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
132 It also shows a convention that we follow in this document. When
133 demonstrating instructions, we will follow an instruction with a comment
134 that defines the type and name of value produced.
142 LLVM programs are composed of ``Module``'s, each of which is a
143 translation unit of the input programs. Each module consists of
144 functions, global variables, and symbol table entries. Modules may be
145 combined together with the LLVM linker, which merges function (and
146 global variable) definitions, resolves forward declarations, and merges
147 symbol table entries. Here is an example of the "hello world" module:
151 ; Declare the string constant as a global constant.
152 @.str = private unnamed_addr constant [13 x i8] c"hello world\0A\00"
154 ; External declaration of the puts function
155 declare i32 @puts(i8* nocapture) nounwind
157 ; Definition of main function
158 define i32 @main() { ; i32()*
159 ; Convert [13 x i8]* to i8 *...
160 %cast210 = getelementptr [13 x i8]* @.str, i64 0, i64 0
162 ; Call puts function to write out the string to stdout.
163 call i32 @puts(i8* %cast210)
168 !1 = metadata !{i32 42}
171 This example is made up of a :ref:`global variable <globalvars>` named
172 "``.str``", an external declaration of the "``puts``" function, a
173 :ref:`function definition <functionstructure>` for "``main``" and
174 :ref:`named metadata <namedmetadatastructure>` "``foo``".
176 In general, a module is made up of a list of global values (where both
177 functions and global variables are global values). Global values are
178 represented by a pointer to a memory location (in this case, a pointer
179 to an array of char, and a pointer to a function), and have one of the
180 following :ref:`linkage types <linkage>`.
187 All Global Variables and Functions have one of the following types of
191 Global values with "``private``" linkage are only directly
192 accessible by objects in the current module. In particular, linking
193 code into a module with an private global value may cause the
194 private to be renamed as necessary to avoid collisions. Because the
195 symbol is private to the module, all references can be updated. This
196 doesn't show up in any symbol table in the object file.
198 Similar to ``private``, but the symbol is passed through the
199 assembler and evaluated by the linker. Unlike normal strong symbols,
200 they are removed by the linker from the final linked image
201 (executable or dynamic library).
202 ``linker_private_weak``
203 Similar to "``linker_private``", but the symbol is weak. Note that
204 ``linker_private_weak`` symbols are subject to coalescing by the
205 linker. The symbols are removed by the linker from the final linked
206 image (executable or dynamic library).
208 Similar to private, but the value shows as a local symbol
209 (``STB_LOCAL`` in the case of ELF) in the object file. This
210 corresponds to the notion of the '``static``' keyword in C.
211 ``available_externally``
212 Globals with "``available_externally``" linkage are never emitted
213 into the object file corresponding to the LLVM module. They exist to
214 allow inlining and other optimizations to take place given knowledge
215 of the definition of the global, which is known to be somewhere
216 outside the module. Globals with ``available_externally`` linkage
217 are allowed to be discarded at will, and are otherwise the same as
218 ``linkonce_odr``. This linkage type is only allowed on definitions,
221 Globals with "``linkonce``" linkage are merged with other globals of
222 the same name when linkage occurs. This can be used to implement
223 some forms of inline functions, templates, or other code which must
224 be generated in each translation unit that uses it, but where the
225 body may be overridden with a more definitive definition later.
226 Unreferenced ``linkonce`` globals are allowed to be discarded. Note
227 that ``linkonce`` linkage does not actually allow the optimizer to
228 inline the body of this function into callers because it doesn't
229 know if this definition of the function is the definitive definition
230 within the program or whether it will be overridden by a stronger
231 definition. To enable inlining and other optimizations, use
232 "``linkonce_odr``" linkage.
234 "``weak``" linkage has the same merging semantics as ``linkonce``
235 linkage, except that unreferenced globals with ``weak`` linkage may
236 not be discarded. This is used for globals that are declared "weak"
239 "``common``" linkage is most similar to "``weak``" linkage, but they
240 are used for tentative definitions in C, such as "``int X;``" at
241 global scope. Symbols with "``common``" linkage are merged in the
242 same way as ``weak symbols``, and they may not be deleted if
243 unreferenced. ``common`` symbols may not have an explicit section,
244 must have a zero initializer, and may not be marked
245 ':ref:`constant <globalvars>`'. Functions and aliases may not have
248 .. _linkage_appending:
251 "``appending``" linkage may only be applied to global variables of
252 pointer to array type. When two global variables with appending
253 linkage are linked together, the two global arrays are appended
254 together. This is the LLVM, typesafe, equivalent of having the
255 system linker append together "sections" with identical names when
258 The semantics of this linkage follow the ELF object file model: the
259 symbol is weak until linked, if not linked, the symbol becomes null
260 instead of being an undefined reference.
261 ``linkonce_odr``, ``weak_odr``
262 Some languages allow differing globals to be merged, such as two
263 functions with different semantics. Other languages, such as
264 ``C++``, ensure that only equivalent globals are ever merged (the
265 "one definition rule" --- "ODR"). Such languages can use the
266 ``linkonce_odr`` and ``weak_odr`` linkage types to indicate that the
267 global will only be merged with equivalent globals. These linkage
268 types are otherwise the same as their non-``odr`` versions.
269 ``linkonce_odr_auto_hide``
270 Similar to "``linkonce_odr``", but nothing in the translation unit
271 takes the address of this definition. For instance, functions that
272 had an inline definition, but the compiler decided not to inline it.
273 ``linkonce_odr_auto_hide`` may have only ``default`` visibility. The
274 symbols are removed by the linker from the final linked image
275 (executable or dynamic library).
277 If none of the above identifiers are used, the global is externally
278 visible, meaning that it participates in linkage and can be used to
279 resolve external symbol references.
281 The next two types of linkage are targeted for Microsoft Windows
282 platform only. They are designed to support importing (exporting)
283 symbols from (to) DLLs (Dynamic Link Libraries).
286 "``dllimport``" linkage causes the compiler to reference a function
287 or variable via a global pointer to a pointer that is set up by the
288 DLL exporting the symbol. On Microsoft Windows targets, the pointer
289 name is formed by combining ``__imp_`` and the function or variable
292 "``dllexport``" linkage causes the compiler to provide a global
293 pointer to a pointer in a DLL, so that it can be referenced with the
294 ``dllimport`` attribute. On Microsoft Windows targets, the pointer
295 name is formed by combining ``__imp_`` and the function or variable
298 For example, since the "``.LC0``" variable is defined to be internal, if
299 another module defined a "``.LC0``" variable and was linked with this
300 one, one of the two would be renamed, preventing a collision. Since
301 "``main``" and "``puts``" are external (i.e., lacking any linkage
302 declarations), they are accessible outside of the current module.
304 It is illegal for a function *declaration* to have any linkage type
305 other than ``external``, ``dllimport`` or ``extern_weak``.
307 Aliases can have only ``external``, ``internal``, ``weak`` or
308 ``weak_odr`` linkages.
315 LLVM :ref:`functions <functionstructure>`, :ref:`calls <i_call>` and
316 :ref:`invokes <i_invoke>` can all have an optional calling convention
317 specified for the call. The calling convention of any pair of dynamic
318 caller/callee must match, or the behavior of the program is undefined.
319 The following calling conventions are supported by LLVM, and more may be
322 "``ccc``" - The C calling convention
323 This calling convention (the default if no other calling convention
324 is specified) matches the target C calling conventions. This calling
325 convention supports varargs function calls and tolerates some
326 mismatch in the declared prototype and implemented declaration of
327 the function (as does normal C).
328 "``fastcc``" - The fast calling convention
329 This calling convention attempts to make calls as fast as possible
330 (e.g. by passing things in registers). This calling convention
331 allows the target to use whatever tricks it wants to produce fast
332 code for the target, without having to conform to an externally
333 specified ABI (Application Binary Interface). `Tail calls can only
334 be optimized when this, the GHC or the HiPE convention is
335 used. <CodeGenerator.html#id80>`_ This calling convention does not
336 support varargs and requires the prototype of all callees to exactly
337 match the prototype of the function definition.
338 "``coldcc``" - The cold calling convention
339 This calling convention attempts to make code in the caller as
340 efficient as possible under the assumption that the call is not
341 commonly executed. As such, these calls often preserve all registers
342 so that the call does not break any live ranges in the caller side.
343 This calling convention does not support varargs and requires the
344 prototype of all callees to exactly match the prototype of the
346 "``cc 10``" - GHC convention
347 This calling convention has been implemented specifically for use by
348 the `Glasgow Haskell Compiler (GHC) <http://www.haskell.org/ghc>`_.
349 It passes everything in registers, going to extremes to achieve this
350 by disabling callee save registers. This calling convention should
351 not be used lightly but only for specific situations such as an
352 alternative to the *register pinning* performance technique often
353 used when implementing functional programming languages. At the
354 moment only X86 supports this convention and it has the following
357 - On *X86-32* only supports up to 4 bit type parameters. No
358 floating point types are supported.
359 - On *X86-64* only supports up to 10 bit type parameters and 6
360 floating point parameters.
362 This calling convention supports `tail call
363 optimization <CodeGenerator.html#id80>`_ but requires both the
364 caller and callee are using it.
365 "``cc 11``" - The HiPE calling convention
366 This calling convention has been implemented specifically for use by
367 the `High-Performance Erlang
368 (HiPE) <http://www.it.uu.se/research/group/hipe/>`_ compiler, *the*
369 native code compiler of the `Ericsson's Open Source Erlang/OTP
370 system <http://www.erlang.org/download.shtml>`_. It uses more
371 registers for argument passing than the ordinary C calling
372 convention and defines no callee-saved registers. The calling
373 convention properly supports `tail call
374 optimization <CodeGenerator.html#id80>`_ but requires that both the
375 caller and the callee use it. It uses a *register pinning*
376 mechanism, similar to GHC's convention, for keeping frequently
377 accessed runtime components pinned to specific hardware registers.
378 At the moment only X86 supports this convention (both 32 and 64
380 "``cc <n>``" - Numbered convention
381 Any calling convention may be specified by number, allowing
382 target-specific calling conventions to be used. Target specific
383 calling conventions start at 64.
385 More calling conventions can be added/defined on an as-needed basis, to
386 support Pascal conventions or any other well-known target-independent
392 All Global Variables and Functions have one of the following visibility
395 "``default``" - Default style
396 On targets that use the ELF object file format, default visibility
397 means that the declaration is visible to other modules and, in
398 shared libraries, means that the declared entity may be overridden.
399 On Darwin, default visibility means that the declaration is visible
400 to other modules. Default visibility corresponds to "external
401 linkage" in the language.
402 "``hidden``" - Hidden style
403 Two declarations of an object with hidden visibility refer to the
404 same object if they are in the same shared object. Usually, hidden
405 visibility indicates that the symbol will not be placed into the
406 dynamic symbol table, so no other module (executable or shared
407 library) can reference it directly.
408 "``protected``" - Protected style
409 On ELF, protected visibility indicates that the symbol will be
410 placed in the dynamic symbol table, but that references within the
411 defining module will bind to the local symbol. That is, the symbol
412 cannot be overridden by another module.
417 LLVM IR allows you to specify name aliases for certain types. This can
418 make it easier to read the IR and make the IR more condensed
419 (particularly when recursive types are involved). An example of a name
424 %mytype = type { %mytype*, i32 }
426 You may give a name to any :ref:`type <typesystem>` except
427 ":ref:`void <t_void>`". Type name aliases may be used anywhere a type is
428 expected with the syntax "%mytype".
430 Note that type names are aliases for the structural type that they
431 indicate, and that you can therefore specify multiple names for the same
432 type. This often leads to confusing behavior when dumping out a .ll
433 file. Since LLVM IR uses structural typing, the name is not part of the
434 type. When printing out LLVM IR, the printer will pick *one name* to
435 render all types of a particular shape. This means that if you have code
436 where two different source types end up having the same LLVM type, that
437 the dumper will sometimes print the "wrong" or unexpected type. This is
438 an important design point and isn't going to change.
445 Global variables define regions of memory allocated at compilation time
446 instead of run-time. Global variables may optionally be initialized, may
447 have an explicit section to be placed in, and may have an optional
448 explicit alignment specified.
450 A variable may be defined as ``thread_local``, which means that it will
451 not be shared by threads (each thread will have a separated copy of the
452 variable). Not all targets support thread-local variables. Optionally, a
453 TLS model may be specified:
456 For variables that are only used within the current shared library.
458 For variables in modules that will not be loaded dynamically.
460 For variables defined in the executable and only used within it.
462 The models correspond to the ELF TLS models; see `ELF Handling For
463 Thread-Local Storage <http://people.redhat.com/drepper/tls.pdf>`_ for
464 more information on under which circumstances the different models may
465 be used. The target may choose a different TLS model if the specified
466 model is not supported, or if a better choice of model can be made.
468 A variable may be defined as a global ``constant``, which indicates that
469 the contents of the variable will **never** be modified (enabling better
470 optimization, allowing the global data to be placed in the read-only
471 section of an executable, etc). Note that variables that need runtime
472 initialization cannot be marked ``constant`` as there is a store to the
475 LLVM explicitly allows *declarations* of global variables to be marked
476 constant, even if the final definition of the global is not. This
477 capability can be used to enable slightly better optimization of the
478 program, but requires the language definition to guarantee that
479 optimizations based on the 'constantness' are valid for the translation
480 units that do not include the definition.
482 As SSA values, global variables define pointer values that are in scope
483 (i.e. they dominate) all basic blocks in the program. Global variables
484 always define a pointer to their "content" type because they describe a
485 region of memory, and all memory objects in LLVM are accessed through
488 Global variables can be marked with ``unnamed_addr`` which indicates
489 that the address is not significant, only the content. Constants marked
490 like this can be merged with other constants if they have the same
491 initializer. Note that a constant with significant address *can* be
492 merged with a ``unnamed_addr`` constant, the result being a constant
493 whose address is significant.
495 A global variable may be declared to reside in a target-specific
496 numbered address space. For targets that support them, address spaces
497 may affect how optimizations are performed and/or what target
498 instructions are used to access the variable. The default address space
499 is zero. The address space qualifier must precede any other attributes.
501 LLVM allows an explicit section to be specified for globals. If the
502 target supports it, it will emit globals to the section specified.
504 By default, global initializers are optimized by assuming that global
505 variables defined within the module are not modified from their
506 initial values before the start of the global initializer. This is
507 true even for variables potentially accessible from outside the
508 module, including those with external linkage or appearing in
509 ``@llvm.used``. This assumption may be suppressed by marking the
510 variable with ``externally_initialized``.
512 An explicit alignment may be specified for a global, which must be a
513 power of 2. If not present, or if the alignment is set to zero, the
514 alignment of the global is set by the target to whatever it feels
515 convenient. If an explicit alignment is specified, the global is forced
516 to have exactly that alignment. Targets and optimizers are not allowed
517 to over-align the global if the global has an assigned section. In this
518 case, the extra alignment could be observable: for example, code could
519 assume that the globals are densely packed in their section and try to
520 iterate over them as an array, alignment padding would break this
523 For example, the following defines a global in a numbered address space
524 with an initializer, section, and alignment:
528 @G = addrspace(5) constant float 1.0, section "foo", align 4
530 The following example defines a thread-local global with the
531 ``initialexec`` TLS model:
535 @G = thread_local(initialexec) global i32 0, align 4
537 .. _functionstructure:
542 LLVM function definitions consist of the "``define``" keyword, an
543 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
544 style <visibility>`, an optional :ref:`calling convention <callingconv>`,
545 an optional ``unnamed_addr`` attribute, a return type, an optional
546 :ref:`parameter attribute <paramattrs>` for the return type, a function
547 name, a (possibly empty) argument list (each with optional :ref:`parameter
548 attributes <paramattrs>`), optional :ref:`function attributes <fnattrs>`,
549 an optional section, an optional alignment, an optional :ref:`garbage
550 collector name <gc>`, an opening curly brace, a list of basic blocks,
551 and a closing curly brace.
553 LLVM function declarations consist of the "``declare``" keyword, an
554 optional :ref:`linkage type <linkage>`, an optional :ref:`visibility
555 style <visibility>`, an optional :ref:`calling convention <callingconv>`,
556 an optional ``unnamed_addr`` attribute, a return type, an optional
557 :ref:`parameter attribute <paramattrs>` for the return type, a function
558 name, a possibly empty list of arguments, an optional alignment, and an
559 optional :ref:`garbage collector name <gc>`.
561 A function definition contains a list of basic blocks, forming the CFG
562 (Control Flow Graph) for the function. Each basic block may optionally
563 start with a label (giving the basic block a symbol table entry),
564 contains a list of instructions, and ends with a
565 :ref:`terminator <terminators>` instruction (such as a branch or function
568 The first basic block in a function is special in two ways: it is
569 immediately executed on entrance to the function, and it is not allowed
570 to have predecessor basic blocks (i.e. there can not be any branches to
571 the entry block of a function). Because the block can have no
572 predecessors, it also cannot have any :ref:`PHI nodes <i_phi>`.
574 LLVM allows an explicit section to be specified for functions. If the
575 target supports it, it will emit functions to the section specified.
577 An explicit alignment may be specified for a function. If not present,
578 or if the alignment is set to zero, the alignment of the function is set
579 by the target to whatever it feels convenient. If an explicit alignment
580 is specified, the function is forced to have at least that much
581 alignment. All alignments must be a power of 2.
583 If the ``unnamed_addr`` attribute is given, the address is know to not
584 be significant and two identical functions can be merged.
588 define [linkage] [visibility]
590 <ResultType> @<FunctionName> ([argument list])
591 [fn Attrs] [section "name"] [align N]
597 Aliases act as "second name" for the aliasee value (which can be either
598 function, global variable, another alias or bitcast of global value).
599 Aliases may have an optional :ref:`linkage type <linkage>`, and an optional
600 :ref:`visibility style <visibility>`.
604 @<Name> = alias [Linkage] [Visibility] <AliaseeTy> @<Aliasee>
606 .. _namedmetadatastructure:
611 Named metadata is a collection of metadata. :ref:`Metadata
612 nodes <metadata>` (but not metadata strings) are the only valid
613 operands for a named metadata.
617 ; Some unnamed metadata nodes, which are referenced by the named metadata.
618 !0 = metadata !{metadata !"zero"}
619 !1 = metadata !{metadata !"one"}
620 !2 = metadata !{metadata !"two"}
622 !name = !{!0, !1, !2}
629 The return type and each parameter of a function type may have a set of
630 *parameter attributes* associated with them. Parameter attributes are
631 used to communicate additional information about the result or
632 parameters of a function. Parameter attributes are considered to be part
633 of the function, not of the function type, so functions with different
634 parameter attributes can have the same function type.
636 Parameter attributes are simple keywords that follow the type specified.
637 If multiple parameter attributes are needed, they are space separated.
642 declare i32 @printf(i8* noalias nocapture, ...)
643 declare i32 @atoi(i8 zeroext)
644 declare signext i8 @returns_signed_char()
646 Note that any attributes for the function result (``nounwind``,
647 ``readonly``) come immediately after the argument list.
649 Currently, only the following parameter attributes are defined:
652 This indicates to the code generator that the parameter or return
653 value should be zero-extended to the extent required by the target's
654 ABI (which is usually 32-bits, but is 8-bits for a i1 on x86-64) by
655 the caller (for a parameter) or the callee (for a return value).
657 This indicates to the code generator that the parameter or return
658 value should be sign-extended to the extent required by the target's
659 ABI (which is usually 32-bits) by the caller (for a parameter) or
660 the callee (for a return value).
662 This indicates that this parameter or return value should be treated
663 in a special target-dependent fashion during while emitting code for
664 a function call or return (usually, by putting it in a register as
665 opposed to memory, though some targets use it to distinguish between
666 two different kinds of registers). Use of this attribute is
669 This indicates that the pointer parameter should really be passed by
670 value to the function. The attribute implies that a hidden copy of
671 the pointee is made between the caller and the callee, so the callee
672 is unable to modify the value in the caller. This attribute is only
673 valid on LLVM pointer arguments. It is generally used to pass
674 structs and arrays by value, but is also valid on pointers to
675 scalars. The copy is considered to belong to the caller not the
676 callee (for example, ``readonly`` functions should not write to
677 ``byval`` parameters). This is not a valid attribute for return
680 The byval attribute also supports specifying an alignment with the
681 align attribute. It indicates the alignment of the stack slot to
682 form and the known alignment of the pointer specified to the call
683 site. If the alignment is not specified, then the code generator
684 makes a target-specific assumption.
687 This indicates that the pointer parameter specifies the address of a
688 structure that is the return value of the function in the source
689 program. This pointer must be guaranteed by the caller to be valid:
690 loads and stores to the structure may be assumed by the callee
691 not to trap and to be properly aligned. This may only be applied to
692 the first parameter. This is not a valid attribute for return
695 This indicates that pointer values `*based* <pointeraliasing>` on
696 the argument or return value do not alias pointer values which are
697 not *based* on it, ignoring certain "irrelevant" dependencies. For a
698 call to the parent function, dependencies between memory references
699 from before or after the call and from those during the call are
700 "irrelevant" to the ``noalias`` keyword for the arguments and return
701 value used in that call. The caller shares the responsibility with
702 the callee for ensuring that these requirements are met. For further
703 details, please see the discussion of the NoAlias response in `alias
704 analysis <AliasAnalysis.html#MustMayNo>`_.
706 Note that this definition of ``noalias`` is intentionally similar
707 to the definition of ``restrict`` in C99 for function arguments,
708 though it is slightly weaker.
710 For function return values, C99's ``restrict`` is not meaningful,
711 while LLVM's ``noalias`` is.
713 This indicates that the callee does not make any copies of the
714 pointer that outlive the callee itself. This is not a valid
715 attribute for return values.
720 This indicates that the pointer parameter can be excised using the
721 :ref:`trampoline intrinsics <int_trampoline>`. This is not a valid
722 attribute for return values.
726 Garbage Collector Names
727 -----------------------
729 Each function may specify a garbage collector name, which is simply a
734 define void @f() gc "name" { ... }
736 The compiler declares the supported values of *name*. Specifying a
737 collector which will cause the compiler to alter its output in order to
738 support the named garbage collection algorithm.
745 Attribute groups are groups of attributes that are referenced by objects within
746 the IR. They are important for keeping ``.ll`` files readable, because a lot of
747 functions will use the same set of attributes. In the degenerative case of a
748 ``.ll`` file that corresponds to a single ``.c`` file, the single attribute
749 group will capture the important command line flags used to build that file.
751 An attribute group is a module-level object. To use an attribute group, an
752 object references the attribute group's ID (e.g. ``#37``). An object may refer
753 to more than one attribute group. In that situation, the attributes from the
754 different groups are merged.
756 Here is an example of attribute groups for a function that should always be
757 inlined, has a stack alignment of 4, and which shouldn't use SSE instructions:
761 ; Target-independent attributes:
762 #0 = attributes { alwaysinline alignstack=4 }
764 ; Target-dependent attributes:
765 #1 = attributes { "no-sse" }
767 ; Function @f has attributes: alwaysinline, alignstack=4, and "no-sse".
768 define void @f() #0 #1 { ... }
775 Function attributes are set to communicate additional information about
776 a function. Function attributes are considered to be part of the
777 function, not of the function type, so functions with different function
778 attributes can have the same function type.
780 Function attributes are simple keywords that follow the type specified.
781 If multiple attributes are needed, they are space separated. For
786 define void @f() noinline { ... }
787 define void @f() alwaysinline { ... }
788 define void @f() alwaysinline optsize { ... }
789 define void @f() optsize { ... }
792 This attribute indicates that the address safety analysis is enabled
795 This attribute indicates that, when emitting the prologue and
796 epilogue, the backend should forcibly align the stack pointer.
797 Specify the desired alignment, which must be a power of two, in
800 This attribute indicates that the inliner should attempt to inline
801 this function into callers whenever possible, ignoring any active
802 inlining size threshold for this caller.
804 This attribute suppresses lazy symbol binding for the function. This
805 may make calls to the function faster, at the cost of extra program
806 startup time if the function is not called during program startup.
808 This attribute indicates that the source code contained a hint that
809 inlining this function is desirable (such as the "inline" keyword in
810 C/C++). It is just a hint; it imposes no requirements on the
813 This attribute disables prologue / epilogue emission for the
814 function. This can have very system-specific consequences.
816 This attribute indicates that calls to the function cannot be
817 duplicated. A call to a ``noduplicate`` function may be moved
818 within its parent function, but may not be duplicated within
821 A function containing a ``noduplicate`` call may still
822 be an inlining candidate, provided that the call is not
823 duplicated by inlining. That implies that the function has
824 internal linkage and only has one call site, so the original
825 call is dead after inlining.
827 This attributes disables implicit floating point instructions.
829 This attribute indicates that the inliner should never inline this
830 function in any situation. This attribute may not be used together
831 with the ``alwaysinline`` attribute.
