT }
4976 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
4978 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
4985 In the example above, the first index is indexing into the
4986 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
4987 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
4988 indexes into the third element of the structure, yielding a
4989 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
4990 structure. The third index indexes into the second element of the
4991 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
4992 dimensions of the array are subscripted into, yielding an '``i32``'
4993 type. The '``getelementptr``' instruction returns a pointer to this
4994 element, thus computing a value of '``i32*``' type.
4996 Note that it is perfectly legal to index partially through a structure,
4997 returning a pointer to an inner element. Because of this, the LLVM code
4998 for the given testcase is equivalent to:
5000 .. code-block:: llvm
5002 define i32* @foo(%struct.ST* %s) {
5003 %t1 = getelementptr %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
5004 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
5005 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
5006 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
5007 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
5011 If the ``inbounds`` keyword is present, the result value of the
5012 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
5013 pointer is not an *in bounds* address of an allocated object, or if any
5014 of the addresses that would be formed by successive addition of the
5015 offsets implied by the indices to the base address with infinitely
5016 precise signed arithmetic are not an *in bounds* address of that
5017 allocated object. The *in bounds* addresses for an allocated object are
5018 all the addresses that point into the object, plus the address one byte
5019 past the end. In cases where the base is a vector of pointers the
5020 ``inbounds`` keyword applies to each of the computations element-wise.
5022 If the ``inbounds`` keyword is not present, the offsets are added to the
5023 base address with silently-wrapping two's complement arithmetic. If the
5024 offsets have a different width from the pointer, they are sign-extended
5025 or truncated to the width of the pointer. The result value of the
5026 ``getelementptr`` may be outside the object pointed to by the base
5027 pointer. The result value may not necessarily be used to access memory
5028 though, even if it happens to point into allocated storage. See the
5029 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
5032 The getelementptr instruction is often confusing. For some more insight
5033 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
5038 .. code-block:: llvm
5040 ; yields [12 x i8]*:aptr
5041 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5043 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
5045 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
5047 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
5049 In cases where the pointer argument is a vector of pointers, each index
5050 must be a vector with the same number of elements. For example:
5052 .. code-block:: llvm
5054 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5056 Conversion Operations
5057 ---------------------
5059 The instructions in this category are the conversion instructions
5060 (casting) which all take a single operand and a type. They perform
5061 various bit conversions on the operand.
5063 '``trunc .. to``' Instruction
5064 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5071 <result> = trunc <ty> <value> to <ty2> ; yields ty2
5076 The '``trunc``' instruction truncates its operand to the type ``ty2``.
5081 The '``trunc``' instruction takes a value to trunc, and a type to trunc
5082 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
5083 of the same number of integers. The bit size of the ``value`` must be
5084 larger than the bit size of the destination type, ``ty2``. Equal sized
5085 types are not allowed.
5090 The '``trunc``' instruction truncates the high order bits in ``value``
5091 and converts the remaining bits to ``ty2``. Since the source size must
5092 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
5093 It will always truncate bits.
5098 .. code-block:: llvm
5100 %X = trunc i32 257 to i8 ; yields i8:1
5101 %Y = trunc i32 123 to i1 ; yields i1:true
5102 %Z = trunc i32 122 to i1 ; yields i1:false
5103 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
5105 '``zext .. to``' Instruction
5106 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5113 <result> = zext <ty> <value> to <ty2> ; yields ty2
5118 The '``zext``' instruction zero extends its operand to type ``ty2``.
5123 The '``zext``' instruction takes a value to cast, and a type to cast it
5124 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5125 the same number of integers. The bit size of the ``value`` must be
5126 smaller than the bit size of the destination type, ``ty2``.
5131 The ``zext`` fills the high order bits of the ``value`` with zero bits
5132 until it reaches the size of the destination type, ``ty2``.
5134 When zero extending from i1, the result will always be either 0 or 1.
5139 .. code-block:: llvm
5141 %X = zext i32 257 to i64 ; yields i64:257
5142 %Y = zext i1 true to i32 ; yields i32:1
5143 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5145 '``sext .. to``' Instruction
5146 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5153 <result> = sext <ty> <value> to <ty2> ; yields ty2
5158 The '``sext``' sign extends ``value`` to the type ``ty2``.
5163 The '``sext``' instruction takes a value to cast, and a type to cast it
5164 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5165 the same number of integers. The bit size of the ``value`` must be
5166 smaller than the bit size of the destination type, ``ty2``.
5171 The '``sext``' instruction performs a sign extension by copying the sign
5172 bit (highest order bit) of the ``value`` until it reaches the bit size
5173 of the type ``ty2``.
5175 When sign extending from i1, the extension always results in -1 or 0.
5180 .. code-block:: llvm
5182 %X = sext i8 -1 to i16 ; yields i16 :65535
5183 %Y = sext i1 true to i32 ; yields i32:-1
5184 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5186 '``fptrunc .. to``' Instruction
5187 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5194 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
5199 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
5204 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
5205 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
5206 The size of ``value`` must be larger than the size of ``ty2``. This
5207 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
5212 The '``fptrunc``' instruction truncates a ``value`` from a larger
5213 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
5214 point <t_floating>` type. If the value cannot fit within the
5215 destination type, ``ty2``, then the results are undefined.
5220 .. code-block:: llvm
5222 %X = fptrunc double 123.0 to float ; yields float:123.0
5223 %Y = fptrunc double 1.0E+300 to float ; yields undefined
5225 '``fpext .. to``' Instruction
5226 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5233 <result> = fpext <ty> <value> to <ty2> ; yields ty2
5238 The '``fpext``' extends a floating point ``value`` to a larger floating
5244 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
5245 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
5246 to. The source type must be smaller than the destination type.
5251 The '``fpext``' instruction extends the ``value`` from a smaller
5252 :ref:`floating point <t_floating>` type to a larger :ref:`floating
5253 point <t_floating>` type. The ``fpext`` cannot be used to make a
5254 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
5255 *no-op cast* for a floating point cast.
5260 .. code-block:: llvm
5262 %X = fpext float 3.125 to double ; yields double:3.125000e+00
5263 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
5265 '``fptoui .. to``' Instruction
5266 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5273 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
5278 The '``fptoui``' converts a floating point ``value`` to its unsigned
5279 integer equivalent of type ``ty2``.
5284 The '``fptoui``' instruction takes a value to cast, which must be a
5285 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5286 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5287 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5288 type with the same number of elements as ``ty``
5293 The '``fptoui``' instruction converts its :ref:`floating
5294 point <t_floating>` operand into the nearest (rounding towards zero)
5295 unsigned integer value. If the value cannot fit in ``ty2``, the results
5301 .. code-block:: llvm
5303 %X = fptoui double 123.0 to i32 ; yields i32:123
5304 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
5305 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
5307 '``fptosi .. to``' Instruction
5308 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5315 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
5320 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
5321 ``value`` to type ``ty2``.
5326 The '``fptosi``' instruction takes a value to cast, which must be a
5327 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5328 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5329 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5330 type with the same number of elements as ``ty``
5335 The '``fptosi``' instruction converts its :ref:`floating
5336 point <t_floating>` operand into the nearest (rounding towards zero)
5337 signed integer value. If the value cannot fit in ``ty2``, the results
5343 .. code-block:: llvm
5345 %X = fptosi double -123.0 to i32 ; yields i32:-123
5346 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
5347 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
5349 '``uitofp .. to``' Instruction
5350 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5357 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
5362 The '``uitofp``' instruction regards ``value`` as an unsigned integer
5363 and converts that value to the ``ty2`` type.
5368 The '``uitofp``' instruction takes a value to cast, which must be a
5369 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5370 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5371 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5372 type with the same number of elements as ``ty``
5377 The '``uitofp``' instruction interprets its operand as an unsigned
5378 integer quantity and converts it to the corresponding floating point
5379 value. If the value cannot fit in the floating point value, the results
5385 .. code-block:: llvm
5387 %X = uitofp i32 257 to float ; yields float:257.0
5388 %Y = uitofp i8 -1 to double ; yields double:255.0
5390 '``sitofp .. to``' Instruction
5391 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5398 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
5403 The '``sitofp``' instruction regards ``value`` as a signed integer and
5404 converts that value to the ``ty2`` type.
5409 The '``sitofp``' instruction takes a value to cast, which must be a
5410 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5411 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5412 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5413 type with the same number of elements as ``ty``
5418 The '``sitofp``' instruction interprets its operand as a signed integer
5419 quantity and converts it to the corresponding floating point value. If
5420 the value cannot fit in the floating point value, the results are
5426 .. code-block:: llvm
5428 %X = sitofp i32 257 to float ; yields float:257.0
5429 %Y = sitofp i8 -1 to double ; yields double:-1.0
5433 '``ptrtoint .. to``' Instruction
5434 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5441 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
5446 The '``ptrtoint``' instruction converts the pointer or a vector of
5447 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
5452 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
5453 a a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
5454 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
5455 a vector of integers type.
5460 The '``ptrtoint``' instruction converts ``value`` to integer type
5461 ``ty2`` by interpreting the pointer value as an integer and either
5462 truncating or zero extending that value to the size of the integer type.
5463 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
5464 ``value`` is larger than ``ty2`` then a truncation is done. If they are
5465 the same size, then nothing is done (*no-op cast*) other than a type
5471 .. code-block:: llvm
5473 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
5474 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
5475 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
5479 '``inttoptr .. to``' Instruction
5480 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5487 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
5492 The '``inttoptr``' instruction converts an integer ``value`` to a
5493 pointer type, ``ty2``.
5498 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
5499 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
5505 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
5506 applying either a zero extension or a truncation depending on the size
5507 of the integer ``value``. If ``value`` is larger than the size of a
5508 pointer then a truncation is done. If ``value`` is smaller than the size
5509 of a pointer then a zero extension is done. If they are the same size,
5510 nothing is done (*no-op cast*).
5515 .. code-block:: llvm
5517 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
5518 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
5519 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
5520 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
5524 '``bitcast .. to``' Instruction
5525 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5532 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
5537 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
5543 The '``bitcast``' instruction takes a value to cast, which must be a
5544 non-aggregate first class value, and a type to cast it to, which must
5545 also be a non-aggregate :ref:`first class <t_firstclass>` type. The bit
5546 sizes of ``value`` and the destination type, ``ty2``, must be identical.
5547 If the source type is a pointer, the destination type must also be a
5548 pointer. This instruction supports bitwise conversion of vectors to
5549 integers and to vectors of other types (as long as they have the same
5555 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It is
5556 always a *no-op cast* because no bits change with this conversion. The
5557 conversion is done as if the ``value`` had been stored to memory and
5558 read back as type ``ty2``. Pointer (or vector of pointers) types may
5559 only be converted to other pointer (or vector of pointers) types with
5560 this instruction. To convert pointers to other types, use the
5561 :ref:`inttoptr <i_inttoptr>` or :ref:`ptrtoint <i_ptrtoint>` instructions
5567 .. code-block:: llvm
5569 %X = bitcast i8 255 to i8 ; yields i8 :-1
5570 %Y = bitcast i32* %x to sint* ; yields sint*:%x
5571 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
5572 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
5579 The instructions in this category are the "miscellaneous" instructions,
5580 which defy better classification.
5584 '``icmp``' Instruction
5585 ^^^^^^^^^^^^^^^^^^^^^^
5592 <result> = icmp <cond> <ty> <op1>, <op2> ; yields {i1} or {<N x i1>}:result
5597 The '``icmp``' instruction returns a boolean value or a vector of
5598 boolean values based on comparison of its two integer, integer vector,
5599 pointer, or pointer vector operands.
5604 The '``icmp``' instruction takes three operands. The first operand is
5605 the condition code indicating the kind of comparison to perform. It is
5606 not a value, just a keyword. The possible condition code are:
5609 #. ``ne``: not equal
5610 #. ``ugt``: unsigned greater than
5611 #. ``uge``: unsigned greater or equal
5612 #. ``ult``: unsigned less than
5613 #. ``ule``: unsigned less or equal
5614 #. ``sgt``: signed greater than
5615 #. ``sge``: signed greater or equal
5616 #. ``slt``: signed less than
5617 #. ``sle``: signed less or equal
5619 The remaining two arguments must be :ref:`integer <t_integer>` or
5620 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
5621 must also be identical types.
5626 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
5627 code given as ``cond``. The comparison performed always yields either an
5628 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
5630 #. ``eq``: yields ``true`` if the operands are equal, ``false``
5631 otherwise. No sign interpretation is necessary or performed.
5632 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
5633 otherwise. No sign interpretation is necessary or performed.
5634 #. ``ugt``: interprets the operands as unsigned values and yields
5635 ``true`` if ``op1`` is greater than ``op2``.
5636 #. ``uge``: interprets the operands as unsigned values and yields
5637 ``true`` if ``op1`` is greater than or equal to ``op2``.
5638 #. ``ult``: interprets the operands as unsigned values and yields
5639 ``true`` if ``op1`` is less than ``op2``.
5640 #. ``ule``: interprets the operands as unsigned values and yields
5641 ``true`` if ``op1`` is less than or equal to ``op2``.
5642 #. ``sgt``: interprets the operands as signed values and yields ``true``
5643 if ``op1`` is greater than ``op2``.
5644 #. ``sge``: interprets the operands as signed values and yields ``true``
5645 if ``op1`` is greater than or equal to ``op2``.
5646 #. ``slt``: interprets the operands as signed values and yields ``true``
5647 if ``op1`` is less than ``op2``.
5648 #. ``sle``: interprets the operands as signed values and yields ``true``
5649 if ``op1`` is less than or equal to ``op2``.
5651 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
5652 are compared as if they were integers.
5654 If the operands are integer vectors, then they are compared element by
5655 element. The result is an ``i1`` vector with the same number of elements
5656 as the values being compared. Otherwise, the result is an ``i1``.
5661 .. code-block:: llvm
5663 <result> = icmp eq i32 4, 5 ; yields: result=false
5664 <result> = icmp ne float* %X, %X ; yields: result=false
5665 <result> = icmp ult i16 4, 5 ; yields: result=true
5666 <result> = icmp sgt i16 4, 5 ; yields: result=false
5667 <result> = icmp ule i16 -4, 5 ; yields: result=false
5668 <result> = icmp sge i16 4, 5 ; yields: result=false
5670 Note that the code generator does not yet support vector types with the
5671 ``icmp`` instruction.
5675 '``fcmp``' Instruction
5676 ^^^^^^^^^^^^^^^^^^^^^^
5683 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields {i1} or {<N x i1>}:result
5688 The '``fcmp``' instruction returns a boolean value or vector of boolean
5689 values based on comparison of its operands.
5691 If the operands are floating point scalars, then the result type is a
5692 boolean (:ref:`i1 <t_integer>`).
5694 If the operands are floating point vectors, then the result type is a
5695 vector of boolean with the same number of elements as the operands being
5701 The '``fcmp``' instruction takes three operands. The first operand is
5702 the condition code indicating the kind of comparison to perform. It is
5703 not a value, just a keyword. The possible condition code are:
5705 #. ``false``: no comparison, always returns false
5706 #. ``oeq``: ordered and equal
5707 #. ``ogt``: ordered and greater than
5708 #. ``oge``: ordered and greater than or equal
5709 #. ``olt``: ordered and less than
5710 #. ``ole``: ordered and less than or equal
5711 #. ``one``: ordered and not equal
5712 #. ``ord``: ordered (no nans)
5713 #. ``ueq``: unordered or equal
5714 #. ``ugt``: unordered or greater than
5715 #. ``uge``: unordered or greater than or equal
5716 #. ``ult``: unordered or less than
5717 #. ``ule``: unordered or less than or equal
5718 #. ``une``: unordered or not equal
5719 #. ``uno``: unordered (either nans)
5720 #. ``true``: no comparison, always returns true
5722 *Ordered* means that neither operand is a QNAN while *unordered* means
5723 that either operand may be a QNAN.
5725 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
5726 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
5727 type. They must have identical types.
5732 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
5733 condition code given as ``cond``. If the operands are vectors, then the
5734 vectors are compared element by element. Each comparison performed
5735 always yields an :ref:`i1 <t_integer>` result, as follows:
5737 #. ``false``: always yields ``false``, regardless of operands.
5738 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
5739 is equal to ``op2``.
5740 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
5741 is greater than ``op2``.
5742 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
5743 is greater than or equal to ``op2``.
5744 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
5745 is less than ``op2``.
5746 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
5747 is less than or equal to ``op2``.
5748 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
5749 is not equal to ``op2``.
5750 #. ``ord``: yields ``true`` if both operands are not a QNAN.
5751 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
5753 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
5754 greater than ``op2``.
5755 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
5756 greater than or equal to ``op2``.
5757 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
5759 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
5760 less than or equal to ``op2``.
5761 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
5762 not equal to ``op2``.
5763 #. ``uno``: yields ``true`` if either operand is a QNAN.
5764 #. ``true``: always yields ``true``, regardless of operands.
5769 .. code-block:: llvm
5771 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
5772 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
5773 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
5774 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
5776 Note that the code generator does not yet support vector types with the
5777 ``fcmp`` instruction.
5781 '``phi``' Instruction
5782 ^^^^^^^^^^^^^^^^^^^^^
5789 <result> = phi <ty> [ <val0>, <label0>], ...
5794 The '``phi``' instruction is used to implement the φ node in the SSA
5795 graph representing the function.
5800 The type of the incoming values is specified with the first type field.
5801 After this, the '``phi``' instruction takes a list of pairs as
5802 arguments, with one pair for each predecessor basic block of the current
5803 block. Only values of :ref:`first class <t_firstclass>` type may be used as
5804 the value arguments to the PHI node. Only labels may be used as the
5807 There must be no non-phi instructions between the start of a basic block
5808 and the PHI instructions: i.e. PHI instructions must be first in a basic
5811 For the purposes of the SSA form, the use of each incoming value is
5812 deemed to occur on the edge from the corresponding predecessor block to
5813 the current block (but after any definition of an '``invoke``'
5814 instruction's return value on the same edge).
5819 At runtime, the '``phi``' instruction logically takes on the value
5820 specified by the pair corresponding to the predecessor basic block that
5821 executed just prior to the current block.
5826 .. code-block:: llvm
5828 Loop: ; Infinite loop that counts from 0 on up...
5829 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5830 %nextindvar = add i32 %indvar, 1
5835 '``select``' Instruction
5836 ^^^^^^^^^^^^^^^^^^^^^^^^
5843 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
5845 selty is either i1 or {<N x i1>}
5850 The '``select``' instruction is used to choose one value based on a
5851 condition, without branching.