833 This attribute indicates that the code generator should not use a
834 red zone, even if the target-specific ABI normally permits it.
836 This function attribute indicates that the function never returns
837 normally. This produces undefined behavior at runtime if the
838 function ever does dynamically return.
840 This function attribute indicates that the function never returns
841 with an unwind or exceptional control flow. If the function does
842 unwind, its runtime behavior is undefined.
844 This attribute suggests that optimization passes and code generator
845 passes make choices that keep the code size of this function low,
846 and otherwise do optimizations specifically to reduce code size.
848 This attribute indicates that the function computes its result (or
849 decides to unwind an exception) based strictly on its arguments,
850 without dereferencing any pointer arguments or otherwise accessing
851 any mutable state (e.g. memory, control registers, etc) visible to
852 caller functions. It does not write through any pointer arguments
853 (including ``byval`` arguments) and never changes any state visible
854 to callers. This means that it cannot unwind exceptions by calling
855 the ``C++`` exception throwing methods.
857 This attribute indicates that the function does not write through
858 any pointer arguments (including ``byval`` arguments) or otherwise
859 modify any state (e.g. memory, control registers, etc) visible to
860 caller functions. It may dereference pointer arguments and read
861 state that may be set in the caller. A readonly function always
862 returns the same value (or unwinds an exception identically) when
863 called with the same set of arguments and global state. It cannot
864 unwind an exception by calling the ``C++`` exception throwing
867 This attribute indicates that this function can return twice. The C
868 ``setjmp`` is an example of such a function. The compiler disables
869 some optimizations (like tail calls) in the caller of these
872 This attribute indicates that the function should emit a stack
873 smashing protector. It is in the form of a "canary" --- a random value
874 placed on the stack before the local variables that's checked upon
875 return from the function to see if it has been overwritten. A
876 heuristic is used to determine if a function needs stack protectors
877 or not. The heuristic used will enable protectors for functions with:
879 - Character arrays larger than ``ssp-buffer-size`` (default 8).
880 - Aggregates containing character arrays larger than ``ssp-buffer-size``.
881 - Calls to alloca() with variable sizes or constant sizes greater than
884 If a function that has an ``ssp`` attribute is inlined into a
885 function that doesn't have an ``ssp`` attribute, then the resulting
886 function will have an ``ssp`` attribute.
888 This attribute indicates that the function should *always* emit a
889 stack smashing protector. This overrides the ``ssp`` function
892 If a function that has an ``sspreq`` attribute is inlined into a
893 function that doesn't have an ``sspreq`` attribute or which has an
894 ``ssp`` or ``sspstrong`` attribute, then the resulting function will have
895 an ``sspreq`` attribute.
897 This attribute indicates that the function should emit a stack smashing
898 protector. This attribute causes a strong heuristic to be used when
899 determining if a function needs stack protectors. The strong heuristic
900 will enable protectors for functions with:
902 - Arrays of any size and type
903 - Aggregates containing an array of any size and type.
905 - Local variables that have had their address taken.
907 This overrides the ``ssp`` function attribute.
909 If a function that has an ``sspstrong`` attribute is inlined into a
910 function that doesn't have an ``sspstrong`` attribute, then the
911 resulting function will have an ``sspstrong`` attribute.
913 This attribute indicates that the thread safety analysis is enabled
915 ``uninitialized_checks``
916 This attribute indicates that the checks for uses of uninitialized
919 This attribute indicates that the ABI being targeted requires that
920 an unwind table entry be produce for this function even if we can
921 show that no exceptions passes by it. This is normally the case for
922 the ELF x86-64 abi, but it can be disabled for some compilation
927 Module-Level Inline Assembly
928 ----------------------------
930 Modules may contain "module-level inline asm" blocks, which corresponds
931 to the GCC "file scope inline asm" blocks. These blocks are internally
932 concatenated by LLVM and treated as a single unit, but may be separated
933 in the ``.ll`` file if desired. The syntax is very simple:
937 module asm "inline asm code goes here"
938 module asm "more can go here"
940 The strings can contain any character by escaping non-printable
941 characters. The escape sequence used is simply "\\xx" where "xx" is the
942 two digit hex code for the number.
944 The inline asm code is simply printed to the machine code .s file when
945 assembly code is generated.
950 A module may specify a target specific data layout string that specifies
951 how data is to be laid out in memory. The syntax for the data layout is
956 target datalayout = "layout specification"
958 The *layout specification* consists of a list of specifications
959 separated by the minus sign character ('-'). Each specification starts
960 with a letter and may include other information after the letter to
961 define some aspect of the data layout. The specifications accepted are
965 Specifies that the target lays out data in big-endian form. That is,
966 the bits with the most significance have the lowest address
969 Specifies that the target lays out data in little-endian form. That
970 is, the bits with the least significance have the lowest address
973 Specifies the natural alignment of the stack in bits. Alignment
974 promotion of stack variables is limited to the natural stack
975 alignment to avoid dynamic stack realignment. The stack alignment
976 must be a multiple of 8-bits. If omitted, the natural stack
977 alignment defaults to "unspecified", which does not prevent any
978 alignment promotions.
979 ``p[n]:<size>:<abi>:<pref>``
980 This specifies the *size* of a pointer and its ``<abi>`` and
981 ``<pref>``\erred alignments for address space ``n``. All sizes are in
982 bits. Specifying the ``<pref>`` alignment is optional. If omitted, the
983 preceding ``:`` should be omitted too. The address space, ``n`` is
984 optional, and if not specified, denotes the default address space 0.
985 The value of ``n`` must be in the range [1,2^23).
986 ``i<size>:<abi>:<pref>``
987 This specifies the alignment for an integer type of a given bit
988 ``<size>``. The value of ``<size>`` must be in the range [1,2^23).
989 ``v<size>:<abi>:<pref>``
990 This specifies the alignment for a vector type of a given bit
992 ``f<size>:<abi>:<pref>``
993 This specifies the alignment for a floating point type of a given bit
994 ``<size>``. Only values of ``<size>`` that are supported by the target
995 will work. 32 (float) and 64 (double) are supported on all targets; 80
996 or 128 (different flavors of long double) are also supported on some
998 ``a<size>:<abi>:<pref>``
999 This specifies the alignment for an aggregate type of a given bit
1001 ``s<size>:<abi>:<pref>``
1002 This specifies the alignment for a stack object of a given bit
1004 ``n<size1>:<size2>:<size3>...``
1005 This specifies a set of native integer widths for the target CPU in
1006 bits. For example, it might contain ``n32`` for 32-bit PowerPC,
1007 ``n32:64`` for PowerPC 64, or ``n8:16:32:64`` for X86-64. Elements of
1008 this set are considered to support most general arithmetic operations
1011 When constructing the data layout for a given target, LLVM starts with a
1012 default set of specifications which are then (possibly) overridden by
1013 the specifications in the ``datalayout`` keyword. The default
1014 specifications are given in this list:
1016 - ``E`` - big endian
1017 - ``p:64:64:64`` - 64-bit pointers with 64-bit alignment
1018 - ``S0`` - natural stack alignment is unspecified
1019 - ``i1:8:8`` - i1 is 8-bit (byte) aligned
1020 - ``i8:8:8`` - i8 is 8-bit (byte) aligned
1021 - ``i16:16:16`` - i16 is 16-bit aligned
1022 - ``i32:32:32`` - i32 is 32-bit aligned
1023 - ``i64:32:64`` - i64 has ABI alignment of 32-bits but preferred
1024 alignment of 64-bits
1025 - ``f16:16:16`` - half is 16-bit aligned
1026 - ``f32:32:32`` - float is 32-bit aligned
1027 - ``f64:64:64`` - double is 64-bit aligned
1028 - ``f128:128:128`` - quad is 128-bit aligned
1029 - ``v64:64:64`` - 64-bit vector is 64-bit aligned
1030 - ``v128:128:128`` - 128-bit vector is 128-bit aligned
1031 - ``a0:0:64`` - aggregates are 64-bit aligned
1033 When LLVM is determining the alignment for a given type, it uses the
1036 #. If the type sought is an exact match for one of the specifications,
1037 that specification is used.
1038 #. If no match is found, and the type sought is an integer type, then
1039 the smallest integer type that is larger than the bitwidth of the
1040 sought type is used. If none of the specifications are larger than
1041 the bitwidth then the largest integer type is used. For example,
1042 given the default specifications above, the i7 type will use the
1043 alignment of i8 (next largest) while both i65 and i256 will use the
1044 alignment of i64 (largest specified).
1045 #. If no match is found, and the type sought is a vector type, then the
1046 largest vector type that is smaller than the sought vector type will
1047 be used as a fall back. This happens because <128 x double> can be
1048 implemented in terms of 64 <2 x double>, for example.
1050 The function of the data layout string may not be what you expect.
1051 Notably, this is not a specification from the frontend of what alignment
1052 the code generator should use.
1054 Instead, if specified, the target data layout is required to match what
1055 the ultimate *code generator* expects. This string is used by the
1056 mid-level optimizers to improve code, and this only works if it matches
1057 what the ultimate code generator uses. If you would like to generate IR
1058 that does not embed this target-specific detail into the IR, then you
1059 don't have to specify the string. This will disable some optimizations
1060 that require precise layout information, but this also prevents those
1061 optimizations from introducing target specificity into the IR.
1063 .. _pointeraliasing:
1065 Pointer Aliasing Rules
1066 ----------------------
1068 Any memory access must be done through a pointer value associated with
1069 an address range of the memory access, otherwise the behavior is
1070 undefined. Pointer values are associated with address ranges according
1071 to the following rules:
1073 - A pointer value is associated with the addresses associated with any
1074 value it is *based* on.
1075 - An address of a global variable is associated with the address range
1076 of the variable's storage.
1077 - The result value of an allocation instruction is associated with the
1078 address range of the allocated storage.
1079 - A null pointer in the default address-space is associated with no
1081 - An integer constant other than zero or a pointer value returned from
1082 a function not defined within LLVM may be associated with address
1083 ranges allocated through mechanisms other than those provided by
1084 LLVM. Such ranges shall not overlap with any ranges of addresses
1085 allocated by mechanisms provided by LLVM.
1087 A pointer value is *based* on another pointer value according to the
1090 - A pointer value formed from a ``getelementptr`` operation is *based*
1091 on the first operand of the ``getelementptr``.
1092 - The result value of a ``bitcast`` is *based* on the operand of the
1094 - A pointer value formed by an ``inttoptr`` is *based* on all pointer
1095 values that contribute (directly or indirectly) to the computation of
1096 the pointer's value.
1097 - The "*based* on" relationship is transitive.
1099 Note that this definition of *"based"* is intentionally similar to the
1100 definition of *"based"* in C99, though it is slightly weaker.
1102 LLVM IR does not associate types with memory. The result type of a
1103 ``load`` merely indicates the size and alignment of the memory from
1104 which to load, as well as the interpretation of the value. The first
1105 operand type of a ``store`` similarly only indicates the size and
1106 alignment of the store.
1108 Consequently, type-based alias analysis, aka TBAA, aka
1109 ``-fstrict-aliasing``, is not applicable to general unadorned LLVM IR.
1110 :ref:`Metadata <metadata>` may be used to encode additional information
1111 which specialized optimization passes may use to implement type-based
1116 Volatile Memory Accesses
1117 ------------------------
1119 Certain memory accesses, such as :ref:`load <i_load>`'s,
1120 :ref:`store <i_store>`'s, and :ref:`llvm.memcpy <int_memcpy>`'s may be
1121 marked ``volatile``. The optimizers must not change the number of
1122 volatile operations or change their order of execution relative to other
1123 volatile operations. The optimizers *may* change the order of volatile
1124 operations relative to non-volatile operations. This is not Java's
1125 "volatile" and has no cross-thread synchronization behavior.
1127 IR-level volatile loads and stores cannot safely be optimized into
1128 llvm.memcpy or llvm.memmove intrinsics even when those intrinsics are
1129 flagged volatile. Likewise, the backend should never split or merge
1130 target-legal volatile load/store instructions.
1132 .. admonition:: Rationale
1134 Platforms may rely on volatile loads and stores of natively supported
1135 data width to be executed as single instruction. For example, in C
1136 this holds for an l-value of volatile primitive type with native
1137 hardware support, but not necessarily for aggregate types. The
1138 frontend upholds these expectations, which are intentionally
1139 unspecified in the IR. The rules above ensure that IR transformation
1140 do not violate the frontend's contract with the language.
1144 Memory Model for Concurrent Operations
1145 --------------------------------------
1147 The LLVM IR does not define any way to start parallel threads of
1148 execution or to register signal handlers. Nonetheless, there are
1149 platform-specific ways to create them, and we define LLVM IR's behavior
1150 in their presence. This model is inspired by the C++0x memory model.
1152 For a more informal introduction to this model, see the :doc:`Atomics`.
1154 We define a *happens-before* partial order as the least partial order
1157 - Is a superset of single-thread program order, and
1158 - When a *synchronizes-with* ``b``, includes an edge from ``a`` to
1159 ``b``. *Synchronizes-with* pairs are introduced by platform-specific
1160 techniques, like pthread locks, thread creation, thread joining,
1161 etc., and by atomic instructions. (See also :ref:`Atomic Memory Ordering
1162 Constraints <ordering>`).
1164 Note that program order does not introduce *happens-before* edges
1165 between a thread and signals executing inside that thread.
1167 Every (defined) read operation (load instructions, memcpy, atomic
1168 loads/read-modify-writes, etc.) R reads a series of bytes written by
1169 (defined) write operations (store instructions, atomic
1170 stores/read-modify-writes, memcpy, etc.). For the purposes of this
1171 section, initialized globals are considered to have a write of the
1172 initializer which is atomic and happens before any other read or write
1173 of the memory in question. For each byte of a read R, R\ :sub:`byte`
1174 may see any write to the same byte, except:
1176 - If write\ :sub:`1` happens before write\ :sub:`2`, and
1177 write\ :sub:`2` happens before R\ :sub:`byte`, then
1178 R\ :sub:`byte` does not see write\ :sub:`1`.
1179 - If R\ :sub:`byte` happens before write\ :sub:`3`, then
1180 R\ :sub:`byte` does not see write\ :sub:`3`.
1182 Given that definition, R\ :sub:`byte` is defined as follows:
1184 - If R is volatile, the result is target-dependent. (Volatile is
1185 supposed to give guarantees which can support ``sig_atomic_t`` in
1186 C/C++, and may be used for accesses to addresses which do not behave
1187 like normal memory. It does not generally provide cross-thread
1189 - Otherwise, if there is no write to the same byte that happens before
1190 R\ :sub:`byte`, R\ :sub:`byte` returns ``undef`` for that byte.
1191 - Otherwise, if R\ :sub:`byte` may see exactly one write,
1192 R\ :sub:`byte` returns the value written by that write.
1193 - Otherwise, if R is atomic, and all the writes R\ :sub:`byte` may
1194 see are atomic, it chooses one of the values written. See the :ref:`Atomic
1195 Memory Ordering Constraints <ordering>` section for additional
1196 constraints on how the choice is made.
1197 - Otherwise R\ :sub:`byte` returns ``undef``.
1199 R returns the value composed of the series of bytes it read. This
1200 implies that some bytes within the value may be ``undef`` **without**
1201 the entire value being ``undef``. Note that this only defines the
1202 semantics of the operation; it doesn't mean that targets will emit more
1203 than one instruction to read the series of bytes.
1205 Note that in cases where none of the atomic intrinsics are used, this
1206 model places only one restriction on IR transformations on top of what
1207 is required for single-threaded execution: introducing a store to a byte
1208 which might not otherwise be stored is not allowed in general.
1209 (Specifically, in the case where another thread might write to and read
1210 from an address, introducing a store can change a load that may see
1211 exactly one write into a load that may see multiple writes.)
1215 Atomic Memory Ordering Constraints
1216 ----------------------------------
1218 Atomic instructions (:ref:`cmpxchg <i_cmpxchg>`,
1219 :ref:`atomicrmw <i_atomicrmw>`, :ref:`fence <i_fence>`,
1220 :ref:`atomic load <i_load>`, and :ref:`atomic store <i_store>`) take
1221 an ordering parameter that determines which other atomic instructions on
1222 the same address they *synchronize with*. These semantics are borrowed
1223 from Java and C++0x, but are somewhat more colloquial. If these
1224 descriptions aren't precise enough, check those specs (see spec
1225 references in the :doc:`atomics guide <Atomics>`).
1226 :ref:`fence <i_fence>` instructions treat these orderings somewhat
1227 differently since they don't take an address. See that instruction's
1228 documentation for details.
1230 For a simpler introduction to the ordering constraints, see the
1234 The set of values that can be read is governed by the happens-before
1235 partial order. A value cannot be read unless some operation wrote
1236 it. This is intended to provide a guarantee strong enough to model
1237 Java's non-volatile shared variables. This ordering cannot be
1238 specified for read-modify-write operations; it is not strong enough
1239 to make them atomic in any interesting way.
1241 In addition to the guarantees of ``unordered``, there is a single
1242 total order for modifications by ``monotonic`` operations on each
1243 address. All modification orders must be compatible with the
1244 happens-before order. There is no guarantee that the modification
1245 orders can be combined to a global total order for the whole program
1246 (and this often will not be possible). The read in an atomic
1247 read-modify-write operation (:ref:`cmpxchg <i_cmpxchg>` and
1248 :ref:`atomicrmw <i_atomicrmw>`) reads the value in the modification
1249 order immediately before the value it writes. If one atomic read
1250 happens before another atomic read of the same address, the later
1251 read must see the same value or a later value in the address's
1252 modification order. This disallows reordering of ``monotonic`` (or
1253 stronger) operations on the same address. If an address is written
1254 ``monotonic``-ally by one thread, and other threads ``monotonic``-ally
1255 read that address repeatedly, the other threads must eventually see
1256 the write. This corresponds to the C++0x/C1x
1257 ``memory_order_relaxed``.
1259 In addition to the guarantees of ``monotonic``, a
1260 *synchronizes-with* edge may be formed with a ``release`` operation.
1261 This is intended to model C++'s ``memory_order_acquire``.
1263 In addition to the guarantees of ``monotonic``, if this operation
1264 writes a value which is subsequently read by an ``acquire``
1265 operation, it *synchronizes-with* that operation. (This isn't a
1266 complete description; see the C++0x definition of a release
1267 sequence.) This corresponds to the C++0x/C1x
1268 ``memory_order_release``.
1269 ``acq_rel`` (acquire+release)
1270 Acts as both an ``acquire`` and ``release`` operation on its
1271 address. This corresponds to the C++0x/C1x ``memory_order_acq_rel``.
1272 ``seq_cst`` (sequentially consistent)
1273 In addition to the guarantees of ``acq_rel`` (``acquire`` for an
1274 operation which only reads, ``release`` for an operation which only
1275 writes), there is a global total order on all
1276 sequentially-consistent operations on all addresses, which is
1277 consistent with the *happens-before* partial order and with the
1278 modification orders of all the affected addresses. Each
1279 sequentially-consistent read sees the last preceding write to the
1280 same address in this global order. This corresponds to the C++0x/C1x
1281 ``memory_order_seq_cst`` and Java volatile.
1285 If an atomic operation is marked ``singlethread``, it only *synchronizes
1286 with* or participates in modification and seq\_cst total orderings with
1287 other operations running in the same thread (for example, in signal
1295 LLVM IR floating-point binary ops (:ref:`fadd <i_fadd>`,
1296 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
1297 :ref:`frem <i_frem>`) have the following flags that can set to enable
1298 otherwise unsafe floating point operations
1301 No NaNs - Allow optimizations to assume the arguments and result are not
1302 NaN. Such optimizations are required to retain defined behavior over
1303 NaNs, but the value of the result is undefined.
1306 No Infs - Allow optimizations to assume the arguments and result are not
1307 +/-Inf. Such optimizations are required to retain defined behavior over
1308 +/-Inf, but the value of the result is undefined.
1311 No Signed Zeros - Allow optimizations to treat the sign of a zero
1312 argument or result as insignificant.
1315 Allow Reciprocal - Allow optimizations to use the reciprocal of an
1316 argument rather than perform division.
1319 Fast - Allow algebraically equivalent transformations that may
1320 dramatically change results in floating point (e.g. reassociate). This
1321 flag implies all the others.
1328 The LLVM type system is one of the most important features of the
1329 intermediate representation. Being typed enables a number of
1330 optimizations to be performed on the intermediate representation
1331 directly, without having to do extra analyses on the side before the
1332 transformation. A strong type system makes it easier to read the
1333 generated code and enables novel analyses and transformations that are
1334 not feasible to perform on normal three address code representations.
1336 Type Classifications
1337 --------------------
1339 The types fall into a few useful classifications:
1348 * - :ref:`integer <t_integer>`
1349 - ``i1``, ``i2``, ``i3``, ... ``i8``, ... ``i16``, ... ``i32``, ...
1352 * - :ref:`floating point <t_floating>`
1353 - ``half``, ``float``, ``double``, ``x86_fp80``, ``fp128``,
1361 - :ref:`integer <t_integer>`, :ref:`floating point <t_floating>`,
1362 :ref:`pointer <t_pointer>`, :ref:`vector <t_vector>`,
1363 :ref:`structure <t_struct>`, :ref:`array <t_array>`,
1364 :ref:`label <t_label>`, :ref:`metadata <t_metadata>`.
1366 * - :ref:`primitive <t_primitive>`
1367 - :ref:`label <t_label>`,
1368 :ref:`void <t_void>`,
1369 :ref:`integer <t_integer>`,
1370 :ref:`floating point <t_floating>`,
1371 :ref:`x86mmx <t_x86mmx>`,
1372 :ref:`metadata <t_metadata>`.
1374 * - :ref:`derived <t_derived>`
1375 - :ref:`array <t_array>`,
1376 :ref:`function <t_function>`,
1377 :ref:`pointer <t_pointer>`,
1378 :ref:`structure <t_struct>`,
1379 :ref:`vector <t_vector>`,
1380 :ref:`opaque <t_opaque>`.
1382 The :ref:`first class <t_firstclass>` types are perhaps the most important.
1383 Values of these types are the only ones which can be produced by
1391 The primitive types are the fundamental building blocks of the LLVM
1402 The integer type is a very simple type that simply specifies an
1403 arbitrary bit width for the integer type desired. Any bit width from 1
1404 bit to 2\ :sup:`23`\ -1 (about 8 million) can be specified.
1413 The number of bits the integer will occupy is specified by the ``N``
1419 +----------------+------------------------------------------------+
1420 | ``i1`` | a single-bit integer. |
1421 +----------------+------------------------------------------------+
1422 | ``i32`` | a 32-bit integer. |
1423 +----------------+------------------------------------------------+
1424 | ``i1942652`` | a really big integer of over 1 million bits. |
1425 +----------------+------------------------------------------------+
1429 Floating Point Types
1430 ^^^^^^^^^^^^^^^^^^^^
1439 - 16-bit floating point value
1442 - 32-bit floating point value
1445 - 64-bit floating point value
1448 - 128-bit floating point value (112-bit mantissa)
1451 - 80-bit floating point value (X87)
1454 - 128-bit floating point value (two 64-bits)
1464 The x86mmx type represents a value held in an MMX register on an x86
1465 machine. The operations allowed on it are quite limited: parameters and
1466 return values, load and store, and bitcast. User-specified MMX
1467 instructions are represented as intrinsic or asm calls with arguments
1468 and/or results of this type. There are no arrays, vectors or constants
1486 The void type does not represent any value and has no size.
1503 The label type represents code labels.
1520 The metadata type represents embedded metadata. No derived types may be
1521 created from metadata except for :ref:`function <t_function>` arguments.
1535 The real power in LLVM comes from the derived types in the system. This
1536 is what allows a programmer to represent arrays, functions, pointers,
1537 and other useful types. Each of these types contain one or more element
1538 types which may be a primitive type, or another derived type. For
1539 example, it is possible to have a two dimensional array, using an array
1540 as the element type of another array.
1547 Aggregate Types are a subset of derived types that can contain multiple
1548 member types. :ref:`Arrays <t_array>` and :ref:`structs <t_struct>` are
1549 aggregate types. :ref:`Vectors <t_vector>` are not considered to be
1560 The array type is a very simple derived type that arranges elements
1561 sequentially in memory. The array type requires a size (number of
1562 elements) and an underlying data type.
1569 [<# elements> x <elementtype>]
1571 The number of elements is a constant integer value; ``elementtype`` may
1572 be any type with a size.
1577 +------------------+--------------------------------------+
1578 | ``[40 x i32]`` | Array of 40 32-bit integer values. |
1579 +------------------+--------------------------------------+
1580 | ``[41 x i32]`` | Array of 41 32-bit integer values. |
1581 +------------------+--------------------------------------+
1582 | ``[4 x i8]`` | Array of 4 8-bit integer values. |
1583 +------------------+--------------------------------------+
1585 Here are some examples of multidimensional arrays:
1587 +-----------------------------+----------------------------------------------------------+
1588 | ``[3 x [4 x i32]]`` | 3x4 array of 32-bit integer values. |
1589 +-----------------------------+----------------------------------------------------------+
1590 | ``[12 x [10 x float]]`` | 12x10 array of single precision floating point values. |
1591 +-----------------------------+----------------------------------------------------------+
1592 | ``[2 x [3 x [4 x i16]]]`` | 2x3x4 array of 16-bit integer values. |
1593 +-----------------------------+----------------------------------------------------------+
1595 There is no restriction on indexing beyond the end of the array implied
1596 by a static type (though there are restrictions on indexing beyond the
1597 bounds of an allocated object in some cases). This means that
1598 single-dimension 'variable sized array' addressing can be implemented in
1599 LLVM with a zero length array type. An implementation of 'pascal style
1600 arrays' in LLVM could use the type "``{ i32, [0 x float]}``", for
1611 The function type can be thought of as a function signature. It consists
1612 of a return type and a list of formal parameter types. The return type
1613 of a function type is a first class type or a void type.
1620 <returntype> (<parameter list>)
1622 ...where '``<parameter list>``' is a comma-separated list of type
1623 specifiers. Optionally, the parameter list may include a type ``...``,
1624 which indicates that the function takes a variable number of arguments.
1625 Variable argument functions can access their arguments with the
1626 :ref:`variable argument handling intrinsic <int_varargs>` functions.
1627 '``<returntype>``' is any type except :ref:`label <t_label>`.
1632 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1633 | ``i32 (i32)`` | function taking an ``i32``, returning an ``i32`` |
1634 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1635 | ``float (i16, i32 *) *`` | :ref:`Pointer <t_pointer>` to a function that takes an ``i16`` and a :ref:`pointer <t_pointer>` to ``i32``, returning ``float``. |
1636 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1637 | ``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. |
1638 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1639 | ``{i32, i32} (i32)`` | A function taking an ``i32``, returning a :ref:`structure <t_struct>` containing two ``i32`` values |
1640 +---------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1650 The structure type is used to represent a collection of data members
1651 together in memory. The elements of a structure may be any type that has
1654 Structures in memory are accessed using '``load``' and '``store``' by
1655 getting a pointer to a field with the '``getelementptr``' instruction.
1656 Structures in registers are accessed using the '``extractvalue``' and
1657 '``insertvalue``' instructions.
1659 Structures may optionally be "packed" structures, which indicate that
1660 the alignment of the struct is one byte, and that there is no padding
1661 between the elements. In non-packed structs, padding between field types
1662 is inserted as defined by the DataLayout string in the module, which is
1663 required to match what the underlying code generator expects.
1665 Structures can either be "literal" or "identified". A literal structure
1666 is defined inline with other types (e.g. ``{i32, i32}*``) whereas
1667 identified types are always defined at the top level with a name.
1668 Literal types are uniqued by their contents and can never be recursive
1669 or opaque since there is no way to write one. Identified types can be
1670 recursive, can be opaqued, and are never uniqued.
1677 %T1 = type { <type list> } ; Identified normal struct type
1678 %T2 = type <{ <type list> }> ; Identified packed struct type
1683 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1684 | ``{ i32, i32, i32 }`` | A triple of three ``i32`` values |
1685 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1686 | ``{ 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``. |
1687 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1688 | ``<{ i8, i32 }>`` | A packed struct known to be 5 bytes in size. |
1689 +------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1693 Opaque Structure Types
1694 ^^^^^^^^^^^^^^^^^^^^^^
1699 Opaque structure types are used to represent named structure types that
1700 do not have a body specified. This corresponds (for example) to the C
1701 notion of a forward declared structure.
1714 +--------------+-------------------+
1715 | ``opaque`` | An opaque type. |
1716 +--------------+-------------------+
1726 The pointer type is used to specify memory locations. Pointers are
1727 commonly used to reference objects in memory.
1729 Pointer types may have an optional address space attribute defining the
1730 numbered address space where the pointed-to object resides. The default
1731 address space is number zero. The semantics of non-zero address spaces
1732 are target-specific.
1734 Note that LLVM does not permit pointers to void (``void*``) nor does it
1735 permit pointers to labels (``label*``). Use ``i8*`` instead.
1747 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1748 | ``[4 x i32]*`` | A :ref:`pointer <t_pointer>` to :ref:`array <t_array>` of four ``i32`` values. |
1749 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1750 | ``i32 (i32*) *`` | A :ref:`pointer <t_pointer>` to a :ref:`function <t_function>` that takes an ``i32*``, returning an ``i32``. |
1751 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1752 | ``i32 addrspace(5)*`` | A :ref:`pointer <t_pointer>` to an ``i32`` value that resides in address space #5. |
1753 +-------------------------+--------------------------------------------------------------------------------------------------------------+
1763 A vector type is a simple derived type that represents a vector of
1764 elements. Vector types are used when multiple primitive data are
1765 operated in parallel using a single instruction (SIMD). A vector type
1766 requires a size (number of elements) and an underlying primitive data
1767 type. Vector types are considered :ref:`first class <t_firstclass>`.
1774 < <# elements> x <elementtype> >
1776 The number of elements is a constant integer value larger than 0;
1777 elementtype may be any integer or floating point type, or a pointer to
1778 these types. Vectors of size zero are not allowed.
1783 +-------------------+--------------------------------------------------+
1784 | ``<4 x i32>`` | Vector of 4 32-bit integer values. |
1785 +-------------------+--------------------------------------------------+
1786 | ``<8 x float>`` | Vector of 8 32-bit floating-point values. |
1787 +-------------------+--------------------------------------------------+
1788 | ``<2 x i64>`` | Vector of 2 64-bit integer values. |
1789 +-------------------+--------------------------------------------------+
1790 | ``<4 x i64*>`` | Vector of 4 pointers to 64-bit integer values. |
1791 +-------------------+--------------------------------------------------+
1796 LLVM has several different basic types of constants. This section
1797 describes them all and their syntax.
1802 **Boolean constants**
1803 The two strings '``true``' and '``false``' are both valid constants
1805 **Integer constants**
1806 Standard integers (such as '4') are constants of the
1807 :ref:`integer <t_integer>` type. Negative numbers may be used with
1809 **Floating point constants**
1810 Floating point constants use standard decimal notation (e.g.
1811 123.421), exponential notation (e.g. 1.23421e+2), or a more precise
1812 hexadecimal notation (see below). The assembler requires the exact
1813 decimal value of a floating-point constant. For example, the
1814 assembler accepts 1.25 but rejects 1.3 because 1.3 is a repeating
1815 decimal in binary. Floating point constants must have a :ref:`floating
1816 point <t_floating>` type.
1817 **Null pointer constants**
1818 The identifier '``null``' is recognized as a null pointer constant
1819 and must be of :ref:`pointer type <t_pointer>`.
1821 The one non-intuitive notation for constants is the hexadecimal form of
1822 floating point constants. For example, the form
1823 '``double 0x432ff973cafa8000``' is equivalent to (but harder to read
1824 than) '``double 4.5e+15``'. The only time hexadecimal floating point
1825 constants are required (and the only time that they are generated by the
1826 disassembler) is when a floating point constant must be emitted but it
1827 cannot be represented as a decimal floating point number in a reasonable
1828 number of digits. For example, NaN's, infinities, and other special
1829 values are represented in their IEEE hexadecimal format so that assembly
1830 and disassembly do not cause any bits to change in the constants.
1832 When using the hexadecimal form, constants of types half, float, and
1833 double are represented using the 16-digit form shown above (which
1834 matches the IEEE754 representation for double); half and float values
1835 must, however, be exactly representable as IEEE 754 half and single
1836 precision, respectively. Hexadecimal format is always used for long
1837 double, and there are three forms of long double. The 80-bit format used
1838 by x86 is represented as ``0xK`` followed by 20 hexadecimal digits. The
1839 128-bit format used by PowerPC (two adjacent doubles) is represented by
1840 ``0xM`` followed by 32 hexadecimal digits. The IEEE 128-bit format is
1841 represented by ``0xL`` followed by 32 hexadecimal digits; no currently
1842 supported target uses this format. Long doubles will only work if they
1843 match the long double format on your target. The IEEE 16-bit format
1844 (half precision) is represented by ``0xH`` followed by 4 hexadecimal
1845 digits. All hexadecimal formats are big-endian (sign bit at the left).
1847 There are no constants of type x86mmx.
1852 Complex constants are a (potentially recursive) combination of simple
1853 constants and smaller complex constants.
1855 **Structure constants**
1856 Structure constants are represented with notation similar to
1857 structure type definitions (a comma separated list of elements,
1858 surrounded by braces (``{}``)). For example:
1859 "``{ i32 4, float 17.0, i32* @G }``", where "``@G``" is declared as
1860 "``@G = external global i32``". Structure constants must have
1861 :ref:`structure type <t_struct>`, and the number and types of elements
1862 must match those specified by the type.
1864 Array constants are represented with notation similar to array type
1865 definitions (a comma separated list of elements, surrounded by
1866 square brackets (``[]``)). For example:
1867 "``[ i32 42, i32 11, i32 74 ]``". Array constants must have
1868 :ref:`array type <t_array>`, and the number and types of elements must
1869 match those specified by the type.
1870 **Vector constants**
1871 Vector constants are represented with notation similar to vector
1872 type definitions (a comma separated list of elements, surrounded by
1873 less-than/greater-than's (``<>``)). For example:
1874 "``< i32 42, i32 11, i32 74, i32 100 >``". Vector constants
1875 must have :ref:`vector type <t_vector>`, and the number and types of
1876 elements must match those specified by the type.
1877 **Zero initialization**
1878 The string '``zeroinitializer``' can be used to zero initialize a
1879 value to zero of *any* type, including scalar and
1880 :ref:`aggregate <t_aggregate>` types. This is often used to avoid
1881 having to print large zero initializers (e.g. for large arrays) and
1882 is always exactly equivalent to using explicit zero initializers.
1884 A metadata node is a structure-like constant with :ref:`metadata
1885 type <t_metadata>`. For example:
1886 "``metadata !{ i32 0, metadata !"test" }``". Unlike other
1887 constants that are meant to be interpreted as part of the
1888 instruction stream, metadata is a place to attach additional
1889 information such as debug info.
1891 Global Variable and Function Addresses
1892 --------------------------------------
1894 The addresses of :ref:`global variables <globalvars>` and
1895 :ref:`functions <functionstructure>` are always implicitly valid
1896 (link-time) constants. These constants are explicitly referenced when
1897 the :ref:`identifier for the global <identifiers>` is used and always have
1898 :ref:`pointer <t_pointer>` type. For example, the following is a legal LLVM
1901 .. code-block:: llvm
1905 @Z = global [2 x i32*] [ i32* @X, i32* @Y ]
1912 The string '``undef``' can be used anywhere a constant is expected, and
1913 indicates that the user of the value may receive an unspecified
1914 bit-pattern. Undefined values may be of any type (other than '``label``'
1915 or '``void``') and be used anywhere a constant is permitted.
1917 Undefined values are useful because they indicate to the compiler that
1918 the program is well defined no matter what value is used. This gives the
1919 compiler more freedom to optimize. Here are some examples of
1920 (potentially surprising) transformations that are valid (in pseudo IR):
1922 .. code-block:: llvm
1932 This is safe because all of the output bits are affected by the undef
1933 bits. Any output bit can have a zero or one depending on the input bits.
1935 .. code-block:: llvm
1946 These logical operations have bits that are not always affected by the
1947 input. For example, if ``%X`` has a zero bit, then the output of the
1948 '``and``' operation will always be a zero for that bit, no matter what
1949 the corresponding bit from the '``undef``' is. As such, it is unsafe to
1950 optimize or assume that the result of the '``and``' is '``undef``'.
1951 However, it is safe to assume that all bits of the '``undef``' could be
1952 0, and optimize the '``and``' to 0. Likewise, it is safe to assume that
1953 all the bits of the '``undef``' operand to the '``or``' could be set,
1954 allowing the '``or``' to be folded to -1.
1956 .. code-block:: llvm
1958 %A = select undef, %X, %Y
1959 %B = select undef, 42, %Y
1960 %C = select %X, %Y, undef
1970 This set of examples shows that undefined '``select``' (and conditional
1971 branch) conditions can go *either way*, but they have to come from one
1972 of the two operands. In the ``%A`` example, if ``%X`` and ``%Y`` were
1973 both known to have a clear low bit, then ``%A`` would have to have a
1974 cleared low bit. However, in the ``%C`` example, the optimizer is
1975 allowed to assume that the '``undef``' operand could be the same as
1976 ``%Y``, allowing the whole '``select``' to be eliminated.
1978 .. code-block:: llvm
1980 %A = xor undef, undef
1997 This example points out that two '``undef``' operands are not
1998 necessarily the same. This can be surprising to people (and also matches
1999 C semantics) where they assume that "``X^X``" is always zero, even if
2000 ``X`` is undefined. This isn't true for a number of reasons, but the
2001 short answer is that an '``undef``' "variable" can arbitrarily change
2002 its value over its "live range". This is true because the variable
2003 doesn't actually *have a live range*. Instead, the value is logically
2004 read from arbitrary registers that happen to be around when needed, so
2005 the value is not necessarily consistent over time. In fact, ``%A`` and
2006 ``%C`` need to have the same semantics or the core LLVM "replace all
2007 uses with" concept would not hold.
2009 .. code-block:: llvm
2017 These examples show the crucial difference between an *undefined value*
2018 and *undefined behavior*. An undefined value (like '``undef``') is
2019 allowed to have an arbitrary bit-pattern. This means that the ``%A``
2020 operation can be constant folded to '``undef``', because the '``undef``'
2021 could be an SNaN, and ``fdiv`` is not (currently) defined on SNaN's.
2022 However, in the second example, we can make a more aggressive
2023 assumption: because the ``undef`` is allowed to be an arbitrary value,
2024 we are allowed to assume that it could be zero. Since a divide by zero
2025 has *undefined behavior*, we are allowed to assume that the operation
2026 does not execute at all. This allows us to delete the divide and all
2027 code after it. Because the undefined operation "can't happen", the
2028 optimizer can assume that it occurs in dead code.
2030 .. code-block:: llvm
2032 a: store undef -> %X
2033 b: store %X -> undef
2038 These examples reiterate the ``fdiv`` example: a store *of* an undefined
2039 value can be assumed to not have any effect; we can assume that the
2040 value is overwritten with bits that happen to match what was already
2041 there. However, a store *to* an undefined location could clobber
2042 arbitrary memory, therefore, it has undefined behavior.
2049 Poison values are similar to :ref:`undef values <undefvalues>`, however
2050 they also represent the fact that an instruction or constant expression
2051 which cannot evoke side effects has nevertheless detected a condition
2052 which results in undefined behavior.
2054 There is currently no way of representing a poison value in the IR; they
2055 only exist when produced by operations such as :ref:`add <i_add>` with
2058 Poison value behavior is defined in terms of value *dependence*:
2060 - Values other than :ref:`phi <i_phi>` nodes depend on their operands.
2061 - :ref:`Phi <i_phi>` nodes depend on the operand corresponding to
2062 their dynamic predecessor basic block.
2063 - Function arguments depend on the corresponding actual argument values
2064 in the dynamic callers of their functions.
2065 - :ref:`Call <i_call>` instructions depend on the :ref:`ret <i_ret>`
2066 instructions that dynamically transfer control back to them.
2067 - :ref:`Invoke <i_invoke>` instructions depend on the
2068 :ref:`ret <i_ret>`, :ref:`resume <i_resume>`, or exception-throwing
2069 call instructions that dynamically transfer control back to them.
2070 - Non-volatile loads and stores depend on the most recent stores to all
2071 of the referenced memory addresses, following the order in the IR
2072 (including loads and stores implied by intrinsics such as
2073 :ref:`@llvm.memcpy <int_memcpy>`.)
2074 - An instruction with externally visible side effects depends on the
2075 most recent preceding instruction with externally visible side
2076 effects, following the order in the IR. (This includes :ref:`volatile
2077 operations <volatile>`.)
2078 - An instruction *control-depends* on a :ref:`terminator
2079 instruction <terminators>` if the terminator instruction has
2080 multiple successors and the instruction is always executed when
2081 control transfers to one of the successors, and may not be executed
2082 when control is transferred to another.
2083 - Additionally, an instruction also *control-depends* on a terminator
2084 instruction if the set of instructions it otherwise depends on would
2085 be different if the terminator had transferred control to a different
2087 - Dependence is transitive.
2089 Poison Values have the same behavior as :ref:`undef values <undefvalues>`,
2090 with the additional affect that any instruction which has a *dependence*
2091 on a poison value has undefined behavior.
2093 Here are some examples:
2095 .. code-block:: llvm
2098 %poison = sub nuw i32 0, 1 ; Results in a poison value.
2099 %still_poison = and i32 %poison, 0 ; 0, but also poison.
2100 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
2101 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
2103 store i32 %poison, i32* @g ; Poison value stored to memory.
2104 %poison2 = load i32* @g ; Poison value loaded back from memory.
2106 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
2108 %narrowaddr = bitcast i32* @g to i16*
2109 %wideaddr = bitcast i32* @g to i64*
2110 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2111 %poison4 = load i64* %wideaddr ; Returns a poison value.
2113 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2114 br i1 %cmp, label %true, label %end ; Branch to either destination.
2117 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2118 ; it has undefined behavior.
2122 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2123 ; Both edges into this PHI are
2124 ; control-dependent on %cmp, so this
2125 ; always results in a poison value.
2127 store volatile i32 0, i32* @g ; This would depend on the store in %true
2128 ; if %cmp is true, or the store in %entry
2129 ; otherwise, so this is undefined behavior.
2131 br i1 %cmp, label %second_true, label %second_end
2132 ; The same branch again, but this time the
2133 ; true block doesn't have side effects.
2140 store volatile i32 0, i32* @g ; This time, the instruction always depends
2141 ; on the store in %end. Also, it is
2142 ; control-equivalent to %end, so this is
2143 ; well-defined (ignoring earlier undefined
2144 ; behavior in this example).
2148 Addresses of Basic Blocks
2149 -------------------------
2151 ``blockaddress(@function, %block)``
2153 The '``blockaddress``' constant computes the address of the specified
2154 basic block in the specified function, and always has an ``i8*`` type.
2155 Taking the address of the entry block is illegal.
2157 This value only has defined behavior when used as an operand to the
2158 ':ref:`indirectbr <i_indirectbr>`' instruction, or for comparisons
2159 against null. Pointer equality tests between labels addresses results in
2160 undefined behavior --- though, again, comparison against null is ok, and
2161 no label is equal to the null pointer. This may be passed around as an
2162 opaque pointer sized value as long as the bits are not inspected. This
2163 allows ``ptrtoint`` and arithmetic to be performed on these values so
2164 long as the original value is reconstituted before the ``indirectbr``
2167 Finally, some targets may provide defined semantics when using the value
2168 as the operand to an inline assembly, but that is target specific.
2170 Constant Expressions
2171 --------------------
2173 Constant expressions are used to allow expressions involving other
2174 constants to be used as constants. Constant expressions may be of any
2175 :ref:`first class <t_firstclass>` type and may involve any LLVM operation
2176 that does not have side effects (e.g. load and call are not supported).
2177 The following is the syntax for constant expressions:
2179 ``trunc (CST to TYPE)``
2180 Truncate a constant to another type. The bit size of CST must be
2181 larger than the bit size of TYPE. Both types must be integers.
2182 ``zext (CST to TYPE)``
2183 Zero extend a constant to another type. The bit size of CST must be
2184 smaller than the bit size of TYPE. Both types must be integers.
2185 ``sext (CST to TYPE)``
2186 Sign extend a constant to another type. The bit size of CST must be
2187 smaller than the bit size of TYPE. Both types must be integers.
2188 ``fptrunc (CST to TYPE)``
2189 Truncate a floating point constant to another floating point type.
2190 The size of CST must be larger than the size of TYPE. Both types
2191 must be floating point.
2192 ``fpext (CST to TYPE)``
2193 Floating point extend a constant to another type. The size of CST
2194 must be smaller or equal to the size of TYPE. Both types must be
2196 ``fptoui (CST to TYPE)``
2197 Convert a floating point constant to the corresponding unsigned
2198 integer constant. TYPE must be a scalar or vector integer type. CST
2199 must be of scalar or vector floating point type. Both CST and TYPE
2200 must be scalars, or vectors of the same number of elements. If the
2201 value won't fit in the integer type, the results are undefined.
2202 ``fptosi (CST to TYPE)``
2203 Convert a floating point constant to the corresponding signed
2204 integer constant. TYPE must be a scalar or vector integer type. CST
2205 must be of scalar or vector floating point type. Both CST and TYPE
2206 must be scalars, or vectors of the same number of elements. If the
2207 value won't fit in the integer type, the results are undefined.
2208 ``uitofp (CST to TYPE)``
2209 Convert an unsigned integer constant to the corresponding floating
2210 point constant. TYPE must be a scalar or vector floating point type.
2211 CST must be of scalar or vector integer type. Both CST and TYPE must
2212 be scalars, or vectors of the same number of elements. If the value
2213 won't fit in the floating point type, the results are undefined.
2214 ``sitofp (CST to TYPE)``
2215 Convert a signed integer constant to the corresponding floating
2216 point constant. TYPE must be a scalar or vector floating point type.
2217 CST must be of scalar or vector integer type. Both CST and TYPE must
2218 be scalars, or vectors of the same number of elements. If the value
2219 won't fit in the floating point type, the results are undefined.
2220 ``ptrtoint (CST to TYPE)``
2221 Convert a pointer typed constant to the corresponding integer
2222 constant ``TYPE`` must be an integer type. ``CST`` must be of
2223 pointer type. The ``CST`` value is zero extended, truncated, or
2224 unchanged to make it fit in ``TYPE``.
2225 ``inttoptr (CST to TYPE)``
2226 Convert an integer constant to a pointer constant. TYPE must be a
2227 pointer type. CST must be of integer type. The CST value is zero
2228 extended, truncated, or unchanged to make it fit in a pointer size.
2229 This one is *really* dangerous!
2230 ``bitcast (CST to TYPE)``
2231 Convert a constant, CST, to another TYPE. The constraints of the
2232 operands are the same as those for the :ref:`bitcast
2233 instruction <i_bitcast>`.
2234 ``getelementptr (CSTPTR, IDX0, IDX1, ...)``, ``getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)``
2235 Perform the :ref:`getelementptr operation <i_getelementptr>` on
2236 constants. As with the :ref:`getelementptr <i_getelementptr>`
2237 instruction, the index list may have zero or more indexes, which are
2238 required to make sense for the type of "CSTPTR".
2239 ``select (COND, VAL1, VAL2)``
2240 Perform the :ref:`select operation <i_select>` on constants.
2241 ``icmp COND (VAL1, VAL2)``
2242 Performs the :ref:`icmp operation <i_icmp>` on constants.
2243 ``fcmp COND (VAL1, VAL2)``
2244 Performs the :ref:`fcmp operation <i_fcmp>` on constants.
2245 ``extractelement (VAL, IDX)``
2246 Perform the :ref:`extractelement operation <i_extractelement>` on
2248 ``insertelement (VAL, ELT, IDX)``
2249 Perform the :ref:`insertelement operation <i_insertelement>` on
2251 ``shufflevector (VEC1, VEC2, IDXMASK)``
2252 Perform the :ref:`shufflevector operation <i_shufflevector>` on
2254 ``extractvalue (VAL, IDX0, IDX1, ...)``
2255 Perform the :ref:`extractvalue operation <i_extractvalue>` on
2256 constants. The index list is interpreted in a similar manner as
2257 indices in a ':ref:`getelementptr <i_getelementptr>`' operation. At
2258 least one index value must be specified.
2259 ``insertvalue (VAL, ELT, IDX0, IDX1, ...)``
2260 Perform the :ref:`insertvalue operation <i_insertvalue>` on constants.
2261 The index list is interpreted in a similar manner as indices in a
2262 ':ref:`getelementptr <i_getelementptr>`' operation. At least one index
2263 value must be specified.
2264 ``OPCODE (LHS, RHS)``
2265 Perform the specified operation of the LHS and RHS constants. OPCODE
2266 may be any of the :ref:`binary <binaryops>` or :ref:`bitwise
2267 binary <bitwiseops>` operations. The constraints on operands are
2268 the same as those for the corresponding instruction (e.g. no bitwise
2269 operations on floating point values are allowed).