5856 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
5857 values indicating the condition, and two values of the same :ref:`first
5858 class <t_firstclass>` type. If the val1/val2 are vectors and the
5859 condition is a scalar, then entire vectors are selected, not individual
5865 If the condition is an i1 and it evaluates to 1, the instruction returns
5866 the first value argument; otherwise, it returns the second value
5869 If the condition is a vector of i1, then the value arguments must be
5870 vectors of the same size, and the selection is done element by element.
5875 .. code-block:: llvm
5877 %X = select i1 true, i8 17, i8 42 ; yields i8:17
5881 '``call``' Instruction
5882 ^^^^^^^^^^^^^^^^^^^^^^
5889 <result> = [tail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
5894 The '``call``' instruction represents a simple function call.
5899 This instruction requires several arguments:
5901 #. The optional "tail" marker indicates that the callee function does
5902 not access any allocas or varargs in the caller. Note that calls may
5903 be marked "tail" even if they do not occur before a
5904 :ref:`ret <i_ret>` instruction. If the "tail" marker is present, the
5905 function call is eligible for tail call optimization, but `might not
5906 in fact be optimized into a jump <CodeGenerator.html#tailcallopt>`_.
5907 The code generator may optimize calls marked "tail" with either 1)
5908 automatic `sibling call
5909 optimization <CodeGenerator.html#sibcallopt>`_ when the caller and
5910 callee have matching signatures, or 2) forced tail call optimization
5911 when the following extra requirements are met:
5913 - Caller and callee both have the calling convention ``fastcc``.
5914 - The call is in tail position (ret immediately follows call and ret
5915 uses value of call or is void).
5916 - Option ``-tailcallopt`` is enabled, or
5917 ``llvm::GuaranteedTailCallOpt`` is ``true``.
5918 - `Platform specific constraints are
5919 met. <CodeGenerator.html#tailcallopt>`_
5921 #. The optional "cconv" marker indicates which :ref:`calling
5922 convention <callingconv>` the call should use. If none is
5923 specified, the call defaults to using C calling conventions. The
5924 calling convention of the call must match the calling convention of
5925 the target function, or else the behavior is undefined.
5926 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
5927 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
5929 #. '``ty``': the type of the call instruction itself which is also the
5930 type of the return value. Functions that return no value are marked
5932 #. '``fnty``': shall be the signature of the pointer to function value
5933 being invoked. The argument types must match the types implied by
5934 this signature. This type can be omitted if the function is not
5935 varargs and if the function type does not return a pointer to a
5937 #. '``fnptrval``': An LLVM value containing a pointer to a function to
5938 be invoked. In most cases, this is a direct function invocation, but
5939 indirect ``call``'s are just as possible, calling an arbitrary pointer
5941 #. '``function args``': argument list whose types match the function
5942 signature argument types and parameter attributes. All arguments must
5943 be of :ref:`first class <t_firstclass>` type. If the function signature
5944 indicates the function accepts a variable number of arguments, the
5945 extra arguments can be specified.
5946 #. The optional :ref:`function attributes <fnattrs>` list. Only
5947 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
5948 attributes are valid here.
5953 The '``call``' instruction is used to cause control flow to transfer to
5954 a specified function, with its incoming arguments bound to the specified
5955 values. Upon a '``ret``' instruction in the called function, control
5956 flow continues with the instruction after the function call, and the
5957 return value of the function is bound to the result argument.
5962 .. code-block:: llvm
5964 %retval = call i32 @test(i32 %argc)
5965 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
5966 %X = tail call i32 @foo() ; yields i32
5967 %Y = tail call fastcc i32 @foo() ; yields i32
5968 call void %foo(i8 97 signext)
5970 %struct.A = type { i32, i8 }
5971 %r = call %struct.A @foo() ; yields { 32, i8 }
5972 %gr = extractvalue %struct.A %r, 0 ; yields i32
5973 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
5974 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
5975 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
5977 llvm treats calls to some functions with names and arguments that match
5978 the standard C99 library as being the C99 library functions, and may
5979 perform optimizations or generate code for them under that assumption.
5980 This is something we'd like to change in the future to provide better
5981 support for freestanding environments and non-C-based languages.
5985 '``va_arg``' Instruction
5986 ^^^^^^^^^^^^^^^^^^^^^^^^
5993 <resultval> = va_arg <va_list*> <arglist>, <argty>
5998 The '``va_arg``' instruction is used to access arguments passed through
5999 the "variable argument" area of a function call. It is used to implement
6000 the ``va_arg`` macro in C.
6005 This instruction takes a ``va_list*`` value and the type of the
6006 argument. It returns a value of the specified argument type and
6007 increments the ``va_list`` to point to the next argument. The actual
6008 type of ``va_list`` is target specific.
6013 The '``va_arg``' instruction loads an argument of the specified type
6014 from the specified ``va_list`` and causes the ``va_list`` to point to
6015 the next argument. For more information, see the variable argument
6016 handling :ref:`Intrinsic Functions <int_varargs>`.
6018 It is legal for this instruction to be called in a function which does
6019 not take a variable number of arguments, for example, the ``vfprintf``
6022 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
6023 function <intrinsics>` because it takes a type as an argument.
6028 See the :ref:`variable argument processing <int_varargs>` section.
6030 Note that the code generator does not yet fully support va\_arg on many
6031 targets. Also, it does not currently support va\_arg with aggregate
6032 types on any target.
6036 '``landingpad``' Instruction
6037 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6044 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
6045 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
6047 <clause> := catch <type> <value>
6048 <clause> := filter <array constant type> <array constant>
6053 The '``landingpad``' instruction is used by `LLVM's exception handling
6054 system <ExceptionHandling.html#overview>`_ to specify that a basic block
6055 is a landing pad --- one where the exception lands, and corresponds to the
6056 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
6057 defines values supplied by the personality function (``pers_fn``) upon
6058 re-entry to the function. The ``resultval`` has the type ``resultty``.
6063 This instruction takes a ``pers_fn`` value. This is the personality
6064 function associated with the unwinding mechanism. The optional
6065 ``cleanup`` flag indicates that the landing pad block is a cleanup.
6067 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
6068 contains the global variable representing the "type" that may be caught
6069 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
6070 clause takes an array constant as its argument. Use
6071 "``[0 x i8**] undef``" for a filter which cannot throw. The
6072 '``landingpad``' instruction must contain *at least* one ``clause`` or
6073 the ``cleanup`` flag.
6078 The '``landingpad``' instruction defines the values which are set by the
6079 personality function (``pers_fn``) upon re-entry to the function, and
6080 therefore the "result type" of the ``landingpad`` instruction. As with
6081 calling conventions, how the personality function results are
6082 represented in LLVM IR is target specific.
6084 The clauses are applied in order from top to bottom. If two
6085 ``landingpad`` instructions are merged together through inlining, the
6086 clauses from the calling function are appended to the list of clauses.
6087 When the call stack is being unwound due to an exception being thrown,
6088 the exception is compared against each ``clause`` in turn. If it doesn't
6089 match any of the clauses, and the ``cleanup`` flag is not set, then
6090 unwinding continues further up the call stack.
6092 The ``landingpad`` instruction has several restrictions:
6094 - A landing pad block is a basic block which is the unwind destination
6095 of an '``invoke``' instruction.
6096 - A landing pad block must have a '``landingpad``' instruction as its
6097 first non-PHI instruction.
6098 - There can be only one '``landingpad``' instruction within the landing
6100 - A basic block that is not a landing pad block may not include a
6101 '``landingpad``' instruction.
6102 - All '``landingpad``' instructions in a function must have the same
6103 personality function.
6108 .. code-block:: llvm
6110 ;; A landing pad which can catch an integer.
6111 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6113 ;; A landing pad that is a cleanup.
6114 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6116 ;; A landing pad which can catch an integer and can only throw a double.
6117 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6119 filter [1 x i8**] [@_ZTId]
6126 LLVM supports the notion of an "intrinsic function". These functions
6127 have well known names and semantics and are required to follow certain
6128 restrictions. Overall, these intrinsics represent an extension mechanism
6129 for the LLVM language that does not require changing all of the
6130 transformations in LLVM when adding to the language (or the bitcode
6131 reader/writer, the parser, etc...).
6133 Intrinsic function names must all start with an "``llvm.``" prefix. This
6134 prefix is reserved in LLVM for intrinsic names; thus, function names may
6135 not begin with this prefix. Intrinsic functions must always be external
6136 functions: you cannot define the body of intrinsic functions. Intrinsic
6137 functions may only be used in call or invoke instructions: it is illegal
6138 to take the address of an intrinsic function. Additionally, because
6139 intrinsic functions are part of the LLVM language, it is required if any
6140 are added that they be documented here.
6142 Some intrinsic functions can be overloaded, i.e., the intrinsic
6143 represents a family of functions that perform the same operation but on
6144 different data types. Because LLVM can represent over 8 million
6145 different integer types, overloading is used commonly to allow an
6146 intrinsic function to operate on any integer type. One or more of the
6147 argument types or the result type can be overloaded to accept any
6148 integer type. Argument types may also be defined as exactly matching a
6149 previous argument's type or the result type. This allows an intrinsic
6150 function which accepts multiple arguments, but needs all of them to be
6151 of the same type, to only be overloaded with respect to a single
6152 argument or the result.