2274 Inline Assembler Expressions
2275 ----------------------------
2277 LLVM supports inline assembler expressions (as opposed to :ref:`Module-Level
2278 Inline Assembly <moduleasm>`) through the use of a special value. This
2279 value represents the inline assembler as a string (containing the
2280 instructions to emit), a list of operand constraints (stored as a
2281 string), a flag that indicates whether or not the inline asm expression
2282 has side effects, and a flag indicating whether the function containing
2283 the asm needs to align its stack conservatively. An example inline
2284 assembler expression is:
2286 .. code-block:: llvm
2288 i32 (i32) asm "bswap $0", "=r,r"
2290 Inline assembler expressions may **only** be used as the callee operand
2291 of a :ref:`call <i_call>` or an :ref:`invoke <i_invoke>` instruction.
2292 Thus, typically we have:
2294 .. code-block:: llvm
2296 %X = call i32 asm "bswap $0", "=r,r"(i32 %Y)
2298 Inline asms with side effects not visible in the constraint list must be
2299 marked as having side effects. This is done through the use of the
2300 '``sideeffect``' keyword, like so:
2302 .. code-block:: llvm
2304 call void asm sideeffect "eieio", ""()
2306 In some cases inline asms will contain code that will not work unless
2307 the stack is aligned in some way, such as calls or SSE instructions on
2308 x86, yet will not contain code that does that alignment within the asm.
2309 The compiler should make conservative assumptions about what the asm
2310 might contain and should generate its usual stack alignment code in the
2311 prologue if the '``alignstack``' keyword is present:
2313 .. code-block:: llvm
2315 call void asm alignstack "eieio", ""()
2317 Inline asms also support using non-standard assembly dialects. The
2318 assumed dialect is ATT. When the '``inteldialect``' keyword is present,
2319 the inline asm is using the Intel dialect. Currently, ATT and Intel are
2320 the only supported dialects. An example is:
2322 .. code-block:: llvm
2324 call void asm inteldialect "eieio", ""()
2326 If multiple keywords appear the '``sideeffect``' keyword must come
2327 first, the '``alignstack``' keyword second and the '``inteldialect``'
2333 The call instructions that wrap inline asm nodes may have a
2334 "``!srcloc``" MDNode attached to it that contains a list of constant
2335 integers. If present, the code generator will use the integer as the
2336 location cookie value when report errors through the ``LLVMContext``
2337 error reporting mechanisms. This allows a front-end to correlate backend
2338 errors that occur with inline asm back to the source code that produced
2341 .. code-block:: llvm
2343 call void asm sideeffect "something bad", ""(), !srcloc !42
2345 !42 = !{ i32 1234567 }
2347 It is up to the front-end to make sense of the magic numbers it places
2348 in the IR. If the MDNode contains multiple constants, the code generator
2349 will use the one that corresponds to the line of the asm that the error
2354 Metadata Nodes and Metadata Strings
2355 -----------------------------------
2357 LLVM IR allows metadata to be attached to instructions in the program
2358 that can convey extra information about the code to the optimizers and
2359 code generator. One example application of metadata is source-level
2360 debug information. There are two metadata primitives: strings and nodes.
2361 All metadata has the ``metadata`` type and is identified in syntax by a
2362 preceding exclamation point ('``!``').
2364 A metadata string is a string surrounded by double quotes. It can
2365 contain any character by escaping non-printable characters with
2366 "``\xx``" where "``xx``" is the two digit hex code. For example:
2369 Metadata nodes are represented with notation similar to structure
2370 constants (a comma separated list of elements, surrounded by braces and
2371 preceded by an exclamation point). Metadata nodes can have any values as
2372 their operand. For example:
2374 .. code-block:: llvm
2376 !{ metadata !"test\00", i32 10}
2378 A :ref:`named metadata <namedmetadatastructure>` is a collection of
2379 metadata nodes, which can be looked up in the module symbol table. For
2382 .. code-block:: llvm
2384 !foo = metadata !{!4, !3}
2386 Metadata can be used as function arguments. Here ``llvm.dbg.value``
2387 function is using two metadata arguments:
2389 .. code-block:: llvm
2391 call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2393 Metadata can be attached with an instruction. Here metadata ``!21`` is
2394 attached to the ``add`` instruction using the ``!dbg`` identifier:
2396 .. code-block:: llvm
2398 %indvar.next = add i64 %indvar, 1, !dbg !21
2400 More information about specific metadata nodes recognized by the
2401 optimizers and code generator is found below.
2406 In LLVM IR, memory does not have types, so LLVM's own type system is not
2407 suitable for doing TBAA. Instead, metadata is added to the IR to
2408 describe a type system of a higher level language. This can be used to
2409 implement typical C/C++ TBAA, but it can also be used to implement
2410 custom alias analysis behavior for other languages.
2412 The current metadata format is very simple. TBAA metadata nodes have up
2413 to three fields, e.g.:
2415 .. code-block:: llvm
2417 !0 = metadata !{ metadata !"an example type tree" }
2418 !1 = metadata !{ metadata !"int", metadata !0 }
2419 !2 = metadata !{ metadata !"float", metadata !0 }
2420 !3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2422 The first field is an identity field. It can be any value, usually a
2423 metadata string, which uniquely identifies the type. The most important
2424 name in the tree is the name of the root node. Two trees with different
2425 root node names are entirely disjoint, even if they have leaves with
2428 The second field identifies the type's parent node in the tree, or is
2429 null or omitted for a root node. A type is considered to alias all of
2430 its descendants and all of its ancestors in the tree. Also, a type is
2431 considered to alias all types in other trees, so that bitcode produced
2432 from multiple front-ends is handled conservatively.
2434 If the third field is present, it's an integer which if equal to 1
2435 indicates that the type is "constant" (meaning
2436 ``pointsToConstantMemory`` should return true; see `other useful
2437 AliasAnalysis methods <AliasAnalysis.html#OtherItfs>`_).
2439 '``tbaa.struct``' Metadata
2440 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2442 The :ref:`llvm.memcpy <int_memcpy>` is often used to implement
2443 aggregate assignment operations in C and similar languages, however it
2444 is defined to copy a contiguous region of memory, which is more than
2445 strictly necessary for aggregate types which contain holes due to
2446 padding. Also, it doesn't contain any TBAA information about the fields
2449 ``!tbaa.struct`` metadata can describe which memory subregions in a
2450 memcpy are padding and what the TBAA tags of the struct are.
2452 The current metadata format is very simple. ``!tbaa.struct`` metadata
2453 nodes are a list of operands which are in conceptual groups of three.
2454 For each group of three, the first operand gives the byte offset of a
2455 field in bytes, the second gives its size in bytes, and the third gives
2458 .. code-block:: llvm
2460 !4 = metadata !{ i64 0, i64 4, metadata !1, i64 8, i64 4, metadata !2 }
2462 This describes a struct with two fields. The first is at offset 0 bytes
2463 with size 4 bytes, and has tbaa tag !1. The second is at offset 8 bytes
2464 and has size 4 bytes and has tbaa tag !2.
2466 Note that the fields need not be contiguous. In this example, there is a
2467 4 byte gap between the two fields. This gap represents padding which
2468 does not carry useful data and need not be preserved.
2470 '``fpmath``' Metadata
2471 ^^^^^^^^^^^^^^^^^^^^^
2473 ``fpmath`` metadata may be attached to any instruction of floating point
2474 type. It can be used to express the maximum acceptable error in the
2475 result of that instruction, in ULPs, thus potentially allowing the
2476 compiler to use a more efficient but less accurate method of computing
2477 it. ULP is defined as follows:
2479 If ``x`` is a real number that lies between two finite consecutive
2480 floating-point numbers ``a`` and ``b``, without being equal to one
2481 of them, then ``ulp(x) = |b - a|``, otherwise ``ulp(x)`` is the
2482 distance between the two non-equal finite floating-point numbers
2483 nearest ``x``. Moreover, ``ulp(NaN)`` is ``NaN``.
2485 The metadata node shall consist of a single positive floating point
2486 number representing the maximum relative error, for example:
2488 .. code-block:: llvm
2490 !0 = metadata !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
2492 '``range``' Metadata
2493 ^^^^^^^^^^^^^^^^^^^^
2495 ``range`` metadata may be attached only to loads of integer types. It
2496 expresses the possible ranges the loaded value is in. The ranges are
2497 represented with a flattened list of integers. The loaded value is known
2498 to be in the union of the ranges defined by each consecutive pair. Each
2499 pair has the following properties:
2501 - The type must match the type loaded by the instruction.
2502 - The pair ``a,b`` represents the range ``[a,b)``.
2503 - Both ``a`` and ``b`` are constants.
2504 - The range is allowed to wrap.
2505 - The range should not represent the full or empty set. That is,
2508 In addition, the pairs must be in signed order of the lower bound and
2509 they must be non-contiguous.
2513 .. code-block:: llvm
2515 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
2516 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
2517 %c = load i8* %z, align 1, !range !2 ; Can only be 0, 1, 3, 4 or 5
2518 %d = load i8* %z, align 1, !range !3 ; Can only be -2, -1, 3, 4 or 5
2520 !0 = metadata !{ i8 0, i8 2 }
2521 !1 = metadata !{ i8 255, i8 2 }
2522 !2 = metadata !{ i8 0, i8 2, i8 3, i8 6 }
2523 !3 = metadata !{ i8 -2, i8 0, i8 3, i8 6 }
2528 It is sometimes useful to attach information to loop constructs. Currently,
2529 loop metadata is implemented as metadata attached to the branch instruction
2530 in the loop latch block. This type of metadata refer to a metadata node that is
2531 guaranteed to be separate for each loop. The loop-level metadata is prefixed
2534 The loop identifier metadata is implemented using a metadata that refers to
2537 .. code-block:: llvm
2538 !0 = metadata !{ metadata !0 }
2540 '``llvm.loop.parallel``' Metadata
2541 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2543 This loop metadata can be used to communicate that a loop should be considered
2544 a parallel loop. The semantics of parallel loops in this case is the one
2545 with the strongest cross-iteration instruction ordering freedom: the
2546 iterations in the loop can be considered completely independent of each
2547 other (also known as embarrassingly parallel loops).
2549 This metadata can originate from a programming language with parallel loop
2550 constructs. In such a case it is completely the programmer's responsibility
2551 to ensure the instructions from the different iterations of the loop can be
2552 executed in an arbitrary order, in parallel, or intertwined. No loop-carried
2553 dependency checking at all must be expected from the compiler.
2555 In order to fulfill the LLVM requirement for metadata to be safely ignored,
2556 it is important to ensure that a parallel loop is converted to
2557 a sequential loop in case an optimization (agnostic of the parallel loop
2558 semantics) converts the loop back to such. This happens when new memory
2559 accesses that do not fulfill the requirement of free ordering across iterations
2560 are added to the loop. Therefore, this metadata is required, but not
2561 sufficient, to consider the loop at hand a parallel loop. For a loop
2562 to be parallel, all its memory accessing instructions need to be
2563 marked with the ``llvm.mem.parallel_loop_access`` metadata that refer
2564 to the same loop identifier metadata that identify the loop at hand.
2569 Metadata types used to annotate memory accesses with information helpful
2570 for optimizations are prefixed with ``llvm.mem``.
2572 '``llvm.mem.parallel_loop_access``' Metadata
2573 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2575 For a loop to be parallel, in addition to using
2576 the ``llvm.loop.parallel`` metadata to mark the loop latch branch instruction,
2577 also all of the memory accessing instructions in the loop body need to be
2578 marked with the ``llvm.mem.parallel_loop_access`` metadata. If there
2579 is at least one memory accessing instruction not marked with the metadata,
2580 the loop, despite it possibly using the ``llvm.loop.parallel`` metadata,
2581 must be considered a sequential loop. This causes parallel loops to be
2582 converted to sequential loops due to optimization passes that are unaware of
2583 the parallel semantics and that insert new memory instructions to the loop
2586 Example of a loop that is considered parallel due to its correct use of
2587 both ``llvm.loop.parallel`` and ``llvm.mem.parallel_loop_access``
2588 metadata types that refer to the same loop identifier metadata.
2590 .. code-block:: llvm
2594 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2596 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2598 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop.parallel !0
2602 !0 = metadata !{ metadata !0 }
2604 It is also possible to have nested parallel loops. In that case the
2605 memory accesses refer to a list of loop identifier metadata nodes instead of
2606 the loop identifier metadata node directly:
2608 .. code-block:: llvm
2615 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2617 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2619 br i1 %exitcond, label %inner.for.end, label %inner.for.body, !llvm.loop.parallel !1
2623 %0 = load i32* %arrayidx, align 4, !llvm.mem.parallel_loop_access !0
2625 store i32 %0, i32* %arrayidx4, align 4, !llvm.mem.parallel_loop_access !0
2627 br i1 %exitcond, label %outer.for.end, label %outer.for.body, !llvm.loop.parallel !2
2629 outer.for.end: ; preds = %for.body
2631 !0 = metadata !{ metadata !1, metadata !2 } ; a list of parallel loop identifiers
2632 !1 = metadata !{ metadata !1 } ; an identifier for the inner parallel loop
2633 !2 = metadata !{ metadata !2 } ; an identifier for the outer parallel loop
2636 Module Flags Metadata
2637 =====================
2639 Information about the module as a whole is difficult to convey to LLVM's
2640 subsystems. The LLVM IR isn't sufficient to transmit this information.
2641 The ``llvm.module.flags`` named metadata exists in order to facilitate
2642 this. These flags are in the form of key / value pairs --- much like a
2643 dictionary --- making it easy for any subsystem who cares about a flag to
2646 The ``llvm.module.flags`` metadata contains a list of metadata triplets.
2647 Each triplet has the following form:
2649 - The first element is a *behavior* flag, which specifies the behavior
2650 when two (or more) modules are merged together, and it encounters two
2651 (or more) metadata with the same ID. The supported behaviors are
2653 - The second element is a metadata string that is a unique ID for the
2654 metadata. Each module may only have one flag entry for each unique ID (not
2655 including entries with the **Require** behavior).
2656 - The third element is the value of the flag.
2658 When two (or more) modules are merged together, the resulting
2659 ``llvm.module.flags`` metadata is the union of the modules' flags. That is, for
2660 each unique metadata ID string, there will be exactly one entry in the merged
2661 modules ``llvm.module.flags`` metadata table, and the value for that entry will
2662 be determined by the merge behavior flag, as described below. The only exception
2663 is that entries with the *Require* behavior are always preserved.
2665 The following behaviors are supported:
2676 Emits an error if two values disagree, otherwise the resulting value
2677 is that of the operands.
2681 Emits a warning if two values disagree. The result value will be the
2682 operand for the flag from the first module being linked.
2686 Adds a requirement that another module flag be present and have a
2687 specified value after linking is performed. The value must be a
2688 metadata pair, where the first element of the pair is the ID of the
2689 module flag to be restricted, and the second element of the pair is
2690 the value the module flag should be restricted to. This behavior can
2691 be used to restrict the allowable results (via triggering of an
2692 error) of linking IDs with the **Override** behavior.
2696 Uses the specified value, regardless of the behavior or value of the
2697 other module. If both modules specify **Override**, but the values
2698 differ, an error will be emitted.
2702 Appends the two values, which are required to be metadata nodes.
2706 Appends the two values, which are required to be metadata
2707 nodes. However, duplicate entries in the second list are dropped
2708 during the append operation.
2710 It is an error for a particular unique flag ID to have multiple behaviors,
2711 except in the case of **Require** (which adds restrictions on another metadata
2712 value) or **Override**.
2714 An example of module flags:
2716 .. code-block:: llvm
2718 !0 = metadata !{ i32 1, metadata !"foo", i32 1 }
2719 !1 = metadata !{ i32 4, metadata !"bar", i32 37 }
2720 !2 = metadata !{ i32 2, metadata !"qux", i32 42 }
2721 !3 = metadata !{ i32 3, metadata !"qux",
2723 metadata !"foo", i32 1
2726 !llvm.module.flags = !{ !0, !1, !2, !3 }
2728 - Metadata ``!0`` has the ID ``!"foo"`` and the value '1'. The behavior
2729 if two or more ``!"foo"`` flags are seen is to emit an error if their
2730 values are not equal.
2732 - Metadata ``!1`` has the ID ``!"bar"`` and the value '37'. The
2733 behavior if two or more ``!"bar"`` flags are seen is to use the value
2736 - Metadata ``!2`` has the ID ``!"qux"`` and the value '42'. The
2737 behavior if two or more ``!"qux"`` flags are seen is to emit a
2738 warning if their values are not equal.
2740 - Metadata ``!3`` has the ID ``!"qux"`` and the value:
2744 metadata !{ metadata !"foo", i32 1 }
2746 The behavior is to emit an error if the ``llvm.module.flags`` does not
2747 contain a flag with the ID ``!"foo"`` that has the value '1' after linking is
2750 Objective-C Garbage Collection Module Flags Metadata
2751 ----------------------------------------------------
2753 On the Mach-O platform, Objective-C stores metadata about garbage
2754 collection in a special section called "image info". The metadata
2755 consists of a version number and a bitmask specifying what types of
2756 garbage collection are supported (if any) by the file. If two or more
2757 modules are linked together their garbage collection metadata needs to
2758 be merged rather than appended together.
2760 The Objective-C garbage collection module flags metadata consists of the
2761 following key-value pairs:
2770 * - ``Objective-C Version``
2771 - **[Required]** --- The Objective-C ABI version. Valid values are 1 and 2.
2773 * - ``Objective-C Image Info Version``
2774 - **[Required]** --- The version of the image info section. Currently
2777 * - ``Objective-C Image Info Section``
2778 - **[Required]** --- The section to place the metadata. Valid values are
2779 ``"__OBJC, __image_info, regular"`` for Objective-C ABI version 1, and
2780 ``"__DATA,__objc_imageinfo, regular, no_dead_strip"`` for
2781 Objective-C ABI version 2.
2783 * - ``Objective-C Garbage Collection``
2784 - **[Required]** --- Specifies whether garbage collection is supported or
2785 not. Valid values are 0, for no garbage collection, and 2, for garbage
2786 collection supported.
2788 * - ``Objective-C GC Only``
2789 - **[Optional]** --- Specifies that only garbage collection is supported.
2790 If present, its value must be 6. This flag requires that the
2791 ``Objective-C Garbage Collection`` flag have the value 2.
2793 Some important flag interactions:
2795 - If a module with ``Objective-C Garbage Collection`` set to 0 is
2796 merged with a module with ``Objective-C Garbage Collection`` set to
2797 2, then the resulting module has the
2798 ``Objective-C Garbage Collection`` flag set to 0.
2799 - A module with ``Objective-C Garbage Collection`` set to 0 cannot be
2800 merged with a module with ``Objective-C GC Only`` set to 6.
2802 Automatic Linker Flags Module Flags Metadata
2803 --------------------------------------------
2805 Some targets support embedding flags to the linker inside individual object
2806 files. Typically this is used in conjunction with language extensions which
2807 allow source files to explicitly declare the libraries they depend on, and have
2808 these automatically be transmitted to the linker via object files.
2810 These flags are encoded in the IR using metadata in the module flags section,
2811 using the ``Linker Options`` key. The merge behavior for this flag is required
2812 to be ``AppendUnique``, and the value for the key is expected to be a metadata
2813 node which should be a list of other metadata nodes, each of which should be a
2814 list of metadata strings defining linker options.
2816 For example, the following metadata section specifies two separate sets of
2817 linker options, presumably to link against ``libz`` and the ``Cocoa``
2820 !0 = metadata !{ i32 6, metadata !"Linker Options",
2822 metadata !{ metadata !"-lz" },
2823 metadata !{ metadata !"-framework", metadata !"Cocoa" } } }
2824 !llvm.module.flags = !{ !0 }
2826 The metadata encoding as lists of lists of options, as opposed to a collapsed
2827 list of options, is chosen so that the IR encoding can use multiple option
2828 strings to specify e.g., a single library, while still having that specifier be
2829 preserved as an atomic element that can be recognized by a target specific
2830 assembly writer or object file emitter.
2832 Each individual option is required to be either a valid option for the target's
2833 linker, or an option that is reserved by the target specific assembly writer or
2834 object file emitter. No other aspect of these options is defined by the IR.
2836 Intrinsic Global Variables
2837 ==========================
2839 LLVM has a number of "magic" global variables that contain data that
2840 affect code generation or other IR semantics. These are documented here.
2841 All globals of this sort should have a section specified as
2842 "``llvm.metadata``". This section and all globals that start with
2843 "``llvm.``" are reserved for use by LLVM.
2845 The '``llvm.used``' Global Variable
2846 -----------------------------------
2848 The ``@llvm.used`` global is an array with i8\* element type which has
2849 :ref:`appending linkage <linkage_appending>`. This array contains a list of
2850 pointers to global variables and functions which may optionally have a
2851 pointer cast formed of bitcast or getelementptr. For example, a legal
2854 .. code-block:: llvm
2859 @llvm.used = appending global [2 x i8*] [
2861 i8* bitcast (i32* @Y to i8*)
2862 ], section "llvm.metadata"
2864 If a global variable appears in the ``@llvm.used`` list, then the
2865 compiler, assembler, and linker are required to treat the symbol as if
2866 there is a reference to the global that it cannot see. For example, if a
2867 variable has internal linkage and no references other than that from the
2868 ``@llvm.used`` list, it cannot be deleted. This is commonly used to
2869 represent references from inline asms and other things the compiler
2870 cannot "see", and corresponds to "``attribute((used))``" in GNU C.
2872 On some targets, the code generator must emit a directive to the
2873 assembler or object file to prevent the assembler and linker from
2874 molesting the symbol.
2876 The '``llvm.compiler.used``' Global Variable
2877 --------------------------------------------
2879 The ``@llvm.compiler.used`` directive is the same as the ``@llvm.used``
2880 directive, except that it only prevents the compiler from touching the
2881 symbol. On targets that support it, this allows an intelligent linker to
2882 optimize references to the symbol without being impeded as it would be
2885 This is a rare construct that should only be used in rare circumstances,
2886 and should not be exposed to source languages.
2888 The '``llvm.global_ctors``' Global Variable
2889 -------------------------------------------
2891 .. code-block:: llvm
2893 %0 = type { i32, void ()* }
2894 @llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
2896 The ``@llvm.global_ctors`` array contains a list of constructor
2897 functions and associated priorities. The functions referenced by this
2898 array will be called in ascending order of priority (i.e. lowest first)
2899 when the module is loaded. The order of functions with the same priority
2902 The '``llvm.global_dtors``' Global Variable
2903 -------------------------------------------
2905 .. code-block:: llvm
2907 %0 = type { i32, void ()* }
2908 @llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
2910 The ``@llvm.global_dtors`` array contains a list of destructor functions
2911 and associated priorities. The functions referenced by this array will
2912 be called in descending order of priority (i.e. highest first) when the
2913 module is loaded. The order of functions with the same priority is not
2916 Instruction Reference
2917 =====================
2919 The LLVM instruction set consists of several different classifications
2920 of instructions: :ref:`terminator instructions <terminators>`, :ref:`binary
2921 instructions <binaryops>`, :ref:`bitwise binary
2922 instructions <bitwiseops>`, :ref:`memory instructions <memoryops>`, and
2923 :ref:`other instructions <otherops>`.
2927 Terminator Instructions
2928 -----------------------
2930 As mentioned :ref:`previously <functionstructure>`, every basic block in a
2931 program ends with a "Terminator" instruction, which indicates which
2932 block should be executed after the current block is finished. These
2933 terminator instructions typically yield a '``void``' value: they produce
2934 control flow, not values (the one exception being the
2935 ':ref:`invoke <i_invoke>`' instruction).
2937 The terminator instructions are: ':ref:`ret <i_ret>`',
2938 ':ref:`br <i_br>`', ':ref:`switch <i_switch>`',
2939 ':ref:`indirectbr <i_indirectbr>`', ':ref:`invoke <i_invoke>`',
2940 ':ref:`resume <i_resume>`', and ':ref:`unreachable <i_unreachable>`'.
2944 '``ret``' Instruction
2945 ^^^^^^^^^^^^^^^^^^^^^
2952 ret <type> <value> ; Return a value from a non-void function
2953 ret void ; Return from void function
2958 The '``ret``' instruction is used to return control flow (and optionally
2959 a value) from a function back to the caller.
2961 There are two forms of the '``ret``' instruction: one that returns a
2962 value and then causes control flow, and one that just causes control
2968 The '``ret``' instruction optionally accepts a single argument, the
2969 return value. The type of the return value must be a ':ref:`first
2970 class <t_firstclass>`' type.
2972 A function is not :ref:`well formed <wellformed>` if it it has a non-void
2973 return type and contains a '``ret``' instruction with no return value or
2974 a return value with a type that does not match its type, or if it has a
2975 void return type and contains a '``ret``' instruction with a return
2981 When the '``ret``' instruction is executed, control flow returns back to
2982 the calling function's context. If the caller is a
2983 ":ref:`call <i_call>`" instruction, execution continues at the
2984 instruction after the call. If the caller was an
2985 ":ref:`invoke <i_invoke>`" instruction, execution continues at the
2986 beginning of the "normal" destination block. If the instruction returns
2987 a value, that value shall set the call or invoke instruction's return
2993 .. code-block:: llvm
2995 ret i32 5 ; Return an integer value of 5
2996 ret void ; Return from a void function
2997 ret { i32, i8 } { i32 4, i8 2 } ; Return a struct of values 4 and 2
3001 '``br``' Instruction
3002 ^^^^^^^^^^^^^^^^^^^^
3009 br i1 <cond>, label <iftrue>, label <iffalse>
3010 br label <dest> ; Unconditional branch
3015 The '``br``' instruction is used to cause control flow to transfer to a
3016 different basic block in the current function. There are two forms of
3017 this instruction, corresponding to a conditional branch and an
3018 unconditional branch.
3023 The conditional branch form of the '``br``' instruction takes a single
3024 '``i1``' value and two '``label``' values. The unconditional form of the
3025 '``br``' instruction takes a single '``label``' value as a target.
3030 Upon execution of a conditional '``br``' instruction, the '``i1``'
3031 argument is evaluated. If the value is ``true``, control flows to the
3032 '``iftrue``' ``label`` argument. If "cond" is ``false``, control flows
3033 to the '``iffalse``' ``label`` argument.
3038 .. code-block:: llvm
3041 %cond = icmp eq i32 %a, %b
3042 br i1 %cond, label %IfEqual, label %IfUnequal
3050 '``switch``' Instruction
3051 ^^^^^^^^^^^^^^^^^^^^^^^^
3058 switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
3063 The '``switch``' instruction is used to transfer control flow to one of
3064 several different places. It is a generalization of the '``br``'
3065 instruction, allowing a branch to occur to one of many possible
3071 The '``switch``' instruction uses three parameters: an integer
3072 comparison value '``value``', a default '``label``' destination, and an
3073 array of pairs of comparison value constants and '``label``'s. The table
3074 is not allowed to contain duplicate constant entries.
3079 The ``switch`` instruction specifies a table of values and destinations.
3080 When the '``switch``' instruction is executed, this table is searched
3081 for the given value. If the value is found, control flow is transferred
3082 to the corresponding destination; otherwise, control flow is transferred
3083 to the default destination.
3088 Depending on properties of the target machine and the particular
3089 ``switch`` instruction, this instruction may be code generated in
3090 different ways. For example, it could be generated as a series of
3091 chained conditional branches or with a lookup table.
3096 .. code-block:: llvm
3098 ; Emulate a conditional br instruction
3099 %Val = zext i1 %value to i32
3100 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
3102 ; Emulate an unconditional br instruction
3103 switch i32 0, label %dest [ ]
3105 ; Implement a jump table:
3106 switch i32 %val, label %otherwise [ i32 0, label %onzero
3108 i32 2, label %ontwo ]
3112 '``indirectbr``' Instruction
3113 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3120 indirectbr <somety>* <address>, [ label <dest1>, label <dest2>, ... ]
3125 The '``indirectbr``' instruction implements an indirect branch to a
3126 label within the current function, whose address is specified by
3127 "``address``". Address must be derived from a
3128 :ref:`blockaddress <blockaddress>` constant.
3133 The '``address``' argument is the address of the label to jump to. The
3134 rest of the arguments indicate the full set of possible destinations
3135 that the address may point to. Blocks are allowed to occur multiple
3136 times in the destination list, though this isn't particularly useful.
3138 This destination list is required so that dataflow analysis has an
3139 accurate understanding of the CFG.
3144 Control transfers to the block specified in the address argument. All
3145 possible destination blocks must be listed in the label list, otherwise
3146 this instruction has undefined behavior. This implies that jumps to
3147 labels defined in other functions have undefined behavior as well.
3152 This is typically implemented with a jump through a register.
3157 .. code-block:: llvm
3159 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
3163 '``invoke``' Instruction
3164 ^^^^^^^^^^^^^^^^^^^^^^^^
3171 <result> = invoke [cconv] [ret attrs] <ptr to function ty> <function ptr val>(<function args>) [fn attrs]
3172 to label <normal label> unwind label <exception label>
3177 The '``invoke``' instruction causes control to transfer to a specified
3178 function, with the possibility of control flow transfer to either the
3179 '``normal``' label or the '``exception``' label. If the callee function
3180 returns with the "``ret``" instruction, control flow will return to the
3181 "normal" label. If the callee (or any indirect callees) returns via the
3182 ":ref:`resume <i_resume>`" instruction or other exception handling
3183 mechanism, control is interrupted and continued at the dynamically
3184 nearest "exception" label.
3186 The '``exception``' label is a `landing
3187 pad <ExceptionHandling.html#overview>`_ for the exception. As such,
3188 '``exception``' label is required to have the
3189 ":ref:`landingpad <i_landingpad>`" instruction, which contains the
3190 information about the behavior of the program after unwinding happens,
3191 as its first non-PHI instruction. The restrictions on the
3192 "``landingpad``" instruction's tightly couples it to the "``invoke``"
3193 instruction, so that the important information contained within the
3194 "``landingpad``" instruction can't be lost through normal code motion.
3199 This instruction requires several arguments:
3201 #. The optional "cconv" marker indicates which :ref:`calling
3202 convention <callingconv>` the call should use. If none is
3203 specified, the call defaults to using C calling conventions.
3204 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
3205 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
3207 #. '``ptr to function ty``': shall be the signature of the pointer to
3208 function value being invoked. In most cases, this is a direct
3209 function invocation, but indirect ``invoke``'s are just as possible,
3210 branching off an arbitrary pointer to function value.
3211 #. '``function ptr val``': An LLVM value containing a pointer to a
3212 function to be invoked.
3213 #. '``function args``': argument list whose types match the function
3214 signature argument types and parameter attributes. All arguments must
3215 be of :ref:`first class <t_firstclass>` type. If the function signature
3216 indicates the function accepts a variable number of arguments, the
3217 extra arguments can be specified.
3218 #. '``normal label``': the label reached when the called function
3219 executes a '``ret``' instruction.
3220 #. '``exception label``': the label reached when a callee returns via
3221 the :ref:`resume <i_resume>` instruction or other exception handling
3223 #. The optional :ref:`function attributes <fnattrs>` list. Only
3224 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
3225 attributes are valid here.
3230 This instruction is designed to operate as a standard '``call``'
3231 instruction in most regards. The primary difference is that it
3232 establishes an association with a label, which is used by the runtime
3233 library to unwind the stack.
3235 This instruction is used in languages with destructors to ensure that
3236 proper cleanup is performed in the case of either a ``longjmp`` or a
3237 thrown exception. Additionally, this is important for implementation of
3238 '``catch``' clauses in high-level languages that support them.
3240 For the purposes of the SSA form, the definition of the value returned
3241 by the '``invoke``' instruction is deemed to occur on the edge from the
3242 current block to the "normal" label. If the callee unwinds then no
3243 return value is available.
3248 .. code-block:: llvm
3250 %retval = invoke i32 @Test(i32 15) to label %Continue
3251 unwind label %TestCleanup ; {i32}:retval set
3252 %retval = invoke coldcc i32 %Testfnptr(i32 15) to label %Continue
3253 unwind label %TestCleanup ; {i32}:retval set
3257 '``resume``' Instruction
3258 ^^^^^^^^^^^^^^^^^^^^^^^^
3265 resume <type> <value>
3270 The '``resume``' instruction is a terminator instruction that has no
3276 The '``resume``' instruction requires one argument, which must have the
3277 same type as the result of any '``landingpad``' instruction in the same
3283 The '``resume``' instruction resumes propagation of an existing
3284 (in-flight) exception whose unwinding was interrupted with a
3285 :ref:`landingpad <i_landingpad>` instruction.
3290 .. code-block:: llvm
3292 resume { i8*, i32 } %exn
3296 '``unreachable``' Instruction
3297 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3309 The '``unreachable``' instruction has no defined semantics. This
3310 instruction is used to inform the optimizer that a particular portion of
3311 the code is not reachable. This can be used to indicate that the code
3312 after a no-return function cannot be reached, and other facts.
3317 The '``unreachable``' instruction has no defined semantics.
3324 Binary operators are used to do most of the computation in a program.
3325 They require two operands of the same type, execute an operation on
3326 them, and produce a single value. The operands might represent multiple
3327 data, as is the case with the :ref:`vector <t_vector>` data type. The
3328 result value has the same type as its operands.
3330 There are several different binary operators:
3334 '``add``' Instruction
3335 ^^^^^^^^^^^^^^^^^^^^^
3342 <result> = add <ty> <op1>, <op2> ; yields {ty}:result
3343 <result> = add nuw <ty> <op1>, <op2> ; yields {ty}:result
3344 <result> = add nsw <ty> <op1>, <op2> ; yields {ty}:result
3345 <result> = add nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3350 The '``add``' instruction returns the sum of its two operands.
3355 The two arguments to the '``add``' instruction must be
3356 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3357 arguments must have identical types.
3362 The value produced is the integer sum of the two operands.
3364 If the sum has unsigned overflow, the result returned is the
3365 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3368 Because LLVM integers use a two's complement representation, this
3369 instruction is appropriate for both signed and unsigned integers.
3371 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3372 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3373 result value of the ``add`` is a :ref:`poison value <poisonvalues>` if
3374 unsigned and/or signed overflow, respectively, occurs.
3379 .. code-block:: llvm
3381 <result> = add i32 4, %var ; yields {i32}:result = 4 + %var
3385 '``fadd``' Instruction
3386 ^^^^^^^^^^^^^^^^^^^^^^
3393 <result> = fadd [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3398 The '``fadd``' instruction returns the sum of its two operands.
3403 The two arguments to the '``fadd``' instruction must be :ref:`floating
3404 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3405 Both arguments must have identical types.
3410 The value produced is the floating point sum of the two operands. This
3411 instruction can also take any number of :ref:`fast-math flags <fastmath>`,
3412 which are optimization hints to enable otherwise unsafe floating point
3418 .. code-block:: llvm
3420 <result> = fadd float 4.0, %var ; yields {float}:result = 4.0 + %var
3422 '``sub``' Instruction
3423 ^^^^^^^^^^^^^^^^^^^^^
3430 <result> = sub <ty> <op1>, <op2> ; yields {ty}:result
3431 <result> = sub nuw <ty> <op1>, <op2> ; yields {ty}:result
3432 <result> = sub nsw <ty> <op1>, <op2> ; yields {ty}:result
3433 <result> = sub nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3438 The '``sub``' instruction returns the difference of its two operands.
3440 Note that the '``sub``' instruction is used to represent the '``neg``'
3441 instruction present in most other intermediate representations.
3446 The two arguments to the '``sub``' instruction must be
3447 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3448 arguments must have identical types.
3453 The value produced is the integer difference of the two operands.
3455 If the difference has unsigned overflow, the result returned is the
3456 mathematical result modulo 2\ :sup:`n`\ , where n is the bit width of
3459 Because LLVM integers use a two's complement representation, this
3460 instruction is appropriate for both signed and unsigned integers.
3462 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3463 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3464 result value of the ``sub`` is a :ref:`poison value <poisonvalues>` if
3465 unsigned and/or signed overflow, respectively, occurs.
3470 .. code-block:: llvm
3472 <result> = sub i32 4, %var ; yields {i32}:result = 4 - %var
3473 <result> = sub i32 0, %val ; yields {i32}:result = -%var
3477 '``fsub``' Instruction
3478 ^^^^^^^^^^^^^^^^^^^^^^
3485 <result> = fsub [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3490 The '``fsub``' instruction returns the difference of its two operands.
3492 Note that the '``fsub``' instruction is used to represent the '``fneg``'
3493 instruction present in most other intermediate representations.
3498 The two arguments to the '``fsub``' instruction must be :ref:`floating
3499 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3500 Both arguments must have identical types.
3505 The value produced is the floating point difference of the two operands.
3506 This instruction can also take any number of :ref:`fast-math
3507 flags <fastmath>`, which are optimization hints to enable otherwise
3508 unsafe floating point optimizations:
3513 .. code-block:: llvm
3515 <result> = fsub float 4.0, %var ; yields {float}:result = 4.0 - %var
3516 <result> = fsub float -0.0, %val ; yields {float}:result = -%var
3518 '``mul``' Instruction
3519 ^^^^^^^^^^^^^^^^^^^^^
3526 <result> = mul <ty> <op1>, <op2> ; yields {ty}:result
3527 <result> = mul nuw <ty> <op1>, <op2> ; yields {ty}:result
3528 <result> = mul nsw <ty> <op1>, <op2> ; yields {ty}:result
3529 <result> = mul nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3534 The '``mul``' instruction returns the product of its two operands.
3539 The two arguments to the '``mul``' instruction must be
3540 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3541 arguments must have identical types.
3546 The value produced is the integer product of the two operands.
3548 If the result of the multiplication has unsigned overflow, the result
3549 returned is the mathematical result modulo 2\ :sup:`n`\ , where n is the
3550 bit width of the result.
3552 Because LLVM integers use a two's complement representation, and the
3553 result is the same width as the operands, this instruction returns the
3554 correct result for both signed and unsigned integers. If a full product
3555 (e.g. ``i32`` * ``i32`` -> ``i64``) is needed, the operands should be
3556 sign-extended or zero-extended as appropriate to the width of the full
3559 ``nuw`` and ``nsw`` stand for "No Unsigned Wrap" and "No Signed Wrap",
3560 respectively. If the ``nuw`` and/or ``nsw`` keywords are present, the
3561 result value of the ``mul`` is a :ref:`poison value <poisonvalues>` if
3562 unsigned and/or signed overflow, respectively, occurs.
3567 .. code-block:: llvm
3569 <result> = mul i32 4, %var ; yields {i32}:result = 4 * %var
3573 '``fmul``' Instruction
3574 ^^^^^^^^^^^^^^^^^^^^^^
3581 <result> = fmul [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3586 The '``fmul``' instruction returns the product of its two operands.
3591 The two arguments to the '``fmul``' instruction must be :ref:`floating
3592 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3593 Both arguments must have identical types.
3598 The value produced is the floating point product of the two operands.
3599 This instruction can also take any number of :ref:`fast-math
3600 flags <fastmath>`, which are optimization hints to enable otherwise
3601 unsafe floating point optimizations:
3606 .. code-block:: llvm
3608 <result> = fmul float 4.0, %var ; yields {float}:result = 4.0 * %var
3610 '``udiv``' Instruction
3611 ^^^^^^^^^^^^^^^^^^^^^^
3618 <result> = udiv <ty> <op1>, <op2> ; yields {ty}:result
3619 <result> = udiv exact <ty> <op1>, <op2> ; yields {ty}:result
3624 The '``udiv``' instruction returns the quotient of its two operands.
3629 The two arguments to the '``udiv``' instruction must be
3630 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3631 arguments must have identical types.
3636 The value produced is the unsigned integer quotient of the two operands.
3638 Note that unsigned integer division and signed integer division are
3639 distinct operations; for signed integer division, use '``sdiv``'.
3641 Division by zero leads to undefined behavior.
3643 If the ``exact`` keyword is present, the result value of the ``udiv`` is
3644 a :ref:`poison value <poisonvalues>` if %op1 is not a multiple of %op2 (as
3645 such, "((a udiv exact b) mul b) == a").
3650 .. code-block:: llvm
3652 <result> = udiv i32 4, %var ; yields {i32}:result = 4 / %var
3654 '``sdiv``' Instruction
3655 ^^^^^^^^^^^^^^^^^^^^^^
3662 <result> = sdiv <ty> <op1>, <op2> ; yields {ty}:result
3663 <result> = sdiv exact <ty> <op1>, <op2> ; yields {ty}:result
3668 The '``sdiv``' instruction returns the quotient of its two operands.
3673 The two arguments to the '``sdiv``' instruction must be
3674 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3675 arguments must have identical types.
3680 The value produced is the signed integer quotient of the two operands
3681 rounded towards zero.
3683 Note that signed integer division and unsigned integer division are
3684 distinct operations; for unsigned integer division, use '``udiv``'.
3686 Division by zero leads to undefined behavior. Overflow also leads to
3687 undefined behavior; this is a rare case, but can occur, for example, by
3688 doing a 32-bit division of -2147483648 by -1.
3690 If the ``exact`` keyword is present, the result value of the ``sdiv`` is
3691 a :ref:`poison value <poisonvalues>` if the result would be rounded.
3696 .. code-block:: llvm
3698 <result> = sdiv i32 4, %var ; yields {i32}:result = 4 / %var
3702 '``fdiv``' Instruction
3703 ^^^^^^^^^^^^^^^^^^^^^^
3710 <result> = fdiv [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3715 The '``fdiv``' instruction returns the quotient of its two operands.
3720 The two arguments to the '``fdiv``' instruction must be :ref:`floating
3721 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3722 Both arguments must have identical types.
3727 The value produced is the floating point quotient of the two operands.
3728 This instruction can also take any number of :ref:`fast-math
3729 flags <fastmath>`, which are optimization hints to enable otherwise
3730 unsafe floating point optimizations:
3735 .. code-block:: llvm
3737 <result> = fdiv float 4.0, %var ; yields {float}:result = 4.0 / %var
3739 '``urem``' Instruction
3740 ^^^^^^^^^^^^^^^^^^^^^^
3747 <result> = urem <ty> <op1>, <op2> ; yields {ty}:result
3752 The '``urem``' instruction returns the remainder from the unsigned
3753 division of its two arguments.
3758 The two arguments to the '``urem``' instruction must be
3759 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3760 arguments must have identical types.
3765 This instruction returns the unsigned integer *remainder* of a division.
3766 This instruction always performs an unsigned division to get the
3769 Note that unsigned integer remainder and signed integer remainder are
3770 distinct operations; for signed integer remainder, use '``srem``'.
3772 Taking the remainder of a division by zero leads to undefined behavior.
3777 .. code-block:: llvm
3779 <result> = urem i32 4, %var ; yields {i32}:result = 4 % %var
3781 '``srem``' Instruction
3782 ^^^^^^^^^^^^^^^^^^^^^^
3789 <result> = srem <ty> <op1>, <op2> ; yields {ty}:result
3794 The '``srem``' instruction returns the remainder from the signed
3795 division of its two operands. This instruction can also take
3796 :ref:`vector <t_vector>` versions of the values in which case the elements
3802 The two arguments to the '``srem``' instruction must be
3803 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
3804 arguments must have identical types.
3809 This instruction returns the *remainder* of a division (where the result
3810 is either zero or has the same sign as the dividend, ``op1``), not the
3811 *modulo* operator (where the result is either zero or has the same sign
3812 as the divisor, ``op2``) of a value. For more information about the
3813 difference, see `The Math
3814 Forum <http://mathforum.org/dr.math/problems/anne.4.28.99.html>`_. For a
3815 table of how this is implemented in various languages, please see
3817 operation <http://en.wikipedia.org/wiki/Modulo_operation>`_.
3819 Note that signed integer remainder and unsigned integer remainder are
3820 distinct operations; for unsigned integer remainder, use '``urem``'.
3822 Taking the remainder of a division by zero leads to undefined behavior.
3823 Overflow also leads to undefined behavior; this is a rare case, but can
3824 occur, for example, by taking the remainder of a 32-bit division of
3825 -2147483648 by -1. (The remainder doesn't actually overflow, but this
3826 rule lets srem be implemented using instructions that return both the
3827 result of the division and the remainder.)
3832 .. code-block:: llvm
3834 <result> = srem i32 4, %var ; yields {i32}:result = 4 % %var
3838 '``frem``' Instruction
3839 ^^^^^^^^^^^^^^^^^^^^^^
3846 <result> = frem [fast-math flags]* <ty> <op1>, <op2> ; yields {ty}:result
3851 The '``frem``' instruction returns the remainder from the division of
3857 The two arguments to the '``frem``' instruction must be :ref:`floating
3858 point <t_floating>` or :ref:`vector <t_vector>` of floating point values.
3859 Both arguments must have identical types.
3864 This instruction returns the *remainder* of a division. The remainder
3865 has the same sign as the dividend. This instruction can also take any
3866 number of :ref:`fast-math flags <fastmath>`, which are optimization hints
3867 to enable otherwise unsafe floating point optimizations:
3872 .. code-block:: llvm
3874 <result> = frem float 4.0, %var ; yields {float}:result = 4.0 % %var
3878 Bitwise Binary Operations
3879 -------------------------
3881 Bitwise binary operators are used to do various forms of bit-twiddling
3882 in a program. They are generally very efficient instructions and can
3883 commonly be strength reduced from other instructions. They require two
3884 operands of the same type, execute an operation on them, and produce a
3885 single value. The resulting value is the same type as its operands.
3887 '``shl``' Instruction
3888 ^^^^^^^^^^^^^^^^^^^^^
3895 <result> = shl <ty> <op1>, <op2> ; yields {ty}:result
3896 <result> = shl nuw <ty> <op1>, <op2> ; yields {ty}:result
3897 <result> = shl nsw <ty> <op1>, <op2> ; yields {ty}:result
3898 <result> = shl nuw nsw <ty> <op1>, <op2> ; yields {ty}:result
3903 The '``shl``' instruction returns the first operand shifted to the left
3904 a specified number of bits.
3909 Both arguments to the '``shl``' instruction must be the same
3910 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
3911 '``op2``' is treated as an unsigned value.
3916 The value produced is ``op1`` \* 2\ :sup:`op2` mod 2\ :sup:`n`,
3917 where ``n`` is the width of the result. If ``op2`` is (statically or
3918 dynamically) negative or equal to or larger than the number of bits in
3919 ``op1``, the result is undefined. If the arguments are vectors, each
3920 vector element of ``op1`` is shifted by the corresponding shift amount
3923 If the ``nuw`` keyword is present, then the shift produces a :ref:`poison
3924 value <poisonvalues>` if it shifts out any non-zero bits. If the
3925 ``nsw`` keyword is present, then the shift produces a :ref:`poison
3926 value <poisonvalues>` if it shifts out any bits that disagree with the
3927 resultant sign bit. As such, NUW/NSW have the same semantics as they
3928 would if the shift were expressed as a mul instruction with the same
3929 nsw/nuw bits in (mul %op1, (shl 1, %op2)).
3934 .. code-block:: llvm
3936 <result> = shl i32 4, %var ; yields {i32}: 4 << %var
3937 <result> = shl i32 4, 2 ; yields {i32}: 16
3938 <result> = shl i32 1, 10 ; yields {i32}: 1024
3939 <result> = shl i32 1, 32 ; undefined
3940 <result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 2, i32 4>
3942 '``lshr``' Instruction
3943 ^^^^^^^^^^^^^^^^^^^^^^
3950 <result> = lshr <ty> <op1>, <op2> ; yields {ty}:result
3951 <result> = lshr exact <ty> <op1>, <op2> ; yields {ty}:result
3956 The '``lshr``' instruction (logical shift right) returns the first
3957 operand shifted to the right a specified number of bits with zero fill.
3962 Both arguments to the '``lshr``' instruction must be the same
3963 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
3964 '``op2``' is treated as an unsigned value.
3969 This instruction always performs a logical shift right operation. The
3970 most significant bits of the result will be filled with zero bits after
3971 the shift. If ``op2`` is (statically or dynamically) equal to or larger
3972 than the number of bits in ``op1``, the result is undefined. If the
3973 arguments are vectors, each vector element of ``op1`` is shifted by the
3974 corresponding shift amount in ``op2``.
3976 If the ``exact`` keyword is present, the result value of the ``lshr`` is
3977 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
3983 .. code-block:: llvm
3985 <result> = lshr i32 4, 1 ; yields {i32}:result = 2
3986 <result> = lshr i32 4, 2 ; yields {i32}:result = 1
3987 <result> = lshr i8 4, 3 ; yields {i8}:result = 0
3988 <result> = lshr i8 -2, 1 ; yields {i8}:result = 0x7FFFFFFF
3989 <result> = lshr i32 1, 32 ; undefined
3990 <result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> ; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1>
3992 '``ashr``' Instruction
3993 ^^^^^^^^^^^^^^^^^^^^^^
4000 <result> = ashr <ty> <op1>, <op2> ; yields {ty}:result
4001 <result> = ashr exact <ty> <op1>, <op2> ; yields {ty}:result
4006 The '``ashr``' instruction (arithmetic shift right) returns the first
4007 operand shifted to the right a specified number of bits with sign
4013 Both arguments to the '``ashr``' instruction must be the same
4014 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer type.
4015 '``op2``' is treated as an unsigned value.
4020 This instruction always performs an arithmetic shift right operation,
4021 The most significant bits of the result will be filled with the sign bit
4022 of ``op1``. If ``op2`` is (statically or dynamically) equal to or larger
4023 than the number of bits in ``op1``, the result is undefined. If the
4024 arguments are vectors, each vector element of ``op1`` is shifted by the
4025 corresponding shift amount in ``op2``.