6154 Overloaded intrinsics will have the names of its overloaded argument
6155 types encoded into its function name, each preceded by a period. Only
6156 those types which are overloaded result in a name suffix. Arguments
6157 whose type is matched against another type do not. For example, the
6158 ``llvm.ctpop`` function can take an integer of any width and returns an
6159 integer of exactly the same integer width. This leads to a family of
6160 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
6161 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
6162 overloaded, and only one type suffix is required. Because the argument's
6163 type is matched against the return type, it does not require its own
6166 To learn how to add an intrinsic function, please see the `Extending
6167 LLVM Guide <ExtendingLLVM.html>`_.
6171 Variable Argument Handling Intrinsics
6172 -------------------------------------
6174 Variable argument support is defined in LLVM with the
6175 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
6176 functions. These functions are related to the similarly named macros
6177 defined in the ``<stdarg.h>`` header file.
6179 All of these functions operate on arguments that use a target-specific
6180 value type "``va_list``". The LLVM assembly language reference manual
6181 does not define what this type is, so all transformations should be
6182 prepared to handle these functions regardless of the type used.
6184 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
6185 variable argument handling intrinsic functions are used.
6187 .. code-block:: llvm
6189 define i32 @test(i32 %X, ...) {
6190 ; Initialize variable argument processing
6192 %ap2 = bitcast i8** %ap to i8*
6193 call void @llvm.va_start(i8* %ap2)
6195 ; Read a single integer argument
6196 %tmp = va_arg i8** %ap, i32
6198 ; Demonstrate usage of llvm.va_copy and llvm.va_end
6200 %aq2 = bitcast i8** %aq to i8*
6201 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
6202 call void @llvm.va_end(i8* %aq2)
6204 ; Stop processing of arguments.
6205 call void @llvm.va_end(i8* %ap2)
6209 declare void @llvm.va_start(i8*)
6210 declare void @llvm.va_copy(i8*, i8*)
6211 declare void @llvm.va_end(i8*)
6215 '``llvm.va_start``' Intrinsic
6216 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6223 declare void %llvm.va_start(i8* <arglist>)
6228 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
6229 subsequent use by ``va_arg``.
6234 The argument is a pointer to a ``va_list`` element to initialize.
6239 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
6240 available in C. In a target-dependent way, it initializes the
6241 ``va_list`` element to which the argument points, so that the next call
6242 to ``va_arg`` will produce the first variable argument passed to the
6243 function. Unlike the C ``va_start`` macro, this intrinsic does not need
6244 to know the last argument of the function as the compiler can figure
6247 '``llvm.va_end``' Intrinsic
6248 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6255 declare void @llvm.va_end(i8* <arglist>)
6260 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
6261 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
6266 The argument is a pointer to a ``va_list`` to destroy.
6271 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
6272 available in C. In a target-dependent way, it destroys the ``va_list``
6273 element to which the argument points. Calls to
6274 :ref:`llvm.va_start <int_va_start>` and
6275 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
6280 '``llvm.va_copy``' Intrinsic
6281 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6288 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
6293 The '``llvm.va_copy``' intrinsic copies the current argument position
6294 from the source argument list to the destination argument list.
6299 The first argument is a pointer to a ``va_list`` element to initialize.
6300 The second argument is a pointer to a ``va_list`` element to copy from.
6305 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
6306 available in C. In a target-dependent way, it copies the source
6307 ``va_list`` element into the destination ``va_list`` element. This
6308 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
6309 arbitrarily complex and require, for example, memory allocation.
6311 Accurate Garbage Collection Intrinsics
6312 --------------------------------------
6314 LLVM support for `Accurate Garbage Collection <GarbageCollection.html>`_
6315 (GC) requires the implementation and generation of these intrinsics.
6316 These intrinsics allow identification of :ref:`GC roots on the
6317 stack <int_gcroot>`, as well as garbage collector implementations that
6318 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
6319 Front-ends for type-safe garbage collected languages should generate
6320 these intrinsics to make use of the LLVM garbage collectors. For more
6321 details, see `Accurate Garbage Collection with
6322 LLVM <GarbageCollection.html>`_.
6324 The garbage collection intrinsics only operate on objects in the generic
6325 address space (address space zero).
6329 '``llvm.gcroot``' Intrinsic
6330 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6337 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
6342 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
6343 the code generator, and allows some metadata to be associated with it.
6348 The first argument specifies the address of a stack object that contains
6349 the root pointer. The second pointer (which must be either a constant or
6350 a global value address) contains the meta-data to be associated with the
6356 At runtime, a call to this intrinsic stores a null pointer into the
6357 "ptrloc" location. At compile-time, the code generator generates
6358 information to allow the runtime to find the pointer at GC safe points.
6359 The '``llvm.gcroot``' intrinsic may only be used in a function which
6360 :ref:`specifies a GC algorithm <gc>`.
6364 '``llvm.gcread``' Intrinsic
6365 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6372 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
6377 The '``llvm.gcread``' intrinsic identifies reads of references from heap
6378 locations, allowing garbage collector implementations that require read
6384 The second argument is the address to read from, which should be an
6385 address allocated from the garbage collector. The first object is a
6386 pointer to the start of the referenced object, if needed by the language
6387 runtime (otherwise null).
6392 The '``llvm.gcread``' intrinsic has the same semantics as a load
6393 instruction, but may be replaced with substantially more complex code by
6394 the garbage collector runtime, as needed. The '``llvm.gcread``'
6395 intrinsic may only be used in a function which :ref:`specifies a GC
6400 '``llvm.gcwrite``' Intrinsic
6401 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6408 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
6413 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
6414 locations, allowing garbage collector implementations that require write
6415 barriers (such as generational or reference counting collectors).
6420 The first argument is the reference to store, the second is the start of
6421 the object to store it to, and the third is the address of the field of
6422 Obj to store to. If the runtime does not require a pointer to the
6423 object, Obj may be null.
6428 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
6429 instruction, but may be replaced with substantially more complex code by
6430 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
6431 intrinsic may only be used in a function which :ref:`specifies a GC
6434 Code Generator Intrinsics
6435 -------------------------
6437 These intrinsics are provided by LLVM to expose special features that
6438 may only be implemented with code generator support.
6440 '``llvm.returnaddress``' Intrinsic
6441 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6448 declare i8 *@llvm.returnaddress(i32 <level>)
6453 The '``llvm.returnaddress``' intrinsic attempts to compute a
6454 target-specific value indicating the return address of the current
6455 function or one of its callers.
6460 The argument to this intrinsic indicates which function to return the
6461 address for. Zero indicates the calling function, one indicates its
6462 caller, etc. The argument is **required** to be a constant integer
6468 The '``llvm.returnaddress``' intrinsic either returns a pointer
6469 indicating the return address of the specified call frame, or zero if it
6470 cannot be identified. The value returned by this intrinsic is likely to
6471 be incorrect or 0 for arguments other than zero, so it should only be
6472 used for debugging purposes.
6474 Note that calling this intrinsic does not prevent function inlining or
6475 other aggressive transformations, so the value returned may not be that
6476 of the obvious source-language caller.
6478 '``llvm.frameaddress``' Intrinsic
6479 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6486 declare i8* @llvm.frameaddress(i32 <level>)
6491 The '``llvm.frameaddress``' intrinsic attempts to return the
6492 target-specific frame pointer value for the specified stack frame.
6497 The argument to this intrinsic indicates which function to return the
6498 frame pointer for. Zero indicates the calling function, one indicates
6499 its caller, etc. The argument is **required** to be a constant integer
6505 The '``llvm.frameaddress``' intrinsic either returns a pointer
6506 indicating the frame address of the specified call frame, or zero if it
6507 cannot be identified. The value returned by this intrinsic is likely to
6508 be incorrect or 0 for arguments other than zero, so it should only be
6509 used for debugging purposes.
6511 Note that calling this intrinsic does not prevent function inlining or
6512 other aggressive transformations, so the value returned may not be that
6513 of the obvious source-language caller.
6517 '``llvm.stacksave``' Intrinsic
6518 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6525 declare i8* @llvm.stacksave()
6530 The '``llvm.stacksave``' intrinsic is used to remember the current state
6531 of the function stack, for use with
6532 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
6533 implementing language features like scoped automatic variable sized
6539 This intrinsic returns a opaque pointer value that can be passed to
6540 :ref:`llvm.stackrestore <int_stackrestore>`. When an
6541 ``llvm.stackrestore`` intrinsic is executed with a value saved from
6542 ``llvm.stacksave``, it effectively restores the state of the stack to
6543 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
6544 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
6545 were allocated after the ``llvm.stacksave`` was executed.
6547 .. _int_stackrestore:
6549 '``llvm.stackrestore``' Intrinsic
6550 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6557 declare void @llvm.stackrestore(i8* %ptr)
6562 The '``llvm.stackrestore``' intrinsic is used to restore the state of
6563 the function stack to the state it was in when the corresponding
6564 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
6565 useful for implementing language features like scoped automatic variable
6566 sized arrays in C99.
6571 See the description for :ref:`llvm.stacksave <int_stacksave>`.
6573 '``llvm.prefetch``' Intrinsic
6574 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6581 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
6586 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
6587 insert a prefetch instruction if supported; otherwise, it is a noop.
6588 Prefetches have no effect on the behavior of the program but can change
6589 its performance characteristics.
6594 ``address`` is the address to be prefetched, ``rw`` is the specifier
6595 determining if the fetch should be for a read (0) or write (1), and
6596 ``locality`` is a temporal locality specifier ranging from (0) - no
6597 locality, to (3) - extremely local keep in cache. The ``cache type``
6598 specifies whether the prefetch is performed on the data (1) or
6599 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
6600 arguments must be constant integers.