4027 If the ``exact`` keyword is present, the result value of the ``ashr`` is
4028 a :ref:`poison value <poisonvalues>` if any of the bits shifted out are
4034 .. code-block:: llvm
4036 <result> = ashr i32 4, 1 ; yields {i32}:result = 2
4037 <result> = ashr i32 4, 2 ; yields {i32}:result = 1
4038 <result> = ashr i8 4, 3 ; yields {i8}:result = 0
4039 <result> = ashr i8 -2, 1 ; yields {i8}:result = -1
4040 <result> = ashr i32 1, 32 ; undefined
4041 <result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> ; yields: result=<2 x i32> < i32 -1, i32 0>
4043 '``and``' Instruction
4044 ^^^^^^^^^^^^^^^^^^^^^
4051 <result> = and <ty> <op1>, <op2> ; yields {ty}:result
4056 The '``and``' instruction returns the bitwise logical and of its two
4062 The two arguments to the '``and``' instruction must be
4063 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4064 arguments must have identical types.
4069 The truth table used for the '``and``' instruction is:
4086 .. code-block:: llvm
4088 <result> = and i32 4, %var ; yields {i32}:result = 4 & %var
4089 <result> = and i32 15, 40 ; yields {i32}:result = 8
4090 <result> = and i32 4, 8 ; yields {i32}:result = 0
4092 '``or``' Instruction
4093 ^^^^^^^^^^^^^^^^^^^^
4100 <result> = or <ty> <op1>, <op2> ; yields {ty}:result
4105 The '``or``' instruction returns the bitwise logical inclusive or of its
4111 The two arguments to the '``or``' instruction must be
4112 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4113 arguments must have identical types.
4118 The truth table used for the '``or``' instruction is:
4137 <result> = or i32 4, %var ; yields {i32}:result = 4 | %var
4138 <result> = or i32 15, 40 ; yields {i32}:result = 47
4139 <result> = or i32 4, 8 ; yields {i32}:result = 12
4141 '``xor``' Instruction
4142 ^^^^^^^^^^^^^^^^^^^^^
4149 <result> = xor <ty> <op1>, <op2> ; yields {ty}:result
4154 The '``xor``' instruction returns the bitwise logical exclusive or of
4155 its two operands. The ``xor`` is used to implement the "one's
4156 complement" operation, which is the "~" operator in C.
4161 The two arguments to the '``xor``' instruction must be
4162 :ref:`integer <t_integer>` or :ref:`vector <t_vector>` of integer values. Both
4163 arguments must have identical types.
4168 The truth table used for the '``xor``' instruction is:
4185 .. code-block:: llvm
4187 <result> = xor i32 4, %var ; yields {i32}:result = 4 ^ %var
4188 <result> = xor i32 15, 40 ; yields {i32}:result = 39
4189 <result> = xor i32 4, 8 ; yields {i32}:result = 12
4190 <result> = xor i32 %V, -1 ; yields {i32}:result = ~%V
4195 LLVM supports several instructions to represent vector operations in a
4196 target-independent manner. These instructions cover the element-access
4197 and vector-specific operations needed to process vectors effectively.
4198 While LLVM does directly support these vector operations, many
4199 sophisticated algorithms will want to use target-specific intrinsics to
4200 take full advantage of a specific target.
4202 .. _i_extractelement:
4204 '``extractelement``' Instruction
4205 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4212 <result> = extractelement <n x <ty>> <val>, i32 <idx> ; yields <ty>
4217 The '``extractelement``' instruction extracts a single scalar element
4218 from a vector at a specified index.
4223 The first operand of an '``extractelement``' instruction is a value of
4224 :ref:`vector <t_vector>` type. The second operand is an index indicating
4225 the position from which to extract the element. The index may be a
4231 The result is a scalar of the same type as the element type of ``val``.
4232 Its value is the value at position ``idx`` of ``val``. If ``idx``
4233 exceeds the length of ``val``, the results are undefined.
4238 .. code-block:: llvm
4240 <result> = extractelement <4 x i32> %vec, i32 0 ; yields i32
4242 .. _i_insertelement:
4244 '``insertelement``' Instruction
4245 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4252 <result> = insertelement <n x <ty>> <val>, <ty> <elt>, i32 <idx> ; yields <n x <ty>>
4257 The '``insertelement``' instruction inserts a scalar element into a
4258 vector at a specified index.
4263 The first operand of an '``insertelement``' instruction is a value of
4264 :ref:`vector <t_vector>` type. The second operand is a scalar value whose
4265 type must equal the element type of the first operand. The third operand
4266 is an index indicating the position at which to insert the value. The
4267 index may be a variable.
4272 The result is a vector of the same type as ``val``. Its element values
4273 are those of ``val`` except at position ``idx``, where it gets the value
4274 ``elt``. If ``idx`` exceeds the length of ``val``, the results are
4280 .. code-block:: llvm
4282 <result> = insertelement <4 x i32> %vec, i32 1, i32 0 ; yields <4 x i32>
4284 .. _i_shufflevector:
4286 '``shufflevector``' Instruction
4287 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4294 <result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> ; yields <m x <ty>>
4299 The '``shufflevector``' instruction constructs a permutation of elements
4300 from two input vectors, returning a vector with the same element type as
4301 the input and length that is the same as the shuffle mask.
4306 The first two operands of a '``shufflevector``' instruction are vectors
4307 with the same type. The third argument is a shuffle mask whose element
4308 type is always 'i32'. The result of the instruction is a vector whose
4309 length is the same as the shuffle mask and whose element type is the
4310 same as the element type of the first two operands.
4312 The shuffle mask operand is required to be a constant vector with either
4313 constant integer or undef values.
4318 The elements of the two input vectors are numbered from left to right
4319 across both of the vectors. The shuffle mask operand specifies, for each
4320 element of the result vector, which element of the two input vectors the
4321 result element gets. The element selector may be undef (meaning "don't
4322 care") and the second operand may be undef if performing a shuffle from
4328 .. code-block:: llvm
4330 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4331 <4 x i32> <i32 0, i32 4, i32 1, i32 5> ; yields <4 x i32>
4332 <result> = shufflevector <4 x i32> %v1, <4 x i32> undef,
4333 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32> - Identity shuffle.
4334 <result> = shufflevector <8 x i32> %v1, <8 x i32> undef,
4335 <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; yields <4 x i32>
4336 <result> = shufflevector <4 x i32> %v1, <4 x i32> %v2,
4337 <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > ; yields <8 x i32>
4339 Aggregate Operations
4340 --------------------
4342 LLVM supports several instructions for working with
4343 :ref:`aggregate <t_aggregate>` values.
4347 '``extractvalue``' Instruction
4348 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4355 <result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
4360 The '``extractvalue``' instruction extracts the value of a member field
4361 from an :ref:`aggregate <t_aggregate>` value.
4366 The first operand of an '``extractvalue``' instruction is a value of
4367 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The operands are
4368 constant indices to specify which value to extract in a similar manner
4369 as indices in a '``getelementptr``' instruction.
4371 The major differences to ``getelementptr`` indexing are:
4373 - Since the value being indexed is not a pointer, the first index is
4374 omitted and assumed to be zero.
4375 - At least one index must be specified.
4376 - Not only struct indices but also array indices must be in bounds.
4381 The result is the value at the position in the aggregate specified by
4387 .. code-block:: llvm
4389 <result> = extractvalue {i32, float} %agg, 0 ; yields i32
4393 '``insertvalue``' Instruction
4394 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4401 <result> = insertvalue <aggregate type> <val>, <ty> <elt>, <idx>{, <idx>}* ; yields <aggregate type>
4406 The '``insertvalue``' instruction inserts a value into a member field in
4407 an :ref:`aggregate <t_aggregate>` value.
4412 The first operand of an '``insertvalue``' instruction is a value of
4413 :ref:`struct <t_struct>` or :ref:`array <t_array>` type. The second operand is
4414 a first-class value to insert. The following operands are constant
4415 indices indicating the position at which to insert the value in a
4416 similar manner as indices in a '``extractvalue``' instruction. The value
4417 to insert must have the same type as the value identified by the
4423 The result is an aggregate of the same type as ``val``. Its value is
4424 that of ``val`` except that the value at the position specified by the
4425 indices is that of ``elt``.
4430 .. code-block:: llvm
4432 %agg1 = insertvalue {i32, float} undef, i32 1, 0 ; yields {i32 1, float undef}
4433 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 ; yields {i32 1, float %val}
4434 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 ; yields {i32 1, float %val}
4438 Memory Access and Addressing Operations
4439 ---------------------------------------
4441 A key design point of an SSA-based representation is how it represents
4442 memory. In LLVM, no memory locations are in SSA form, which makes things
4443 very simple. This section describes how to read, write, and allocate
4448 '``alloca``' Instruction
4449 ^^^^^^^^^^^^^^^^^^^^^^^^
4456 <result> = alloca <type>[, <ty> <NumElements>][, align <alignment>] ; yields {type*}:result
4461 The '``alloca``' instruction allocates memory on the stack frame of the
4462 currently executing function, to be automatically released when this
4463 function returns to its caller. The object is always allocated in the
4464 generic address space (address space zero).
4469 The '``alloca``' instruction allocates ``sizeof(<type>)*NumElements``
4470 bytes of memory on the runtime stack, returning a pointer of the
4471 appropriate type to the program. If "NumElements" is specified, it is
4472 the number of elements allocated, otherwise "NumElements" is defaulted
4473 to be one. If a constant alignment is specified, the value result of the
4474 allocation is guaranteed to be aligned to at least that boundary. If not
4475 specified, or if zero, the target can choose to align the allocation on
4476 any convenient boundary compatible with the type.
4478 '``type``' may be any sized type.
4483 Memory is allocated; a pointer is returned. The operation is undefined
4484 if there is insufficient stack space for the allocation. '``alloca``'d
4485 memory is automatically released when the function returns. The
4486 '``alloca``' instruction is commonly used to represent automatic
4487 variables that must have an address available. When the function returns
4488 (either with the ``ret`` or ``resume`` instructions), the memory is
4489 reclaimed. Allocating zero bytes is legal, but the result is undefined.
4490 The order in which memory is allocated (ie., which way the stack grows)
4496 .. code-block:: llvm
4498 %ptr = alloca i32 ; yields {i32*}:ptr
4499 %ptr = alloca i32, i32 4 ; yields {i32*}:ptr
4500 %ptr = alloca i32, i32 4, align 1024 ; yields {i32*}:ptr
4501 %ptr = alloca i32, align 1024 ; yields {i32*}:ptr
4505 '``load``' Instruction
4506 ^^^^^^^^^^^^^^^^^^^^^^
4513 <result> = load [volatile] <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>][, !invariant.load !<index>]
4514 <result> = load atomic [volatile] <ty>* <pointer> [singlethread] <ordering>, align <alignment>
4515 !<index> = !{ i32 1 }
4520 The '``load``' instruction is used to read from memory.
4525 The argument to the '``load``' instruction specifies the memory address
4526 from which to load. The pointer must point to a :ref:`first
4527 class <t_firstclass>` type. If the ``load`` is marked as ``volatile``,
4528 then the optimizer is not allowed to modify the number or order of
4529 execution of this ``load`` with other :ref:`volatile
4530 operations <volatile>`.
4532 If the ``load`` is marked as ``atomic``, it takes an extra
4533 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
4534 ``release`` and ``acq_rel`` orderings are not valid on ``load``
4535 instructions. Atomic loads produce :ref:`defined <memmodel>` results
4536 when they may see multiple atomic stores. The type of the pointee must
4537 be an integer type whose bit width is a power of two greater than or
4538 equal to eight and less than or equal to a target-specific size limit.
4539 ``align`` must be explicitly specified on atomic loads, and the load has
4540 undefined behavior if the alignment is not set to a value which is at
4541 least the size in bytes of the pointee. ``!nontemporal`` does not have
4542 any defined semantics for atomic loads.
4544 The optional constant ``align`` argument specifies the alignment of the
4545 operation (that is, the alignment of the memory address). A value of 0
4546 or an omitted ``align`` argument means that the operation has the abi
4547 alignment for the target. It is the responsibility of the code emitter
4548 to ensure that the alignment information is correct. Overestimating the
4549 alignment results in undefined behavior. Underestimating the alignment
4550 may produce less efficient code. An alignment of 1 is always safe.
4552 The optional ``!nontemporal`` metadata must reference a single
4553 metatadata name <index> corresponding to a metadata node with one
4554 ``i32`` entry of value 1. The existence of the ``!nontemporal``
4555 metatadata on the instruction tells the optimizer and code generator
4556 that this load is not expected to be reused in the cache. The code
4557 generator may select special instructions to save cache bandwidth, such
4558 as the ``MOVNT`` instruction on x86.
4560 The optional ``!invariant.load`` metadata must reference a single
4561 metatadata name <index> corresponding to a metadata node with no
4562 entries. The existence of the ``!invariant.load`` metatadata on the
4563 instruction tells the optimizer and code generator that this load
4564 address points to memory which does not change value during program
4565 execution. The optimizer may then move this load around, for example, by
4566 hoisting it out of loops using loop invariant code motion.
4571 The location of memory pointed to is loaded. If the value being loaded
4572 is of scalar type then the number of bytes read does not exceed the
4573 minimum number of bytes needed to hold all bits of the type. For
4574 example, loading an ``i24`` reads at most three bytes. When loading a
4575 value of a type like ``i20`` with a size that is not an integral number
4576 of bytes, the result is undefined if the value was not originally
4577 written using a store of the same type.
4582 .. code-block:: llvm
4584 %ptr = alloca i32 ; yields {i32*}:ptr
4585 store i32 3, i32* %ptr ; yields {void}
4586 %val = load i32* %ptr ; yields {i32}:val = i32 3
4590 '``store``' Instruction
4591 ^^^^^^^^^^^^^^^^^^^^^^^
4598 store [volatile] <ty> <value>, <ty>* <pointer>[, align <alignment>][, !nontemporal !<index>] ; yields {void}
4599 store atomic [volatile] <ty> <value>, <ty>* <pointer> [singlethread] <ordering>, align <alignment> ; yields {void}
4604 The '``store``' instruction is used to write to memory.
4609 There are two arguments to the '``store``' instruction: a value to store
4610 and an address at which to store it. The type of the '``<pointer>``'
4611 operand must be a pointer to the :ref:`first class <t_firstclass>` type of
4612 the '``<value>``' operand. If the ``store`` is marked as ``volatile``,
4613 then the optimizer is not allowed to modify the number or order of
4614 execution of this ``store`` with other :ref:`volatile
4615 operations <volatile>`.
4617 If the ``store`` is marked as ``atomic``, it takes an extra
4618 :ref:`ordering <ordering>` and optional ``singlethread`` argument. The
4619 ``acquire`` and ``acq_rel`` orderings aren't valid on ``store``
4620 instructions. Atomic loads produce :ref:`defined <memmodel>` results
4621 when they may see multiple atomic stores. The type of the pointee must
4622 be an integer type whose bit width is a power of two greater than or
4623 equal to eight and less than or equal to a target-specific size limit.
4624 ``align`` must be explicitly specified on atomic stores, and the store
4625 has undefined behavior if the alignment is not set to a value which is
4626 at least the size in bytes of the pointee. ``!nontemporal`` does not
4627 have any defined semantics for atomic stores.
4629 The optional constant "align" argument specifies the alignment of the
4630 operation (that is, the alignment of the memory address). A value of 0
4631 or an omitted "align" argument means that the operation has the abi
4632 alignment for the target. It is the responsibility of the code emitter
4633 to ensure that the alignment information is correct. Overestimating the
4634 alignment results in an undefined behavior. Underestimating the
4635 alignment may produce less efficient code. An alignment of 1 is always
4638 The optional !nontemporal metadata must reference a single metatadata
4639 name <index> corresponding to a metadata node with one i32 entry of
4640 value 1. The existence of the !nontemporal metatadata on the instruction
4641 tells the optimizer and code generator that this load is not expected to
4642 be reused in the cache. The code generator may select special
4643 instructions to save cache bandwidth, such as the MOVNT instruction on
4649 The contents of memory are updated to contain '``<value>``' at the
4650 location specified by the '``<pointer>``' operand. If '``<value>``' is
4651 of scalar type then the number of bytes written does not exceed the
4652 minimum number of bytes needed to hold all bits of the type. For
4653 example, storing an ``i24`` writes at most three bytes. When writing a
4654 value of a type like ``i20`` with a size that is not an integral number
4655 of bytes, it is unspecified what happens to the extra bits that do not
4656 belong to the type, but they will typically be overwritten.
4661 .. code-block:: llvm
4663 %ptr = alloca i32 ; yields {i32*}:ptr
4664 store i32 3, i32* %ptr ; yields {void}
4665 %val = load i32* %ptr ; yields {i32}:val = i32 3
4669 '``fence``' Instruction
4670 ^^^^^^^^^^^^^^^^^^^^^^^
4677 fence [singlethread] <ordering> ; yields {void}
4682 The '``fence``' instruction is used to introduce happens-before edges
4688 '``fence``' instructions take an :ref:`ordering <ordering>` argument which
4689 defines what *synchronizes-with* edges they add. They can only be given
4690 ``acquire``, ``release``, ``acq_rel``, and ``seq_cst`` orderings.
4695 A fence A which has (at least) ``release`` ordering semantics
4696 *synchronizes with* a fence B with (at least) ``acquire`` ordering
4697 semantics if and only if there exist atomic operations X and Y, both
4698 operating on some atomic object M, such that A is sequenced before X, X
4699 modifies M (either directly or through some side effect of a sequence
4700 headed by X), Y is sequenced before B, and Y observes M. This provides a
4701 *happens-before* dependency between A and B. Rather than an explicit
4702 ``fence``, one (but not both) of the atomic operations X or Y might
4703 provide a ``release`` or ``acquire`` (resp.) ordering constraint and
4704 still *synchronize-with* the explicit ``fence`` and establish the
4705 *happens-before* edge.
4707 A ``fence`` which has ``seq_cst`` ordering, in addition to having both
4708 ``acquire`` and ``release`` semantics specified above, participates in
4709 the global program order of other ``seq_cst`` operations and/or fences.
4711 The optional ":ref:`singlethread <singlethread>`" argument specifies
4712 that the fence only synchronizes with other fences in the same thread.
4713 (This is useful for interacting with signal handlers.)
4718 .. code-block:: llvm
4720 fence acquire ; yields {void}
4721 fence singlethread seq_cst ; yields {void}
4725 '``cmpxchg``' Instruction
4726 ^^^^^^^^^^^^^^^^^^^^^^^^^
4733 cmpxchg [volatile] <ty>* <pointer>, <ty> <cmp>, <ty> <new> [singlethread] <ordering> ; yields {ty}
4738 The '``cmpxchg``' instruction is used to atomically modify memory. It
4739 loads a value in memory and compares it to a given value. If they are
4740 equal, it stores a new value into the memory.
4745 There are three arguments to the '``cmpxchg``' instruction: an address
4746 to operate on, a value to compare to the value currently be at that
4747 address, and a new value to place at that address if the compared values
4748 are equal. The type of '<cmp>' must be an integer type whose bit width
4749 is a power of two greater than or equal to eight and less than or equal
4750 to a target-specific size limit. '<cmp>' and '<new>' must have the same
4751 type, and the type of '<pointer>' must be a pointer to that type. If the
4752 ``cmpxchg`` is marked as ``volatile``, then the optimizer is not allowed
4753 to modify the number or order of execution of this ``cmpxchg`` with
4754 other :ref:`volatile operations <volatile>`.
4756 The :ref:`ordering <ordering>` argument specifies how this ``cmpxchg``
4757 synchronizes with other atomic operations.
4759 The optional "``singlethread``" argument declares that the ``cmpxchg``
4760 is only atomic with respect to code (usually signal handlers) running in
4761 the same thread as the ``cmpxchg``. Otherwise the cmpxchg is atomic with
4762 respect to all other code in the system.
4764 The pointer passed into cmpxchg must have alignment greater than or
4765 equal to the size in memory of the operand.
4770 The contents of memory at the location specified by the '``<pointer>``'
4771 operand is read and compared to '``<cmp>``'; if the read value is the
4772 equal, '``<new>``' is written. The original value at the location is
4775 A successful ``cmpxchg`` is a read-modify-write instruction for the purpose
4776 of identifying release sequences. A failed ``cmpxchg`` is equivalent to an
4777 atomic load with an ordering parameter determined by dropping any
4778 ``release`` part of the ``cmpxchg``'s ordering.
4783 .. code-block:: llvm
4786 %orig = atomic load i32* %ptr unordered ; yields {i32}
4790 %cmp = phi i32 [ %orig, %entry ], [%old, %loop]
4791 %squared = mul i32 %cmp, %cmp
4792 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared ; yields {i32}
4793 %success = icmp eq i32 %cmp, %old
4794 br i1 %success, label %done, label %loop
4801 '``atomicrmw``' Instruction
4802 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
4809 atomicrmw [volatile] <operation> <ty>* <pointer>, <ty> <value> [singlethread] <ordering> ; yields {ty}
4814 The '``atomicrmw``' instruction is used to atomically modify memory.
4819 There are three arguments to the '``atomicrmw``' instruction: an
4820 operation to apply, an address whose value to modify, an argument to the
4821 operation. The operation must be one of the following keywords:
4835 The type of '<value>' must be an integer type whose bit width is a power
4836 of two greater than or equal to eight and less than or equal to a
4837 target-specific size limit. The type of the '``<pointer>``' operand must
4838 be a pointer to that type. If the ``atomicrmw`` is marked as
4839 ``volatile``, then the optimizer is not allowed to modify the number or
4840 order of execution of this ``atomicrmw`` with other :ref:`volatile
4841 operations <volatile>`.
4846 The contents of memory at the location specified by the '``<pointer>``'
4847 operand are atomically read, modified, and written back. The original
4848 value at the location is returned. The modification is specified by the
4851 - xchg: ``*ptr = val``
4852 - add: ``*ptr = *ptr + val``
4853 - sub: ``*ptr = *ptr - val``
4854 - and: ``*ptr = *ptr & val``
4855 - nand: ``*ptr = ~(*ptr & val)``
4856 - or: ``*ptr = *ptr | val``
4857 - xor: ``*ptr = *ptr ^ val``
4858 - max: ``*ptr = *ptr > val ? *ptr : val`` (using a signed comparison)
4859 - min: ``*ptr = *ptr < val ? *ptr : val`` (using a signed comparison)
4860 - umax: ``*ptr = *ptr > val ? *ptr : val`` (using an unsigned
4862 - umin: ``*ptr = *ptr < val ? *ptr : val`` (using an unsigned
4868 .. code-block:: llvm
4870 %old = atomicrmw add i32* %ptr, i32 1 acquire ; yields {i32}
4872 .. _i_getelementptr:
4874 '``getelementptr``' Instruction
4875 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
4882 <result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
4883 <result> = getelementptr inbounds <pty>* <ptrval>{, <ty> <idx>}*
4884 <result> = getelementptr <ptr vector> ptrval, <vector index type> idx
4889 The '``getelementptr``' instruction is used to get the address of a
4890 subelement of an :ref:`aggregate <t_aggregate>` data structure. It performs
4891 address calculation only and does not access memory.
4896 The first argument is always a pointer or a vector of pointers, and
4897 forms the basis of the calculation. The remaining arguments are indices
4898 that indicate which of the elements of the aggregate object are indexed.
4899 The interpretation of each index is dependent on the type being indexed
4900 into. The first index always indexes the pointer value given as the
4901 first argument, the second index indexes a value of the type pointed to
4902 (not necessarily the value directly pointed to, since the first index
4903 can be non-zero), etc. The first type indexed into must be a pointer
4904 value, subsequent types can be arrays, vectors, and structs. Note that
4905 subsequent types being indexed into can never be pointers, since that
4906 would require loading the pointer before continuing calculation.
4908 The type of each index argument depends on the type it is indexing into.
4909 When indexing into a (optionally packed) structure, only ``i32`` integer
4910 **constants** are allowed (when using a vector of indices they must all
4911 be the **same** ``i32`` integer constant). When indexing into an array,
4912 pointer or vector, integers of any width are allowed, and they are not
4913 required to be constant. These integers are treated as signed values
4916 For example, let's consider a C code fragment and how it gets compiled
4932 int *foo(struct ST *s) {
4933 return &s[1].Z.B[5][13];
4936 The LLVM code generated by Clang is:
4938 .. code-block:: llvm
4940 %struct.RT = type { i8, [10 x [20 x i32]], i8 }
4941 %struct.ST = type { i32, double, %struct.RT }
4943 define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
4945 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
4952 In the example above, the first index is indexing into the
4953 '``%struct.ST*``' type, which is a pointer, yielding a '``%struct.ST``'
4954 = '``{ i32, double, %struct.RT }``' type, a structure. The second index
4955 indexes into the third element of the structure, yielding a
4956 '``%struct.RT``' = '``{ i8 , [10 x [20 x i32]], i8 }``' type, another
4957 structure. The third index indexes into the second element of the
4958 structure, yielding a '``[10 x [20 x i32]]``' type, an array. The two
4959 dimensions of the array are subscripted into, yielding an '``i32``'
4960 type. The '``getelementptr``' instruction returns a pointer to this
4961 element, thus computing a value of '``i32*``' type.