6605 This intrinsic does not modify the behavior of the program. In
6606 particular, prefetches cannot trap and do not produce a value. On
6607 targets that support this intrinsic, the prefetch can provide hints to
6608 the processor cache for better performance.
6610 '``llvm.pcmarker``' Intrinsic
6611 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6618 declare void @llvm.pcmarker(i32 <id>)
6623 The '``llvm.pcmarker``' intrinsic is a method to export a Program
6624 Counter (PC) in a region of code to simulators and other tools. The
6625 method is target specific, but it is expected that the marker will use
6626 exported symbols to transmit the PC of the marker. The marker makes no
6627 guarantees that it will remain with any specific instruction after
6628 optimizations. It is possible that the presence of a marker will inhibit
6629 optimizations. The intended use is to be inserted after optimizations to
6630 allow correlations of simulation runs.
6635 ``id`` is a numerical id identifying the marker.
6640 This intrinsic does not modify the behavior of the program. Backends
6641 that do not support this intrinsic may ignore it.
6643 '``llvm.readcyclecounter``' Intrinsic
6644 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6651 declare i64 @llvm.readcyclecounter()
6656 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
6657 counter register (or similar low latency, high accuracy clocks) on those
6658 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6659 should map to RPCC. As the backing counters overflow quickly (on the
6660 order of 9 seconds on alpha), this should only be used for small
6666 When directly supported, reading the cycle counter should not modify any
6667 memory. Implementations are allowed to either return a application
6668 specific value or a system wide value. On backends without support, this
6669 is lowered to a constant 0.
6671 Note that runtime support may be conditional on the privilege-level code is
6672 running at and the host platform.
6674 Standard C Library Intrinsics
6675 -----------------------------
6677 LLVM provides intrinsics for a few important standard C library
6678 functions. These intrinsics allow source-language front-ends to pass
6679 information about the alignment of the pointer arguments to the code
6680 generator, providing opportunity for more efficient code generation.
6684 '``llvm.memcpy``' Intrinsic
6685 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6690 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
6691 integer bit width and for different address spaces. Not all targets
6692 support all bit widths however.
6696 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
6697 i32 <len>, i32 <align>, i1 <isvolatile>)
6698 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
6699 i64 <len>, i32 <align>, i1 <isvolatile>)
6704 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
6705 source location to the destination location.
6707 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
6708 intrinsics do not return a value, takes extra alignment/isvolatile
6709 arguments and the pointers can be in specified address spaces.
6714 The first argument is a pointer to the destination, the second is a
6715 pointer to the source. The third argument is an integer argument
6716 specifying the number of bytes to copy, the fourth argument is the
6717 alignment of the source and destination locations, and the fifth is a
6718 boolean indicating a volatile access.
6720 If the call to this intrinsic has an alignment value that is not 0 or 1,
6721 then the caller guarantees that both the source and destination pointers
6722 are aligned to that boundary.
6724 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
6725 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
6726 very cleanly specified and it is unwise to depend on it.
6731 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
6732 source location to the destination location, which are not allowed to
6733 overlap. It copies "len" bytes of memory over. If the argument is known
6734 to be aligned to some boundary, this can be specified as the fourth
6735 argument, otherwise it should be set to 0 or 1.
6737 '``llvm.memmove``' Intrinsic
6738 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6743 This is an overloaded intrinsic. You can use llvm.memmove on any integer
6744 bit width and for different address space. Not all targets support all
6749 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
6750 i32 <len>, i32 <align>, i1 <isvolatile>)
6751 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
6752 i64 <len>, i32 <align>, i1 <isvolatile>)
6757 The '``llvm.memmove.*``' intrinsics move a block of memory from the
6758 source location to the destination location. It is similar to the
6759 '``llvm.memcpy``' intrinsic but allows the two memory locations to
6762 Note that, unlike the standard libc function, the ``llvm.memmove.*``
6763 intrinsics do not return a value, takes extra alignment/isvolatile
6764 arguments and the pointers can be in specified address spaces.
6769 The first argument is a pointer to the destination, the second is a
6770 pointer to the source. The third argument is an integer argument
6771 specifying the number of bytes to copy, the fourth argument is the
6772 alignment of the source and destination locations, and the fifth is a
6773 boolean indicating a volatile access.
6775 If the call to this intrinsic has an alignment value that is not 0 or 1,
6776 then the caller guarantees that the source and destination pointers are
6777 aligned to that boundary.
6779 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
6780 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
6781 not very cleanly specified and it is unwise to depend on it.
6786 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
6787 source location to the destination location, which may overlap. It
6788 copies "len" bytes of memory over. If the argument is known to be
6789 aligned to some boundary, this can be specified as the fourth argument,
6790 otherwise it should be set to 0 or 1.
6792 '``llvm.memset.*``' Intrinsics
6793 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6798 This is an overloaded intrinsic. You can use llvm.memset on any integer
6799 bit width and for different address spaces. However, not all targets
6800 support all bit widths.
6804 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
6805 i32 <len>, i32 <align>, i1 <isvolatile>)
6806 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
6807 i64 <len>, i32 <align>, i1 <isvolatile>)
6812 The '``llvm.memset.*``' intrinsics fill a block of memory with a
6813 particular byte value.
6815 Note that, unlike the standard libc function, the ``llvm.memset``
6816 intrinsic does not return a value and takes extra alignment/volatile
6817 arguments. Also, the destination can be in an arbitrary address space.
6822 The first argument is a pointer to the destination to fill, the second
6823 is the byte value with which to fill it, the third argument is an
6824 integer argument specifying the number of bytes to fill, and the fourth
6825 argument is the known alignment of the destination location.
6827 If the call to this intrinsic has an alignment value that is not 0 or 1,
6828 then the caller guarantees that the destination pointer is aligned to
6831 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
6832 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
6833 very cleanly specified and it is unwise to depend on it.
6838 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
6839 at the destination location. If the argument is known to be aligned to
6840 some boundary, this can be specified as the fourth argument, otherwise
6841 it should be set to 0 or 1.
6843 '``llvm.sqrt.*``' Intrinsic
6844 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6849 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
6850 floating point or vector of floating point type. Not all targets support
6855 declare float @llvm.sqrt.f32(float %Val)
6856 declare double @llvm.sqrt.f64(double %Val)
6857 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6858 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6859 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
6864 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
6865 returning the same value as the libm '``sqrt``' functions would. Unlike
6866 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
6867 negative numbers other than -0.0 (which allows for better optimization,
6868 because there is no need to worry about errno being set).
6869 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
6874 The argument and return value are floating point numbers of the same
6880 This function returns the sqrt of the specified operand if it is a
6881 nonnegative floating point number.
6883 '``llvm.powi.*``' Intrinsic
6884 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6889 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
6890 floating point or vector of floating point type. Not all targets support
6895 declare float @llvm.powi.f32(float %Val, i32 %power)
6896 declare double @llvm.powi.f64(double %Val, i32 %power)
6897 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6898 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6899 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
6904 The '``llvm.powi.*``' intrinsics return the first operand raised to the
6905 specified (positive or negative) power. The order of evaluation of
6906 multiplications is not defined. When a vector of floating point type is
6907 used, the second argument remains a scalar integer value.
6912 The second argument is an integer power, and the first is a value to
6913 raise to that power.
6918 This function returns the first value raised to the second power with an
6919 unspecified sequence of rounding operations.
6921 '``llvm.sin.*``' Intrinsic
6922 ^^^^^^^^^^^^^^^^^^^^^^^^^^
6927 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
6928 floating point or vector of floating point type. Not all targets support
6933 declare float @llvm.sin.f32(float %Val)
6934 declare double @llvm.sin.f64(double %Val)
6935 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6936 declare fp128 @llvm.sin.f128(fp128 %Val)
6937 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6942 The '``llvm.sin.*``' intrinsics return the sine of the operand.
6947 The argument and return value are floating point numbers of the same
6953 This function returns the sine of the specified operand, returning the
6954 same values as the libm ``sin`` functions would, and handles error
6955 conditions in the same way.
6957 '``llvm.cos.*``' Intrinsic
6958 ^^^^^^^^^^^^^^^^^^^^^^^^^^
6963 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
6964 floating point or vector of floating point type. Not all targets support
6969 declare float @llvm.cos.f32(float %Val)
6970 declare double @llvm.cos.f64(double %Val)
6971 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6972 declare fp128 @llvm.cos.f128(fp128 %Val)
6973 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6978 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
6983 The argument and return value are floating point numbers of the same
6989 This function returns the cosine of the specified operand, returning the
6990 same values as the libm ``cos`` functions would, and handles error
6991 conditions in the same way.
6993 '``llvm.pow.*``' Intrinsic
6994 ^^^^^^^^^^^^^^^^^^^^^^^^^^
6999 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
7000 floating point or vector of floating point type. Not all targets support
7005 declare float @llvm.pow.f32(float %Val, float %Power)
7006 declare double @llvm.pow.f64(double %Val, double %Power)
7007 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7008 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7009 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7014 The '``llvm.pow.*``' intrinsics return the first operand raised to the
7015 specified (positive or negative) power.
7020 The second argument is a floating point power, and the first is a value
7021 to raise to that power.
7026 This function returns the first value raised to the second power,
7027 returning the same values as the libm ``pow`` functions would, and
7028 handles error conditions in the same way.
7030 '``llvm.exp.*``' Intrinsic
7031 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7036 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
7037 floating point or vector of floating point type. Not all targets support
7042 declare float @llvm.exp.f32(float %Val)
7043 declare double @llvm.exp.f64(double %Val)
7044 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7045 declare fp128 @llvm.exp.f128(fp128 %Val)
7046 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7051 The '``llvm.exp.*``' intrinsics perform the exp function.
7056 The argument and return value are floating point numbers of the same
7062 This function returns the same values as the libm ``exp`` functions
7063 would, and handles error conditions in the same way.
7065 '``llvm.exp2.*``' Intrinsic
7066 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7071 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
7072 floating point or vector of floating point type. Not all targets support
7077 declare float @llvm.exp2.f32(float %Val)
7078 declare double @llvm.exp2.f64(double %Val)
7079 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
7080 declare fp128 @llvm.exp2.f128(fp128 %Val)
7081 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
7086 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
7091 The argument and return value are floating point numbers of the same
7097 This function returns the same values as the libm ``exp2`` functions
7098 would, and handles error conditions in the same way.
7100 '``llvm.log.*``' Intrinsic
7101 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7106 This is an overloaded intrinsic. You can use ``llvm.log`` on any
7107 floating point or vector of floating point type. Not all targets support
7112 declare float @llvm.log.f32(float %Val)
7113 declare double @llvm.log.f64(double %Val)
7114 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7115 declare fp128 @llvm.log.f128(fp128 %Val)
7116 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7121 The '``llvm.log.*``' intrinsics perform the log function.
7126 The argument and return value are floating point numbers of the same
7132 This function returns the same values as the libm ``log`` functions
7133 would, and handles error conditions in the same way.
7135 '``llvm.log10.*``' Intrinsic
7136 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7141 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
7142 floating point or vector of floating point type. Not all targets support
7147 declare float @llvm.log10.f32(float %Val)
7148 declare double @llvm.log10.f64(double %Val)
7149 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
7150 declare fp128 @llvm.log10.f128(fp128 %Val)
7151 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
7156 The '``llvm.log10.*``' intrinsics perform the log10 function.
7161 The argument and return value are floating point numbers of the same
7167 This function returns the same values as the libm ``log10`` functions
7168 would, and handles error conditions in the same way.
7170 '``llvm.log2.*``' Intrinsic
7171 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7176 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
7177 floating point or vector of floating point type. Not all targets support
7182 declare float @llvm.log2.f32(float %Val)
7183 declare double @llvm.log2.f64(double %Val)
7184 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
7185 declare fp128 @llvm.log2.f128(fp128 %Val)
7186 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
7191 The '``llvm.log2.*``' intrinsics perform the log2 function.
7196 The argument and return value are floating point numbers of the same
7202 This function returns the same values as the libm ``log2`` functions
7203 would, and handles error conditions in the same way.
7205 '``llvm.fma.*``' Intrinsic
7206 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7211 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
7212 floating point or vector of floating point type. Not all targets support
7217 declare float @llvm.fma.f32(float %a, float %b, float %c)
7218 declare double @llvm.fma.f64(double %a, double %b, double %c)
7219 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7220 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7221 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7226 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
7232 The argument and return value are floating point numbers of the same
7238 This function returns the same values as the libm ``fma`` functions
7241 '``llvm.fabs.*``' Intrinsic
7242 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7247 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
7248 floating point or vector of floating point type. Not all targets support
7253 declare float @llvm.fabs.f32(float %Val)
7254 declare double @llvm.fabs.f64(double %Val)
7255 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
7256 declare fp128 @llvm.fabs.f128(fp128 %Val)
7257 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
7262 The '``llvm.fabs.*``' intrinsics return the absolute value of the
7268 The argument and return value are floating point numbers of the same
7274 This function returns the same values as the libm ``fabs`` functions
7275 would, and handles error conditions in the same way.
7277 '``llvm.floor.*``' Intrinsic
7278 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7283 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
7284 floating point or vector of floating point type. Not all targets support
7289 declare float @llvm.floor.f32(float %Val)
7290 declare double @llvm.floor.f64(double %Val)
7291 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
7292 declare fp128 @llvm.floor.f128(fp128 %Val)
7293 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
7298 The '``llvm.floor.*``' intrinsics return the floor of the operand.
7303 The argument and return value are floating point numbers of the same
7309 This function returns the same values as the libm ``floor`` functions
7310 would, and handles error conditions in the same way.
7312 '``llvm.ceil.*``' Intrinsic
7313 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7318 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
7319 floating point or vector of floating point type. Not all targets support
7324 declare float @llvm.ceil.f32(float %Val)
7325 declare double @llvm.ceil.f64(double %Val)
7326 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
7327 declare fp128 @llvm.ceil.f128(fp128 %Val)
7328 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
7333 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
7338 The argument and return value are floating point numbers of the same
7344 This function returns the same values as the libm ``ceil`` functions
7345 would, and handles error conditions in the same way.
7347 '``llvm.trunc.*``' Intrinsic
7348 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7353 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
7354 floating point or vector of floating point type. Not all targets support
7359 declare float @llvm.trunc.f32(float %Val)
7360 declare double @llvm.trunc.f64(double %Val)
7361 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
7362 declare fp128 @llvm.trunc.f128(fp128 %Val)
7363 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
7368 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
7369 nearest integer not larger in magnitude than the operand.
7374 The argument and return value are floating point numbers of the same
7380 This function returns the same values as the libm ``trunc`` functions
7381 would, and handles error conditions in the same way.
7383 '``llvm.rint.*``' Intrinsic
7384 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7389 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
7390 floating point or vector of floating point type. Not all targets support
7395 declare float @llvm.rint.f32(float %Val)
7396 declare double @llvm.rint.f64(double %Val)
7397 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
7398 declare fp128 @llvm.rint.f128(fp128 %Val)
7399 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
7404 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
7405 nearest integer. It may raise an inexact floating-point exception if the
7406 operand isn't an integer.
7411 The argument and return value are floating point numbers of the same
7417 This function returns the same values as the libm ``rint`` functions
7418 would, and handles error conditions in the same way.
7420 '``llvm.nearbyint.*``' Intrinsic
7421 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7426 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
7427 floating point or vector of floating point type. Not all targets support
7432 declare float @llvm.nearbyint.f32(float %Val)
7433 declare double @llvm.nearbyint.f64(double %Val)
7434 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
7435 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
7436 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
7441 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
7447 The argument and return value are floating point numbers of the same
7453 This function returns the same values as the libm ``nearbyint``
7454 functions would, and handles error conditions in the same way.
7456 Bit Manipulation Intrinsics
7457 ---------------------------
7459 LLVM provides intrinsics for a few important bit manipulation
7460 operations. These allow efficient code generation for some algorithms.
7462 '``llvm.bswap.*``' Intrinsics
7463 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7468 This is an overloaded intrinsic function. You can use bswap on any
7469 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
7473 declare i16 @llvm.bswap.i16(i16 <id>)
7474 declare i32 @llvm.bswap.i32(i32 <id>)
7475 declare i64 @llvm.bswap.i64(i64 <id>)
7480 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
7481 values with an even number of bytes (positive multiple of 16 bits).
7482 These are useful for performing operations on data that is not in the
7483 target's native byte order.
7488 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
7489 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
7490 intrinsic returns an i32 value that has the four bytes of the input i32
7491 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
7492 returned i32 will have its bytes in 3, 2, 1, 0 order. The
7493 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
7494 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
7497 '``llvm.ctpop.*``' Intrinsic
7498 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7503 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
7504 bit width, or on any vector with integer elements. Not all targets
7505 support all bit widths or vector types, however.
7509 declare i8 @llvm.ctpop.i8(i8 <src>)
7510 declare i16 @llvm.ctpop.i16(i16 <src>)
7511 declare i32 @llvm.ctpop.i32(i32 <src>)
7512 declare i64 @llvm.ctpop.i64(i64 <src>)
7513 declare i256 @llvm.ctpop.i256(i256 <src>)
7514 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
7519 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
7525 The only argument is the value to be counted. The argument may be of any
7526 integer type, or a vector with integer elements. The return type must
7527 match the argument type.
7532 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
7533 each element of a vector.
7535 '``llvm.ctlz.*``' Intrinsic
7536 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7541 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
7542 integer bit width, or any vector whose elements are integers. Not all
7543 targets support all bit widths or vector types, however.
7547 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
7548 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
7549 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
7550 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
7551 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
7552 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
7557 The '``llvm.ctlz``' family of intrinsic functions counts the number of
7558 leading zeros in a variable.
7563 The first argument is the value to be counted. This argument may be of
7564 any integer type, or a vectory with integer element type. The return
7565 type must match the first argument type.
7567 The second argument must be a constant and is a flag to indicate whether
7568 the intrinsic should ensure that a zero as the first argument produces a
7569 defined result. Historically some architectures did not provide a
7570 defined result for zero values as efficiently, and many algorithms are
7571 now predicated on avoiding zero-value inputs.
7576 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
7577 zeros in a variable, or within each element of the vector. If
7578 ``src == 0`` then the result is the size in bits of the type of ``src``
7579 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
7580 ``llvm.ctlz(i32 2) = 30``.
7582 '``llvm.cttz.*``' Intrinsic
7583 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7588 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
7589 integer bit width, or any vector of integer elements. Not all targets
7590 support all bit widths or vector types, however.