4963 Note that it is perfectly legal to index partially through a structure,
4964 returning a pointer to an inner element. Because of this, the LLVM code
4965 for the given testcase is equivalent to:
4967 .. code-block:: llvm
4969 define i32* @foo(%struct.ST* %s) {
4970 %t1 = getelementptr %struct.ST* %s, i32 1 ; yields %struct.ST*:%t1
4971 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 ; yields %struct.RT*:%t2
4972 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 ; yields [10 x [20 x i32]]*:%t3
4973 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 ; yields [20 x i32]*:%t4
4974 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 ; yields i32*:%t5
4978 If the ``inbounds`` keyword is present, the result value of the
4979 ``getelementptr`` is a :ref:`poison value <poisonvalues>` if the base
4980 pointer is not an *in bounds* address of an allocated object, or if any
4981 of the addresses that would be formed by successive addition of the
4982 offsets implied by the indices to the base address with infinitely
4983 precise signed arithmetic are not an *in bounds* address of that
4984 allocated object. The *in bounds* addresses for an allocated object are
4985 all the addresses that point into the object, plus the address one byte
4986 past the end. In cases where the base is a vector of pointers the
4987 ``inbounds`` keyword applies to each of the computations element-wise.
4989 If the ``inbounds`` keyword is not present, the offsets are added to the
4990 base address with silently-wrapping two's complement arithmetic. If the
4991 offsets have a different width from the pointer, they are sign-extended
4992 or truncated to the width of the pointer. The result value of the
4993 ``getelementptr`` may be outside the object pointed to by the base
4994 pointer. The result value may not necessarily be used to access memory
4995 though, even if it happens to point into allocated storage. See the
4996 :ref:`Pointer Aliasing Rules <pointeraliasing>` section for more
4999 The getelementptr instruction is often confusing. For some more insight
5000 into how it works, see :doc:`the getelementptr FAQ <GetElementPtr>`.
5005 .. code-block:: llvm
5007 ; yields [12 x i8]*:aptr
5008 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5010 %vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
5012 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
5014 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
5016 In cases where the pointer argument is a vector of pointers, each index
5017 must be a vector with the same number of elements. For example:
5019 .. code-block:: llvm
5021 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5023 Conversion Operations
5024 ---------------------
5026 The instructions in this category are the conversion instructions
5027 (casting) which all take a single operand and a type. They perform
5028 various bit conversions on the operand.
5030 '``trunc .. to``' Instruction
5031 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5038 <result> = trunc <ty> <value> to <ty2> ; yields ty2
5043 The '``trunc``' instruction truncates its operand to the type ``ty2``.
5048 The '``trunc``' instruction takes a value to trunc, and a type to trunc
5049 it to. Both types must be of :ref:`integer <t_integer>` types, or vectors
5050 of the same number of integers. The bit size of the ``value`` must be
5051 larger than the bit size of the destination type, ``ty2``. Equal sized
5052 types are not allowed.
5057 The '``trunc``' instruction truncates the high order bits in ``value``
5058 and converts the remaining bits to ``ty2``. Since the source size must
5059 be larger than the destination size, ``trunc`` cannot be a *no-op cast*.
5060 It will always truncate bits.
5065 .. code-block:: llvm
5067 %X = trunc i32 257 to i8 ; yields i8:1
5068 %Y = trunc i32 123 to i1 ; yields i1:true
5069 %Z = trunc i32 122 to i1 ; yields i1:false
5070 %W = trunc <2 x i16> <i16 8, i16 7> to <2 x i8> ; yields <i8 8, i8 7>
5072 '``zext .. to``' Instruction
5073 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5080 <result> = zext <ty> <value> to <ty2> ; yields ty2
5085 The '``zext``' instruction zero extends its operand to type ``ty2``.
5090 The '``zext``' instruction takes a value to cast, and a type to cast it
5091 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5092 the same number of integers. The bit size of the ``value`` must be
5093 smaller than the bit size of the destination type, ``ty2``.
5098 The ``zext`` fills the high order bits of the ``value`` with zero bits
5099 until it reaches the size of the destination type, ``ty2``.
5101 When zero extending from i1, the result will always be either 0 or 1.
5106 .. code-block:: llvm
5108 %X = zext i32 257 to i64 ; yields i64:257
5109 %Y = zext i1 true to i32 ; yields i32:1
5110 %Z = zext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5112 '``sext .. to``' Instruction
5113 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5120 <result> = sext <ty> <value> to <ty2> ; yields ty2
5125 The '``sext``' sign extends ``value`` to the type ``ty2``.
5130 The '``sext``' instruction takes a value to cast, and a type to cast it
5131 to. Both types must be of :ref:`integer <t_integer>` types, or vectors of
5132 the same number of integers. The bit size of the ``value`` must be
5133 smaller than the bit size of the destination type, ``ty2``.
5138 The '``sext``' instruction performs a sign extension by copying the sign
5139 bit (highest order bit) of the ``value`` until it reaches the bit size
5140 of the type ``ty2``.
5142 When sign extending from i1, the extension always results in -1 or 0.
5147 .. code-block:: llvm
5149 %X = sext i8 -1 to i16 ; yields i16 :65535
5150 %Y = sext i1 true to i32 ; yields i32:-1
5151 %Z = sext <2 x i16> <i16 8, i16 7> to <2 x i32> ; yields <i32 8, i32 7>
5153 '``fptrunc .. to``' Instruction
5154 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5161 <result> = fptrunc <ty> <value> to <ty2> ; yields ty2
5166 The '``fptrunc``' instruction truncates ``value`` to type ``ty2``.
5171 The '``fptrunc``' instruction takes a :ref:`floating point <t_floating>`
5172 value to cast and a :ref:`floating point <t_floating>` type to cast it to.
5173 The size of ``value`` must be larger than the size of ``ty2``. This
5174 implies that ``fptrunc`` cannot be used to make a *no-op cast*.
5179 The '``fptrunc``' instruction truncates a ``value`` from a larger
5180 :ref:`floating point <t_floating>` type to a smaller :ref:`floating
5181 point <t_floating>` type. If the value cannot fit within the
5182 destination type, ``ty2``, then the results are undefined.
5187 .. code-block:: llvm
5189 %X = fptrunc double 123.0 to float ; yields float:123.0
5190 %Y = fptrunc double 1.0E+300 to float ; yields undefined
5192 '``fpext .. to``' Instruction
5193 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5200 <result> = fpext <ty> <value> to <ty2> ; yields ty2
5205 The '``fpext``' extends a floating point ``value`` to a larger floating
5211 The '``fpext``' instruction takes a :ref:`floating point <t_floating>`
5212 ``value`` to cast, and a :ref:`floating point <t_floating>` type to cast it
5213 to. The source type must be smaller than the destination type.
5218 The '``fpext``' instruction extends the ``value`` from a smaller
5219 :ref:`floating point <t_floating>` type to a larger :ref:`floating
5220 point <t_floating>` type. The ``fpext`` cannot be used to make a
5221 *no-op cast* because it always changes bits. Use ``bitcast`` to make a
5222 *no-op cast* for a floating point cast.
5227 .. code-block:: llvm
5229 %X = fpext float 3.125 to double ; yields double:3.125000e+00
5230 %Y = fpext double %X to fp128 ; yields fp128:0xL00000000000000004000900000000000
5232 '``fptoui .. to``' Instruction
5233 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5240 <result> = fptoui <ty> <value> to <ty2> ; yields ty2
5245 The '``fptoui``' converts a floating point ``value`` to its unsigned
5246 integer equivalent of type ``ty2``.
5251 The '``fptoui``' instruction takes a value to cast, which must be a
5252 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5253 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5254 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5255 type with the same number of elements as ``ty``
5260 The '``fptoui``' instruction converts its :ref:`floating
5261 point <t_floating>` operand into the nearest (rounding towards zero)
5262 unsigned integer value. If the value cannot fit in ``ty2``, the results
5268 .. code-block:: llvm
5270 %X = fptoui double 123.0 to i32 ; yields i32:123
5271 %Y = fptoui float 1.0E+300 to i1 ; yields undefined:1
5272 %Z = fptoui float 1.04E+17 to i8 ; yields undefined:1
5274 '``fptosi .. to``' Instruction
5275 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5282 <result> = fptosi <ty> <value> to <ty2> ; yields ty2
5287 The '``fptosi``' instruction converts :ref:`floating point <t_floating>`
5288 ``value`` to type ``ty2``.
5293 The '``fptosi``' instruction takes a value to cast, which must be a
5294 scalar or vector :ref:`floating point <t_floating>` value, and a type to
5295 cast it to ``ty2``, which must be an :ref:`integer <t_integer>` type. If
5296 ``ty`` is a vector floating point type, ``ty2`` must be a vector integer
5297 type with the same number of elements as ``ty``
5302 The '``fptosi``' instruction converts its :ref:`floating
5303 point <t_floating>` operand into the nearest (rounding towards zero)
5304 signed integer value. If the value cannot fit in ``ty2``, the results
5310 .. code-block:: llvm
5312 %X = fptosi double -123.0 to i32 ; yields i32:-123
5313 %Y = fptosi float 1.0E-247 to i1 ; yields undefined:1
5314 %Z = fptosi float 1.04E+17 to i8 ; yields undefined:1
5316 '``uitofp .. to``' Instruction
5317 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5324 <result> = uitofp <ty> <value> to <ty2> ; yields ty2
5329 The '``uitofp``' instruction regards ``value`` as an unsigned integer
5330 and converts that value to the ``ty2`` type.
5335 The '``uitofp``' instruction takes a value to cast, which must be a
5336 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5337 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5338 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5339 type with the same number of elements as ``ty``
5344 The '``uitofp``' instruction interprets its operand as an unsigned
5345 integer quantity and converts it to the corresponding floating point
5346 value. If the value cannot fit in the floating point value, the results
5352 .. code-block:: llvm
5354 %X = uitofp i32 257 to float ; yields float:257.0
5355 %Y = uitofp i8 -1 to double ; yields double:255.0
5357 '``sitofp .. to``' Instruction
5358 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5365 <result> = sitofp <ty> <value> to <ty2> ; yields ty2
5370 The '``sitofp``' instruction regards ``value`` as a signed integer and
5371 converts that value to the ``ty2`` type.
5376 The '``sitofp``' instruction takes a value to cast, which must be a
5377 scalar or vector :ref:`integer <t_integer>` value, and a type to cast it to
5378 ``ty2``, which must be an :ref:`floating point <t_floating>` type. If
5379 ``ty`` is a vector integer type, ``ty2`` must be a vector floating point
5380 type with the same number of elements as ``ty``
5385 The '``sitofp``' instruction interprets its operand as a signed integer
5386 quantity and converts it to the corresponding floating point value. If
5387 the value cannot fit in the floating point value, the results are
5393 .. code-block:: llvm
5395 %X = sitofp i32 257 to float ; yields float:257.0
5396 %Y = sitofp i8 -1 to double ; yields double:-1.0
5400 '``ptrtoint .. to``' Instruction
5401 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5408 <result> = ptrtoint <ty> <value> to <ty2> ; yields ty2
5413 The '``ptrtoint``' instruction converts the pointer or a vector of
5414 pointers ``value`` to the integer (or vector of integers) type ``ty2``.
5419 The '``ptrtoint``' instruction takes a ``value`` to cast, which must be
5420 a a value of type :ref:`pointer <t_pointer>` or a vector of pointers, and a
5421 type to cast it to ``ty2``, which must be an :ref:`integer <t_integer>` or
5422 a vector of integers type.
5427 The '``ptrtoint``' instruction converts ``value`` to integer type
5428 ``ty2`` by interpreting the pointer value as an integer and either
5429 truncating or zero extending that value to the size of the integer type.
5430 If ``value`` is smaller than ``ty2`` then a zero extension is done. If
5431 ``value`` is larger than ``ty2`` then a truncation is done. If they are
5432 the same size, then nothing is done (*no-op cast*) other than a type
5438 .. code-block:: llvm
5440 %X = ptrtoint i32* %P to i8 ; yields truncation on 32-bit architecture
5441 %Y = ptrtoint i32* %P to i64 ; yields zero extension on 32-bit architecture
5442 %Z = ptrtoint <4 x i32*> %P to <4 x i64>; yields vector zero extension for a vector of addresses on 32-bit architecture
5446 '``inttoptr .. to``' Instruction
5447 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5454 <result> = inttoptr <ty> <value> to <ty2> ; yields ty2
5459 The '``inttoptr``' instruction converts an integer ``value`` to a
5460 pointer type, ``ty2``.
5465 The '``inttoptr``' instruction takes an :ref:`integer <t_integer>` value to
5466 cast, and a type to cast it to, which must be a :ref:`pointer <t_pointer>`
5472 The '``inttoptr``' instruction converts ``value`` to type ``ty2`` by
5473 applying either a zero extension or a truncation depending on the size
5474 of the integer ``value``. If ``value`` is larger than the size of a
5475 pointer then a truncation is done. If ``value`` is smaller than the size
5476 of a pointer then a zero extension is done. If they are the same size,
5477 nothing is done (*no-op cast*).
5482 .. code-block:: llvm
5484 %X = inttoptr i32 255 to i32* ; yields zero extension on 64-bit architecture
5485 %Y = inttoptr i32 255 to i32* ; yields no-op on 32-bit architecture
5486 %Z = inttoptr i64 0 to i32* ; yields truncation on 32-bit architecture
5487 %Z = inttoptr <4 x i32> %G to <4 x i8*>; yields truncation of vector G to four pointers
5491 '``bitcast .. to``' Instruction
5492 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
5499 <result> = bitcast <ty> <value> to <ty2> ; yields ty2
5504 The '``bitcast``' instruction converts ``value`` to type ``ty2`` without
5510 The '``bitcast``' instruction takes a value to cast, which must be a
5511 non-aggregate first class value, and a type to cast it to, which must
5512 also be a non-aggregate :ref:`first class <t_firstclass>` type. The bit
5513 sizes of ``value`` and the destination type, ``ty2``, must be identical.
5514 If the source type is a pointer, the destination type must also be a
5515 pointer. This instruction supports bitwise conversion of vectors to
5516 integers and to vectors of other types (as long as they have the same
5522 The '``bitcast``' instruction converts ``value`` to type ``ty2``. It is
5523 always a *no-op cast* because no bits change with this conversion. The
5524 conversion is done as if the ``value`` had been stored to memory and
5525 read back as type ``ty2``. Pointer (or vector of pointers) types may
5526 only be converted to other pointer (or vector of pointers) types with
5527 this instruction. To convert pointers to other types, use the
5528 :ref:`inttoptr <i_inttoptr>` or :ref:`ptrtoint <i_ptrtoint>` instructions
5534 .. code-block:: llvm
5536 %X = bitcast i8 255 to i8 ; yields i8 :-1
5537 %Y = bitcast i32* %x to sint* ; yields sint*:%x
5538 %Z = bitcast <2 x int> %V to i64; ; yields i64: %V
5539 %Z = bitcast <2 x i32*> %V to <2 x i64*> ; yields <2 x i64*>
5546 The instructions in this category are the "miscellaneous" instructions,
5547 which defy better classification.
5551 '``icmp``' Instruction
5552 ^^^^^^^^^^^^^^^^^^^^^^
5559 <result> = icmp <cond> <ty> <op1>, <op2> ; yields {i1} or {<N x i1>}:result
5564 The '``icmp``' instruction returns a boolean value or a vector of
5565 boolean values based on comparison of its two integer, integer vector,
5566 pointer, or pointer vector operands.
5571 The '``icmp``' instruction takes three operands. The first operand is
5572 the condition code indicating the kind of comparison to perform. It is
5573 not a value, just a keyword. The possible condition code are:
5576 #. ``ne``: not equal
5577 #. ``ugt``: unsigned greater than
5578 #. ``uge``: unsigned greater or equal
5579 #. ``ult``: unsigned less than
5580 #. ``ule``: unsigned less or equal
5581 #. ``sgt``: signed greater than
5582 #. ``sge``: signed greater or equal
5583 #. ``slt``: signed less than
5584 #. ``sle``: signed less or equal
5586 The remaining two arguments must be :ref:`integer <t_integer>` or
5587 :ref:`pointer <t_pointer>` or integer :ref:`vector <t_vector>` typed. They
5588 must also be identical types.
5593 The '``icmp``' compares ``op1`` and ``op2`` according to the condition
5594 code given as ``cond``. The comparison performed always yields either an
5595 :ref:`i1 <t_integer>` or vector of ``i1`` result, as follows:
5597 #. ``eq``: yields ``true`` if the operands are equal, ``false``
5598 otherwise. No sign interpretation is necessary or performed.
5599 #. ``ne``: yields ``true`` if the operands are unequal, ``false``
5600 otherwise. No sign interpretation is necessary or performed.
5601 #. ``ugt``: interprets the operands as unsigned values and yields
5602 ``true`` if ``op1`` is greater than ``op2``.
5603 #. ``uge``: interprets the operands as unsigned values and yields
5604 ``true`` if ``op1`` is greater than or equal to ``op2``.
5605 #. ``ult``: interprets the operands as unsigned values and yields
5606 ``true`` if ``op1`` is less than ``op2``.
5607 #. ``ule``: interprets the operands as unsigned values and yields
5608 ``true`` if ``op1`` is less than or equal to ``op2``.
5609 #. ``sgt``: interprets the operands as signed values and yields ``true``
5610 if ``op1`` is greater than ``op2``.
5611 #. ``sge``: interprets the operands as signed values and yields ``true``
5612 if ``op1`` is greater than or equal to ``op2``.
5613 #. ``slt``: interprets the operands as signed values and yields ``true``
5614 if ``op1`` is less than ``op2``.
5615 #. ``sle``: interprets the operands as signed values and yields ``true``
5616 if ``op1`` is less than or equal to ``op2``.
5618 If the operands are :ref:`pointer <t_pointer>` typed, the pointer values
5619 are compared as if they were integers.
5621 If the operands are integer vectors, then they are compared element by
5622 element. The result is an ``i1`` vector with the same number of elements
5623 as the values being compared. Otherwise, the result is an ``i1``.
5628 .. code-block:: llvm
5630 <result> = icmp eq i32 4, 5 ; yields: result=false
5631 <result> = icmp ne float* %X, %X ; yields: result=false
5632 <result> = icmp ult i16 4, 5 ; yields: result=true
5633 <result> = icmp sgt i16 4, 5 ; yields: result=false
5634 <result> = icmp ule i16 -4, 5 ; yields: result=false
5635 <result> = icmp sge i16 4, 5 ; yields: result=false
5637 Note that the code generator does not yet support vector types with the
5638 ``icmp`` instruction.
5642 '``fcmp``' Instruction
5643 ^^^^^^^^^^^^^^^^^^^^^^
5650 <result> = fcmp <cond> <ty> <op1>, <op2> ; yields {i1} or {<N x i1>}:result
5655 The '``fcmp``' instruction returns a boolean value or vector of boolean
5656 values based on comparison of its operands.
5658 If the operands are floating point scalars, then the result type is a
5659 boolean (:ref:`i1 <t_integer>`).
5661 If the operands are floating point vectors, then the result type is a
5662 vector of boolean with the same number of elements as the operands being
5668 The '``fcmp``' instruction takes three operands. The first operand is
5669 the condition code indicating the kind of comparison to perform. It is
5670 not a value, just a keyword. The possible condition code are:
5672 #. ``false``: no comparison, always returns false
5673 #. ``oeq``: ordered and equal
5674 #. ``ogt``: ordered and greater than
5675 #. ``oge``: ordered and greater than or equal
5676 #. ``olt``: ordered and less than
5677 #. ``ole``: ordered and less than or equal
5678 #. ``one``: ordered and not equal
5679 #. ``ord``: ordered (no nans)
5680 #. ``ueq``: unordered or equal
5681 #. ``ugt``: unordered or greater than
5682 #. ``uge``: unordered or greater than or equal
5683 #. ``ult``: unordered or less than
5684 #. ``ule``: unordered or less than or equal
5685 #. ``une``: unordered or not equal
5686 #. ``uno``: unordered (either nans)
5687 #. ``true``: no comparison, always returns true
5689 *Ordered* means that neither operand is a QNAN while *unordered* means
5690 that either operand may be a QNAN.
5692 Each of ``val1`` and ``val2`` arguments must be either a :ref:`floating
5693 point <t_floating>` type or a :ref:`vector <t_vector>` of floating point
5694 type. They must have identical types.
5699 The '``fcmp``' instruction compares ``op1`` and ``op2`` according to the
5700 condition code given as ``cond``. If the operands are vectors, then the
5701 vectors are compared element by element. Each comparison performed
5702 always yields an :ref:`i1 <t_integer>` result, as follows:
5704 #. ``false``: always yields ``false``, regardless of operands.
5705 #. ``oeq``: yields ``true`` if both operands are not a QNAN and ``op1``
5706 is equal to ``op2``.
5707 #. ``ogt``: yields ``true`` if both operands are not a QNAN and ``op1``
5708 is greater than ``op2``.
5709 #. ``oge``: yields ``true`` if both operands are not a QNAN and ``op1``
5710 is greater than or equal to ``op2``.
5711 #. ``olt``: yields ``true`` if both operands are not a QNAN and ``op1``
5712 is less than ``op2``.
5713 #. ``ole``: yields ``true`` if both operands are not a QNAN and ``op1``
5714 is less than or equal to ``op2``.
5715 #. ``one``: yields ``true`` if both operands are not a QNAN and ``op1``
5716 is not equal to ``op2``.
5717 #. ``ord``: yields ``true`` if both operands are not a QNAN.
5718 #. ``ueq``: yields ``true`` if either operand is a QNAN or ``op1`` is
5720 #. ``ugt``: yields ``true`` if either operand is a QNAN or ``op1`` is
5721 greater than ``op2``.
5722 #. ``uge``: yields ``true`` if either operand is a QNAN or ``op1`` is
5723 greater than or equal to ``op2``.
5724 #. ``ult``: yields ``true`` if either operand is a QNAN or ``op1`` is
5726 #. ``ule``: yields ``true`` if either operand is a QNAN or ``op1`` is
5727 less than or equal to ``op2``.
5728 #. ``une``: yields ``true`` if either operand is a QNAN or ``op1`` is
5729 not equal to ``op2``.
5730 #. ``uno``: yields ``true`` if either operand is a QNAN.
5731 #. ``true``: always yields ``true``, regardless of operands.
5736 .. code-block:: llvm
5738 <result> = fcmp oeq float 4.0, 5.0 ; yields: result=false
5739 <result> = fcmp one float 4.0, 5.0 ; yields: result=true
5740 <result> = fcmp olt float 4.0, 5.0 ; yields: result=true
5741 <result> = fcmp ueq double 1.0, 2.0 ; yields: result=false
5743 Note that the code generator does not yet support vector types with the
5744 ``fcmp`` instruction.
5748 '``phi``' Instruction
5749 ^^^^^^^^^^^^^^^^^^^^^
5756 <result> = phi <ty> [ <val0>, <label0>], ...
5761 The '``phi``' instruction is used to implement the φ node in the SSA
5762 graph representing the function.
5767 The type of the incoming values is specified with the first type field.
5768 After this, the '``phi``' instruction takes a list of pairs as
5769 arguments, with one pair for each predecessor basic block of the current
5770 block. Only values of :ref:`first class <t_firstclass>` type may be used as
5771 the value arguments to the PHI node. Only labels may be used as the
5774 There must be no non-phi instructions between the start of a basic block
5775 and the PHI instructions: i.e. PHI instructions must be first in a basic
5778 For the purposes of the SSA form, the use of each incoming value is
5779 deemed to occur on the edge from the corresponding predecessor block to
5780 the current block (but after any definition of an '``invoke``'
5781 instruction's return value on the same edge).
5786 At runtime, the '``phi``' instruction logically takes on the value
5787 specified by the pair corresponding to the predecessor basic block that
5788 executed just prior to the current block.
5793 .. code-block:: llvm
5795 Loop: ; Infinite loop that counts from 0 on up...
5796 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5797 %nextindvar = add i32 %indvar, 1
5802 '``select``' Instruction
5803 ^^^^^^^^^^^^^^^^^^^^^^^^
5810 <result> = select selty <cond>, <ty> <val1>, <ty> <val2> ; yields ty
5812 selty is either i1 or {<N x i1>}
5817 The '``select``' instruction is used to choose one value based on a
5818 condition, without branching.