7594 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
7595 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
7596 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
7597 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
7598 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
7599 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
7604 The '``llvm.cttz``' family of intrinsic functions counts the number of
7610 The first argument is the value to be counted. This argument may be of
7611 any integer type, or a vectory with integer element type. The return
7612 type must match the first argument type.
7614 The second argument must be a constant and is a flag to indicate whether
7615 the intrinsic should ensure that a zero as the first argument produces a
7616 defined result. Historically some architectures did not provide a
7617 defined result for zero values as efficiently, and many algorithms are
7618 now predicated on avoiding zero-value inputs.
7623 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
7624 zeros in a variable, or within each element of a vector. If ``src == 0``
7625 then the result is the size in bits of the type of ``src`` if
7626 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
7627 ``llvm.cttz(2) = 1``.
7629 Arithmetic with Overflow Intrinsics
7630 -----------------------------------
7632 LLVM provides intrinsics for some arithmetic with overflow operations.
7634 '``llvm.sadd.with.overflow.*``' Intrinsics
7635 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7640 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
7641 on any integer bit width.
7645 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7646 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7647 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7652 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
7653 a signed addition of the two arguments, and indicate whether an overflow
7654 occurred during the signed summation.
7659 The arguments (%a and %b) and the first element of the result structure
7660 may be of integer types of any bit width, but they must have the same
7661 bit width. The second element of the result structure must be of type
7662 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
7668 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
7669 a signed addition of the two variables. They return a structure --- the
7670 first element of which is the signed summation, and the second element
7671 of which is a bit specifying if the signed summation resulted in an
7677 .. code-block:: llvm
7679 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7680 %sum = extractvalue {i32, i1} %res, 0
7681 %obit = extractvalue {i32, i1} %res, 1
7682 br i1 %obit, label %overflow, label %normal
7684 '``llvm.uadd.with.overflow.*``' Intrinsics
7685 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7690 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
7691 on any integer bit width.
7695 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7696 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7697 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7702 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
7703 an unsigned addition of the two arguments, and indicate whether a carry
7704 occurred during the unsigned summation.
7709 The arguments (%a and %b) and the first element of the result structure
7710 may be of integer types of any bit width, but they must have the same
7711 bit width. The second element of the result structure must be of type
7712 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
7718 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
7719 an unsigned addition of the two arguments. They return a structure --- the
7720 first element of which is the sum, and the second element of which is a
7721 bit specifying if the unsigned summation resulted in a carry.
7726 .. code-block:: llvm
7728 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7729 %sum = extractvalue {i32, i1} %res, 0
7730 %obit = extractvalue {i32, i1} %res, 1
7731 br i1 %obit, label %carry, label %normal
7733 '``llvm.ssub.with.overflow.*``' Intrinsics
7734 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7739 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
7740 on any integer bit width.
7744 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7745 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7746 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7751 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
7752 a signed subtraction of the two arguments, and indicate whether an
7753 overflow occurred during the signed subtraction.
7758 The arguments (%a and %b) and the first element of the result structure
7759 may be of integer types of any bit width, but they must have the same
7760 bit width. The second element of the result structure must be of type
7761 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
7767 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
7768 a signed subtraction of the two arguments. They return a structure --- the
7769 first element of which is the subtraction, and the second element of
7770 which is a bit specifying if the signed subtraction resulted in an
7776 .. code-block:: llvm
7778 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7779 %sum = extractvalue {i32, i1} %res, 0
7780 %obit = extractvalue {i32, i1} %res, 1
7781 br i1 %obit, label %overflow, label %normal
7783 '``llvm.usub.with.overflow.*``' Intrinsics
7784 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7789 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
7790 on any integer bit width.
7794 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7795 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7796 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7801 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
7802 an unsigned subtraction of the two arguments, and indicate whether an
7803 overflow occurred during the unsigned subtraction.
7808 The arguments (%a and %b) and the first element of the result structure
7809 may be of integer types of any bit width, but they must have the same
7810 bit width. The second element of the result structure must be of type
7811 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
7817 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
7818 an unsigned subtraction of the two arguments. They return a structure ---
7819 the first element of which is the subtraction, and the second element of
7820 which is a bit specifying if the unsigned subtraction resulted in an
7826 .. code-block:: llvm
7828 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7829 %sum = extractvalue {i32, i1} %res, 0
7830 %obit = extractvalue {i32, i1} %res, 1
7831 br i1 %obit, label %overflow, label %normal
7833 '``llvm.smul.with.overflow.*``' Intrinsics
7834 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7839 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
7840 on any integer bit width.
7844 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7845 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7846 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7851 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
7852 a signed multiplication of the two arguments, and indicate whether an
7853 overflow occurred during the signed multiplication.
7858 The arguments (%a and %b) and the first element of the result structure
7859 may be of integer types of any bit width, but they must have the same
7860 bit width. The second element of the result structure must be of type
7861 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
7867 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
7868 a signed multiplication of the two arguments. They return a structure ---
7869 the first element of which is the multiplication, and the second element
7870 of which is a bit specifying if the signed multiplication resulted in an
7876 .. code-block:: llvm
7878 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7879 %sum = extractvalue {i32, i1} %res, 0
7880 %obit = extractvalue {i32, i1} %res, 1
7881 br i1 %obit, label %overflow, label %normal
7883 '``llvm.umul.with.overflow.*``' Intrinsics
7884 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7889 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
7890 on any integer bit width.
7894 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7895 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7896 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7901 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
7902 a unsigned multiplication of the two arguments, and indicate whether an
7903 overflow occurred during the unsigned multiplication.
7908 The arguments (%a and %b) and the first element of the result structure
7909 may be of integer types of any bit width, but they must have the same
7910 bit width. The second element of the result structure must be of type
7911 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
7917 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
7918 an unsigned multiplication of the two arguments. They return a structure ---
7919 the first element of which is the multiplication, and the second
7920 element of which is a bit specifying if the unsigned multiplication
7921 resulted in an overflow.
7926 .. code-block:: llvm
7928 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7929 %sum = extractvalue {i32, i1} %res, 0
7930 %obit = extractvalue {i32, i1} %res, 1
7931 br i1 %obit, label %overflow, label %normal
7933 Specialised Arithmetic Intrinsics
7934 ---------------------------------
7936 '``llvm.fmuladd.*``' Intrinsic
7937 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7944 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
7945 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
7950 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
7951 expressions that can be fused if the code generator determines that (a) the
7952 target instruction set has support for a fused operation, and (b) that the
7953 fused operation is more efficient than the equivalent, separate pair of mul
7954 and add instructions.
7959 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
7960 multiplicands, a and b, and an addend c.
7969 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
7971 is equivalent to the expression a \* b + c, except that rounding will
7972 not be performed between the multiplication and addition steps if the
7973 code generator fuses the operations. Fusion is not guaranteed, even if
7974 the target platform supports it. If a fused multiply-add is required the
7975 corresponding llvm.fma.\* intrinsic function should be used instead.
7980 .. code-block:: llvm
7982 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields {float}:r2 = (a * b) + c
7984 Half Precision Floating Point Intrinsics
7985 ----------------------------------------
7987 For most target platforms, half precision floating point is a
7988 storage-only format. This means that it is a dense encoding (in memory)
7989 but does not support computation in the format.
7991 This means that code must first load the half-precision floating point
7992 value as an i16, then convert it to float with
7993 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
7994 then be performed on the float value (including extending to double
7995 etc). To store the value back to memory, it is first converted to float
7996 if needed, then converted to i16 with
7997 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
8000 .. _int_convert_to_fp16:
8002 '``llvm.convert.to.fp16``' Intrinsic
8003 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8010 declare i16 @llvm.convert.to.fp16(f32 %a)
8015 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion
8016 from single precision floating point format to half precision floating
8022 The intrinsic function contains single argument - the value to be
8028 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion
8029 from single precision floating point format to half precision floating
8030 point format. The return value is an ``i16`` which contains the
8036 .. code-block:: llvm
8038 %res = call i16 @llvm.convert.to.fp16(f32 %a)
8039 store i16 %res, i16* @x, align 2
8041 .. _int_convert_from_fp16:
8043 '``llvm.convert.from.fp16``' Intrinsic
8044 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8051 declare f32 @llvm.convert.from.fp16(i16 %a)
8056 The '``llvm.convert.from.fp16``' intrinsic function performs a
8057 conversion from half precision floating point format to single precision
8058 floating point format.
8063 The intrinsic function contains single argument - the value to be
8069 The '``llvm.convert.from.fp16``' intrinsic function performs a
8070 conversion from half single precision floating point format to single
8071 precision floating point format. The input half-float value is
8072 represented by an ``i16`` value.
8077 .. code-block:: llvm
8079 %a = load i16* @x, align 2
8080 %res = call f32 @llvm.convert.from.fp16(i16 %a)
8085 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
8086 prefix), are described in the `LLVM Source Level
8087 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
8090 Exception Handling Intrinsics
8091 -----------------------------
8093 The LLVM exception handling intrinsics (which all start with
8094 ``llvm.eh.`` prefix), are described in the `LLVM Exception
8095 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
8099 Trampoline Intrinsics
8100 ---------------------
8102 These intrinsics make it possible to excise one parameter, marked with
8103 the :ref:`nest <nest>` attribute, from a function. The result is a
8104 callable function pointer lacking the nest parameter - the caller does
8105 not need to provide a value for it. Instead, the value to use is stored
8106 in advance in a "trampoline", a block of memory usually allocated on the
8107 stack, which also contains code to splice the nest value into the
8108 argument list. This is used to implement the GCC nested function address
8111 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
8112 then the resulting function pointer has signature ``i32 (i32, i32)*``.
8113 It can be created as follows:
8115 .. code-block:: llvm
8117 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
8118 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
8119 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
8120 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
8121 %fp = bitcast i8* %p to i32 (i32, i32)*
8123 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
8124 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
8128 '``llvm.init.trampoline``' Intrinsic
8129 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8136 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
8141 This fills the memory pointed to by ``tramp`` with executable code,
8142 turning it into a trampoline.
8147 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
8148 pointers. The ``tramp`` argument must point to a sufficiently large and
8149 sufficiently aligned block of memory; this memory is written to by the
8150 intrinsic. Note that the size and the alignment are target-specific -
8151 LLVM currently provides no portable way of determining them, so a
8152 front-end that generates this intrinsic needs to have some
8153 target-specific knowledge. The ``func`` argument must hold a function
8154 bitcast to an ``i8*``.
8159 The block of memory pointed to by ``tramp`` is filled with target
8160 dependent code, turning it into a function. Then ``tramp`` needs to be
8161 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
8162 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
8163 function's signature is the same as that of ``func`` with any arguments
8164 marked with the ``nest`` attribute removed. At most one such ``nest``
8165 argument is allowed, and it must be of pointer type. Calling the new
8166 function is equivalent to calling ``func`` with the same argument list,
8167 but with ``nval`` used for the missing ``nest`` argument. If, after
8168 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
8169 modified, then the effect of any later call to the returned function
8170 pointer is undefined.
8174 '``llvm.adjust.trampoline``' Intrinsic
8175 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8182 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
8187 This performs any required machine-specific adjustment to the address of
8188 a trampoline (passed as ``tramp``).
8193 ``tramp`` must point to a block of memory which already has trampoline
8194 code filled in by a previous call to
8195 :ref:`llvm.init.trampoline <int_it>`.
8200 On some architectures the address of the code to be executed needs to be
8201 different to the address where the trampoline is actually stored. This
8202 intrinsic returns the executable address corresponding to ``tramp``
8203 after performing the required machine specific adjustments. The pointer
8204 returned can then be :ref:`bitcast and executed <int_trampoline>`.
8209 This class of intrinsics exists to information about the lifetime of
8210 memory objects and ranges where variables are immutable.
8212 '``llvm.lifetime.start``' Intrinsic
8213 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8220 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
8225 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
8231 The first argument is a constant integer representing the size of the
8232 object, or -1 if it is variable sized. The second argument is a pointer
8238 This intrinsic indicates that before this point in the code, the value
8239 of the memory pointed to by ``ptr`` is dead. This means that it is known
8240 to never be used and has an undefined value. A load from the pointer
8241 that precedes this intrinsic can be replaced with ``'undef'``.
8243 '``llvm.lifetime.end``' Intrinsic
8244 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8251 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
8256 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
8262 The first argument is a constant integer representing the size of the
8263 object, or -1 if it is variable sized. The second argument is a pointer
8269 This intrinsic indicates that after this point in the code, the value of
8270 the memory pointed to by ``ptr`` is dead. This means that it is known to
8271 never be used and has an undefined value. Any stores into the memory
8272 object following this intrinsic may be removed as dead.
8274 '``llvm.invariant.start``' Intrinsic
8275 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8282 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
8287 The '``llvm.invariant.start``' intrinsic specifies that the contents of
8288 a memory object will not change.
8293 The first argument is a constant integer representing the size of the
8294 object, or -1 if it is variable sized. The second argument is a pointer
8300 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
8301 the return value, the referenced memory location is constant and
8304 '``llvm.invariant.end``' Intrinsic
8305 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8312 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
8317 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
8318 memory object are mutable.
8323 The first argument is the matching ``llvm.invariant.start`` intrinsic.
8324 The second argument is a constant integer representing the size of the
8325 object, or -1 if it is variable sized and the third argument is a
8326 pointer to the object.
8331 This intrinsic indicates that the memory is mutable again.
8336 This class of intrinsics is designed to be generic and has no specific
8339 '``llvm.var.annotation``' Intrinsic
8340 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8347 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
8352 The '``llvm.var.annotation``' intrinsic.
8357 The first argument is a pointer to a value, the second is a pointer to a
8358 global string, the third is a pointer to a global string which is the
8359 source file name, and the last argument is the line number.
8364 This intrinsic allows annotation of local variables with arbitrary
8365 strings. This can be useful for special purpose optimizations that want
8366 to look for these annotations. These have no other defined use; they are
8367 ignored by code generation and optimization.
8369 '``llvm.ptr.annotation.*``' Intrinsic
8370 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8375 This is an overloaded intrinsic. You can use '``llvm.ptr.annotation``' on a
8376 pointer to an integer of any width. *NOTE* you must specify an address space for
8377 the pointer. The identifier for the default address space is the integer
8382 declare i8* @llvm.ptr.annotation.p<address space>i8(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
8383 declare i16* @llvm.ptr.annotation.p<address space>i16(i16* <val>, i8* <str>, i8* <str>, i32 <int>)
8384 declare i32* @llvm.ptr.annotation.p<address space>i32(i32* <val>, i8* <str>, i8* <str>, i32 <int>)
8385 declare i64* @llvm.ptr.annotation.p<address space>i64(i64* <val>, i8* <str>, i8* <str>, i32 <int>)
8386 declare i256* @llvm.ptr.annotation.p<address space>i256(i256* <val>, i8* <str>, i8* <str>, i32 <int>)
8391 The '``llvm.ptr.annotation``' intrinsic.
8396 The first argument is a pointer to an integer value of arbitrary bitwidth
8397 (result of some expression), the second is a pointer to a global string, the
8398 third is a pointer to a global string which is the source file name, and the
8399 last argument is the line number. It returns the value of the first argument.
8404 This intrinsic allows annotation of a pointer to an integer with arbitrary
8405 strings. This can be useful for special purpose optimizations that want to look
8406 for these annotations. These have no other defined use; they are ignored by code
8407 generation and optimization.
8409 '``llvm.annotation.*``' Intrinsic
8410 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8415 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
8416 any integer bit width.
8420 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
8421 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
8422 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
8423 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
8424 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
8429 The '``llvm.annotation``' intrinsic.
8434 The first argument is an integer value (result of some expression), the
8435 second is a pointer to a global string, the third is a pointer to a
8436 global string which is the source file name, and the last argument is
8437 the line number. It returns the value of the first argument.
8442 This intrinsic allows annotations to be put on arbitrary expressions
8443 with arbitrary strings. This can be useful for special purpose
8444 optimizations that want to look for these annotations. These have no
8445 other defined use; they are ignored by code generation and optimization.
8447 '``llvm.trap``' Intrinsic
8448 ^^^^^^^^^^^^^^^^^^^^^^^^^
8455 declare void @llvm.trap() noreturn nounwind
8460 The '``llvm.trap``' intrinsic.
8470 This intrinsic is lowered to the target dependent trap instruction. If
8471 the target does not have a trap instruction, this intrinsic will be
8472 lowered to a call of the ``abort()`` function.
8474 '``llvm.debugtrap``' Intrinsic
8475 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8482 declare void @llvm.debugtrap() nounwind
8487 The '``llvm.debugtrap``' intrinsic.
8497 This intrinsic is lowered to code which is intended to cause an
8498 execution trap with the intention of requesting the attention of a
8501 '``llvm.stackprotector``' Intrinsic
8502 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8509 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
8514 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
8515 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
8516 is placed on the stack before local variables.
8521 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
8522 The first argument is the value loaded from the stack guard
8523 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
8524 enough space to hold the value of the guard.
8529 This intrinsic causes the prologue/epilogue inserter to force the
8530 position of the ``AllocaInst`` stack slot to be before local variables
8531 on the stack. This is to ensure that if a local variable on the stack is
8532 overwritten, it will destroy the value of the guard. When the function
8533 exits, the guard on the stack is checked against the original guard. If
8534 they are different, then the program aborts by calling the
8535 ``__stack_chk_fail()`` function.
8537 '``llvm.objectsize``' Intrinsic
8538 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8545 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
8546 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
8551 The ``llvm.objectsize`` intrinsic is designed to provide information to
8552 the optimizers to determine at compile time whether a) an operation
8553 (like memcpy) will overflow a buffer that corresponds to an object, or
8554 b) that a runtime check for overflow isn't necessary. An object in this
8555 context means an allocation of a specific class, structure, array, or
8561 The ``llvm.objectsize`` intrinsic takes two arguments. The first
8562 argument is a pointer to or into the ``object``. The second argument is
8563 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
8564 or -1 (if false) when the object size is unknown. The second argument
8565 only accepts constants.
8570 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
8571 the size of the object concerned. If the size cannot be determined at
8572 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
8573 on the ``min`` argument).
8575 '``llvm.expect``' Intrinsic
8576 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8583 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
8584 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
8589 The ``llvm.expect`` intrinsic provides information about expected (the
8590 most probable) value of ``val``, which can be used by optimizers.
8595 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
8596 a value. The second argument is an expected value, this needs to be a
8597 constant value, variables are not allowed.
8602 This intrinsic is lowered to the ``val``.
8604 '``llvm.donothing``' Intrinsic
8605 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8612 declare void @llvm.donothing() nounwind readnone
8617 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's the
8618 only intrinsic that can be called with an invoke instruction.
8628 This intrinsic does nothing, and it's removed by optimizers and ignored