5823 The '``select``' instruction requires an 'i1' value or a vector of 'i1'
5824 values indicating the condition, and two values of the same :ref:`first
5825 class <t_firstclass>` type. If the val1/val2 are vectors and the
5826 condition is a scalar, then entire vectors are selected, not individual
5832 If the condition is an i1 and it evaluates to 1, the instruction returns
5833 the first value argument; otherwise, it returns the second value
5836 If the condition is a vector of i1, then the value arguments must be
5837 vectors of the same size, and the selection is done element by element.
5842 .. code-block:: llvm
5844 %X = select i1 true, i8 17, i8 42 ; yields i8:17
5848 '``call``' Instruction
5849 ^^^^^^^^^^^^^^^^^^^^^^
5856 <result> = [tail] call [cconv] [ret attrs] <ty> [<fnty>*] <fnptrval>(<function args>) [fn attrs]
5861 The '``call``' instruction represents a simple function call.
5866 This instruction requires several arguments:
5868 #. The optional "tail" marker indicates that the callee function does
5869 not access any allocas or varargs in the caller. Note that calls may
5870 be marked "tail" even if they do not occur before a
5871 :ref:`ret <i_ret>` instruction. If the "tail" marker is present, the
5872 function call is eligible for tail call optimization, but `might not
5873 in fact be optimized into a jump <CodeGenerator.html#tailcallopt>`_.
5874 The code generator may optimize calls marked "tail" with either 1)
5875 automatic `sibling call
5876 optimization <CodeGenerator.html#sibcallopt>`_ when the caller and
5877 callee have matching signatures, or 2) forced tail call optimization
5878 when the following extra requirements are met:
5880 - Caller and callee both have the calling convention ``fastcc``.
5881 - The call is in tail position (ret immediately follows call and ret
5882 uses value of call or is void).
5883 - Option ``-tailcallopt`` is enabled, or
5884 ``llvm::GuaranteedTailCallOpt`` is ``true``.
5885 - `Platform specific constraints are
5886 met. <CodeGenerator.html#tailcallopt>`_
5888 #. The optional "cconv" marker indicates which :ref:`calling
5889 convention <callingconv>` the call should use. If none is
5890 specified, the call defaults to using C calling conventions. The
5891 calling convention of the call must match the calling convention of
5892 the target function, or else the behavior is undefined.
5893 #. The optional :ref:`Parameter Attributes <paramattrs>` list for return
5894 values. Only '``zeroext``', '``signext``', and '``inreg``' attributes
5896 #. '``ty``': the type of the call instruction itself which is also the
5897 type of the return value. Functions that return no value are marked
5899 #. '``fnty``': shall be the signature of the pointer to function value
5900 being invoked. The argument types must match the types implied by
5901 this signature. This type can be omitted if the function is not
5902 varargs and if the function type does not return a pointer to a
5904 #. '``fnptrval``': An LLVM value containing a pointer to a function to
5905 be invoked. In most cases, this is a direct function invocation, but
5906 indirect ``call``'s are just as possible, calling an arbitrary pointer
5908 #. '``function args``': argument list whose types match the function
5909 signature argument types and parameter attributes. All arguments must
5910 be of :ref:`first class <t_firstclass>` type. If the function signature
5911 indicates the function accepts a variable number of arguments, the
5912 extra arguments can be specified.
5913 #. The optional :ref:`function attributes <fnattrs>` list. Only
5914 '``noreturn``', '``nounwind``', '``readonly``' and '``readnone``'
5915 attributes are valid here.
5920 The '``call``' instruction is used to cause control flow to transfer to
5921 a specified function, with its incoming arguments bound to the specified
5922 values. Upon a '``ret``' instruction in the called function, control
5923 flow continues with the instruction after the function call, and the
5924 return value of the function is bound to the result argument.
5929 .. code-block:: llvm
5931 %retval = call i32 @test(i32 %argc)
5932 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) ; yields i32
5933 %X = tail call i32 @foo() ; yields i32
5934 %Y = tail call fastcc i32 @foo() ; yields i32
5935 call void %foo(i8 97 signext)
5937 %struct.A = type { i32, i8 }
5938 %r = call %struct.A @foo() ; yields { 32, i8 }
5939 %gr = extractvalue %struct.A %r, 0 ; yields i32
5940 %gr1 = extractvalue %struct.A %r, 1 ; yields i8
5941 %Z = call void @foo() noreturn ; indicates that %foo never returns normally
5942 %ZZ = call zeroext i32 @bar() ; Return value is %zero extended
5944 llvm treats calls to some functions with names and arguments that match
5945 the standard C99 library as being the C99 library functions, and may
5946 perform optimizations or generate code for them under that assumption.
5947 This is something we'd like to change in the future to provide better
5948 support for freestanding environments and non-C-based languages.
5952 '``va_arg``' Instruction
5953 ^^^^^^^^^^^^^^^^^^^^^^^^
5960 <resultval> = va_arg <va_list*> <arglist>, <argty>
5965 The '``va_arg``' instruction is used to access arguments passed through
5966 the "variable argument" area of a function call. It is used to implement
5967 the ``va_arg`` macro in C.
5972 This instruction takes a ``va_list*`` value and the type of the
5973 argument. It returns a value of the specified argument type and
5974 increments the ``va_list`` to point to the next argument. The actual
5975 type of ``va_list`` is target specific.
5980 The '``va_arg``' instruction loads an argument of the specified type
5981 from the specified ``va_list`` and causes the ``va_list`` to point to
5982 the next argument. For more information, see the variable argument
5983 handling :ref:`Intrinsic Functions <int_varargs>`.
5985 It is legal for this instruction to be called in a function which does
5986 not take a variable number of arguments, for example, the ``vfprintf``
5989 ``va_arg`` is an LLVM instruction instead of an :ref:`intrinsic
5990 function <intrinsics>` because it takes a type as an argument.
5995 See the :ref:`variable argument processing <int_varargs>` section.
5997 Note that the code generator does not yet fully support va\_arg on many
5998 targets. Also, it does not currently support va\_arg with aggregate
5999 types on any target.
6003 '``landingpad``' Instruction
6004 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6011 <resultval> = landingpad <resultty> personality <type> <pers_fn> <clause>+
6012 <resultval> = landingpad <resultty> personality <type> <pers_fn> cleanup <clause>*
6014 <clause> := catch <type> <value>
6015 <clause> := filter <array constant type> <array constant>
6020 The '``landingpad``' instruction is used by `LLVM's exception handling
6021 system <ExceptionHandling.html#overview>`_ to specify that a basic block
6022 is a landing pad --- one where the exception lands, and corresponds to the
6023 code found in the ``catch`` portion of a ``try``/``catch`` sequence. It
6024 defines values supplied by the personality function (``pers_fn``) upon
6025 re-entry to the function. The ``resultval`` has the type ``resultty``.
6030 This instruction takes a ``pers_fn`` value. This is the personality
6031 function associated with the unwinding mechanism. The optional
6032 ``cleanup`` flag indicates that the landing pad block is a cleanup.
6034 A ``clause`` begins with the clause type --- ``catch`` or ``filter`` --- and
6035 contains the global variable representing the "type" that may be caught
6036 or filtered respectively. Unlike the ``catch`` clause, the ``filter``
6037 clause takes an array constant as its argument. Use
6038 "``[0 x i8**] undef``" for a filter which cannot throw. The
6039 '``landingpad``' instruction must contain *at least* one ``clause`` or
6040 the ``cleanup`` flag.
6045 The '``landingpad``' instruction defines the values which are set by the
6046 personality function (``pers_fn``) upon re-entry to the function, and
6047 therefore the "result type" of the ``landingpad`` instruction. As with
6048 calling conventions, how the personality function results are
6049 represented in LLVM IR is target specific.
6051 The clauses are applied in order from top to bottom. If two
6052 ``landingpad`` instructions are merged together through inlining, the
6053 clauses from the calling function are appended to the list of clauses.
6054 When the call stack is being unwound due to an exception being thrown,
6055 the exception is compared against each ``clause`` in turn. If it doesn't
6056 match any of the clauses, and the ``cleanup`` flag is not set, then
6057 unwinding continues further up the call stack.
6059 The ``landingpad`` instruction has several restrictions:
6061 - A landing pad block is a basic block which is the unwind destination
6062 of an '``invoke``' instruction.
6063 - A landing pad block must have a '``landingpad``' instruction as its
6064 first non-PHI instruction.
6065 - There can be only one '``landingpad``' instruction within the landing
6067 - A basic block that is not a landing pad block may not include a
6068 '``landingpad``' instruction.
6069 - All '``landingpad``' instructions in a function must have the same
6070 personality function.
6075 .. code-block:: llvm
6077 ;; A landing pad which can catch an integer.
6078 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6080 ;; A landing pad that is a cleanup.
6081 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6083 ;; A landing pad which can catch an integer and can only throw a double.
6084 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6086 filter [1 x i8**] [@_ZTId]
6093 LLVM supports the notion of an "intrinsic function". These functions
6094 have well known names and semantics and are required to follow certain
6095 restrictions. Overall, these intrinsics represent an extension mechanism
6096 for the LLVM language that does not require changing all of the
6097 transformations in LLVM when adding to the language (or the bitcode
6098 reader/writer, the parser, etc...).
6100 Intrinsic function names must all start with an "``llvm.``" prefix. This
6101 prefix is reserved in LLVM for intrinsic names; thus, function names may
6102 not begin with this prefix. Intrinsic functions must always be external
6103 functions: you cannot define the body of intrinsic functions. Intrinsic
6104 functions may only be used in call or invoke instructions: it is illegal
6105 to take the address of an intrinsic function. Additionally, because
6106 intrinsic functions are part of the LLVM language, it is required if any
6107 are added that they be documented here.
6109 Some intrinsic functions can be overloaded, i.e., the intrinsic
6110 represents a family of functions that perform the same operation but on
6111 different data types. Because LLVM can represent over 8 million
6112 different integer types, overloading is used commonly to allow an
6113 intrinsic function to operate on any integer type. One or more of the
6114 argument types or the result type can be overloaded to accept any
6115 integer type. Argument types may also be defined as exactly matching a
6116 previous argument's type or the result type. This allows an intrinsic
6117 function which accepts multiple arguments, but needs all of them to be
6118 of the same type, to only be overloaded with respect to a single
6119 argument or the result.
6121 Overloaded intrinsics will have the names of its overloaded argument
6122 types encoded into its function name, each preceded by a period. Only
6123 those types which are overloaded result in a name suffix. Arguments
6124 whose type is matched against another type do not. For example, the
6125 ``llvm.ctpop`` function can take an integer of any width and returns an
6126 integer of exactly the same integer width. This leads to a family of
6127 functions such as ``i8 @llvm.ctpop.i8(i8 %val)`` and
6128 ``i29 @llvm.ctpop.i29(i29 %val)``. Only one type, the return type, is
6129 overloaded, and only one type suffix is required. Because the argument's
6130 type is matched against the return type, it does not require its own
6133 To learn how to add an intrinsic function, please see the `Extending
6134 LLVM Guide <ExtendingLLVM.html>`_.
6138 Variable Argument Handling Intrinsics
6139 -------------------------------------
6141 Variable argument support is defined in LLVM with the
6142 :ref:`va_arg <i_va_arg>` instruction and these three intrinsic
6143 functions. These functions are related to the similarly named macros
6144 defined in the ``<stdarg.h>`` header file.
6146 All of these functions operate on arguments that use a target-specific
6147 value type "``va_list``". The LLVM assembly language reference manual
6148 does not define what this type is, so all transformations should be
6149 prepared to handle these functions regardless of the type used.
6151 This example shows how the :ref:`va_arg <i_va_arg>` instruction and the
6152 variable argument handling intrinsic functions are used.
6154 .. code-block:: llvm
6156 define i32 @test(i32 %X, ...) {
6157 ; Initialize variable argument processing
6159 %ap2 = bitcast i8** %ap to i8*
6160 call void @llvm.va_start(i8* %ap2)
6162 ; Read a single integer argument
6163 %tmp = va_arg i8** %ap, i32
6165 ; Demonstrate usage of llvm.va_copy and llvm.va_end
6167 %aq2 = bitcast i8** %aq to i8*
6168 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
6169 call void @llvm.va_end(i8* %aq2)
6171 ; Stop processing of arguments.
6172 call void @llvm.va_end(i8* %ap2)
6176 declare void @llvm.va_start(i8*)
6177 declare void @llvm.va_copy(i8*, i8*)
6178 declare void @llvm.va_end(i8*)
6182 '``llvm.va_start``' Intrinsic
6183 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6190 declare void %llvm.va_start(i8* <arglist>)
6195 The '``llvm.va_start``' intrinsic initializes ``*<arglist>`` for
6196 subsequent use by ``va_arg``.
6201 The argument is a pointer to a ``va_list`` element to initialize.
6206 The '``llvm.va_start``' intrinsic works just like the ``va_start`` macro
6207 available in C. In a target-dependent way, it initializes the
6208 ``va_list`` element to which the argument points, so that the next call
6209 to ``va_arg`` will produce the first variable argument passed to the
6210 function. Unlike the C ``va_start`` macro, this intrinsic does not need
6211 to know the last argument of the function as the compiler can figure
6214 '``llvm.va_end``' Intrinsic
6215 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6222 declare void @llvm.va_end(i8* <arglist>)
6227 The '``llvm.va_end``' intrinsic destroys ``*<arglist>``, which has been
6228 initialized previously with ``llvm.va_start`` or ``llvm.va_copy``.
6233 The argument is a pointer to a ``va_list`` to destroy.
6238 The '``llvm.va_end``' intrinsic works just like the ``va_end`` macro
6239 available in C. In a target-dependent way, it destroys the ``va_list``
6240 element to which the argument points. Calls to
6241 :ref:`llvm.va_start <int_va_start>` and
6242 :ref:`llvm.va_copy <int_va_copy>` must be matched exactly with calls to
6247 '``llvm.va_copy``' Intrinsic
6248 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6255 declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
6260 The '``llvm.va_copy``' intrinsic copies the current argument position
6261 from the source argument list to the destination argument list.
6266 The first argument is a pointer to a ``va_list`` element to initialize.
6267 The second argument is a pointer to a ``va_list`` element to copy from.
6272 The '``llvm.va_copy``' intrinsic works just like the ``va_copy`` macro
6273 available in C. In a target-dependent way, it copies the source
6274 ``va_list`` element into the destination ``va_list`` element. This
6275 intrinsic is necessary because the `` llvm.va_start`` intrinsic may be
6276 arbitrarily complex and require, for example, memory allocation.
6278 Accurate Garbage Collection Intrinsics
6279 --------------------------------------
6281 LLVM support for `Accurate Garbage Collection <GarbageCollection.html>`_
6282 (GC) requires the implementation and generation of these intrinsics.
6283 These intrinsics allow identification of :ref:`GC roots on the
6284 stack <int_gcroot>`, as well as garbage collector implementations that
6285 require :ref:`read <int_gcread>` and :ref:`write <int_gcwrite>` barriers.
6286 Front-ends for type-safe garbage collected languages should generate
6287 these intrinsics to make use of the LLVM garbage collectors. For more
6288 details, see `Accurate Garbage Collection with
6289 LLVM <GarbageCollection.html>`_.
6291 The garbage collection intrinsics only operate on objects in the generic
6292 address space (address space zero).
6296 '``llvm.gcroot``' Intrinsic
6297 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6304 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
6309 The '``llvm.gcroot``' intrinsic declares the existence of a GC root to
6310 the code generator, and allows some metadata to be associated with it.
6315 The first argument specifies the address of a stack object that contains
6316 the root pointer. The second pointer (which must be either a constant or
6317 a global value address) contains the meta-data to be associated with the
6323 At runtime, a call to this intrinsic stores a null pointer into the
6324 "ptrloc" location. At compile-time, the code generator generates
6325 information to allow the runtime to find the pointer at GC safe points.
6326 The '``llvm.gcroot``' intrinsic may only be used in a function which
6327 :ref:`specifies a GC algorithm <gc>`.
6331 '``llvm.gcread``' Intrinsic
6332 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6339 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
6344 The '``llvm.gcread``' intrinsic identifies reads of references from heap
6345 locations, allowing garbage collector implementations that require read
6351 The second argument is the address to read from, which should be an
6352 address allocated from the garbage collector. The first object is a
6353 pointer to the start of the referenced object, if needed by the language
6354 runtime (otherwise null).
6359 The '``llvm.gcread``' intrinsic has the same semantics as a load
6360 instruction, but may be replaced with substantially more complex code by
6361 the garbage collector runtime, as needed. The '``llvm.gcread``'
6362 intrinsic may only be used in a function which :ref:`specifies a GC
6367 '``llvm.gcwrite``' Intrinsic
6368 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6375 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
6380 The '``llvm.gcwrite``' intrinsic identifies writes of references to heap
6381 locations, allowing garbage collector implementations that require write
6382 barriers (such as generational or reference counting collectors).
6387 The first argument is the reference to store, the second is the start of
6388 the object to store it to, and the third is the address of the field of
6389 Obj to store to. If the runtime does not require a pointer to the
6390 object, Obj may be null.
6395 The '``llvm.gcwrite``' intrinsic has the same semantics as a store
6396 instruction, but may be replaced with substantially more complex code by
6397 the garbage collector runtime, as needed. The '``llvm.gcwrite``'
6398 intrinsic may only be used in a function which :ref:`specifies a GC
6401 Code Generator Intrinsics
6402 -------------------------
6404 These intrinsics are provided by LLVM to expose special features that
6405 may only be implemented with code generator support.
6407 '``llvm.returnaddress``' Intrinsic
6408 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6415 declare i8 *@llvm.returnaddress(i32 <level>)
6420 The '``llvm.returnaddress``' intrinsic attempts to compute a
6421 target-specific value indicating the return address of the current
6422 function or one of its callers.
6427 The argument to this intrinsic indicates which function to return the
6428 address for. Zero indicates the calling function, one indicates its
6429 caller, etc. The argument is **required** to be a constant integer
6435 The '``llvm.returnaddress``' intrinsic either returns a pointer
6436 indicating the return address of the specified call frame, or zero if it
6437 cannot be identified. The value returned by this intrinsic is likely to
6438 be incorrect or 0 for arguments other than zero, so it should only be
6439 used for debugging purposes.
6441 Note that calling this intrinsic does not prevent function inlining or
6442 other aggressive transformations, so the value returned may not be that
6443 of the obvious source-language caller.
6445 '``llvm.frameaddress``' Intrinsic
6446 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6453 declare i8* @llvm.frameaddress(i32 <level>)
6458 The '``llvm.frameaddress``' intrinsic attempts to return the
6459 target-specific frame pointer value for the specified stack frame.
6464 The argument to this intrinsic indicates which function to return the
6465 frame pointer for. Zero indicates the calling function, one indicates
6466 its caller, etc. The argument is **required** to be a constant integer
6472 The '``llvm.frameaddress``' intrinsic either returns a pointer
6473 indicating the frame address of the specified call frame, or zero if it
6474 cannot be identified. The value returned by this intrinsic is likely to
6475 be incorrect or 0 for arguments other than zero, so it should only be
6476 used for debugging purposes.
6478 Note that calling this intrinsic does not prevent function inlining or
6479 other aggressive transformations, so the value returned may not be that
6480 of the obvious source-language caller.
6484 '``llvm.stacksave``' Intrinsic
6485 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6492 declare i8* @llvm.stacksave()
6497 The '``llvm.stacksave``' intrinsic is used to remember the current state
6498 of the function stack, for use with
6499 :ref:`llvm.stackrestore <int_stackrestore>`. This is useful for
6500 implementing language features like scoped automatic variable sized
6506 This intrinsic returns a opaque pointer value that can be passed to
6507 :ref:`llvm.stackrestore <int_stackrestore>`. When an
6508 ``llvm.stackrestore`` intrinsic is executed with a value saved from
6509 ``llvm.stacksave``, it effectively restores the state of the stack to
6510 the state it was in when the ``llvm.stacksave`` intrinsic executed. In
6511 practice, this pops any :ref:`alloca <i_alloca>` blocks from the stack that
6512 were allocated after the ``llvm.stacksave`` was executed.
6514 .. _int_stackrestore:
6516 '``llvm.stackrestore``' Intrinsic
6517 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6524 declare void @llvm.stackrestore(i8* %ptr)
6529 The '``llvm.stackrestore``' intrinsic is used to restore the state of
6530 the function stack to the state it was in when the corresponding
6531 :ref:`llvm.stacksave <int_stacksave>` intrinsic executed. This is
6532 useful for implementing language features like scoped automatic variable
6533 sized arrays in C99.
6538 See the description for :ref:`llvm.stacksave <int_stacksave>`.
6540 '``llvm.prefetch``' Intrinsic
6541 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6548 declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>, i32 <cache type>)
6553 The '``llvm.prefetch``' intrinsic is a hint to the code generator to
6554 insert a prefetch instruction if supported; otherwise, it is a noop.
6555 Prefetches have no effect on the behavior of the program but can change
6556 its performance characteristics.
6561 ``address`` is the address to be prefetched, ``rw`` is the specifier
6562 determining if the fetch should be for a read (0) or write (1), and
6563 ``locality`` is a temporal locality specifier ranging from (0) - no
6564 locality, to (3) - extremely local keep in cache. The ``cache type``
6565 specifies whether the prefetch is performed on the data (1) or
6566 instruction (0) cache. The ``rw``, ``locality`` and ``cache type``
6567 arguments must be constant integers.
6572 This intrinsic does not modify the behavior of the program. In
6573 particular, prefetches cannot trap and do not produce a value. On
6574 targets that support this intrinsic, the prefetch can provide hints to
6575 the processor cache for better performance.
6577 '``llvm.pcmarker``' Intrinsic
6578 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6585 declare void @llvm.pcmarker(i32 <id>)
6590 The '``llvm.pcmarker``' intrinsic is a method to export a Program
6591 Counter (PC) in a region of code to simulators and other tools. The
6592 method is target specific, but it is expected that the marker will use
6593 exported symbols to transmit the PC of the marker. The marker makes no
6594 guarantees that it will remain with any specific instruction after
6595 optimizations. It is possible that the presence of a marker will inhibit
6596 optimizations. The intended use is to be inserted after optimizations to
6597 allow correlations of simulation runs.
6602 ``id`` is a numerical id identifying the marker.
6607 This intrinsic does not modify the behavior of the program. Backends
6608 that do not support this intrinsic may ignore it.
6610 '``llvm.readcyclecounter``' Intrinsic
6611 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6618 declare i64 @llvm.readcyclecounter()
6623 The '``llvm.readcyclecounter``' intrinsic provides access to the cycle
6624 counter register (or similar low latency, high accuracy clocks) on those
6625 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6626 should map to RPCC. As the backing counters overflow quickly (on the
6627 order of 9 seconds on alpha), this should only be used for small
6633 When directly supported, reading the cycle counter should not modify any
6634 memory. Implementations are allowed to either return a application
6635 specific value or a system wide value. On backends without support, this
6636 is lowered to a constant 0.
6638 Standard C Library Intrinsics
6639 -----------------------------
6641 LLVM provides intrinsics for a few important standard C library
6642 functions. These intrinsics allow source-language front-ends to pass
6643 information about the alignment of the pointer arguments to the code
6644 generator, providing opportunity for more efficient code generation.
6648 '``llvm.memcpy``' Intrinsic
6649 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6654 This is an overloaded intrinsic. You can use ``llvm.memcpy`` on any
6655 integer bit width and for different address spaces. Not all targets
6656 support all bit widths however.
6660 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
6661 i32 <len>, i32 <align>, i1 <isvolatile>)
6662 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
6663 i64 <len>, i32 <align>, i1 <isvolatile>)
6668 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
6669 source location to the destination location.
6671 Note that, unlike the standard libc function, the ``llvm.memcpy.*``
6672 intrinsics do not return a value, takes extra alignment/isvolatile
6673 arguments and the pointers can be in specified address spaces.
6678 The first argument is a pointer to the destination, the second is a
6679 pointer to the source. The third argument is an integer argument
6680 specifying the number of bytes to copy, the fourth argument is the
6681 alignment of the source and destination locations, and the fifth is a
6682 boolean indicating a volatile access.
6684 If the call to this intrinsic has an alignment value that is not 0 or 1,
6685 then the caller guarantees that both the source and destination pointers
6686 are aligned to that boundary.
6688 If the ``isvolatile`` parameter is ``true``, the ``llvm.memcpy`` call is
6689 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
6690 very cleanly specified and it is unwise to depend on it.
6695 The '``llvm.memcpy.*``' intrinsics copy a block of memory from the
6696 source location to the destination location, which are not allowed to
6697 overlap. It copies "len" bytes of memory over. If the argument is known
6698 to be aligned to some boundary, this can be specified as the fourth
6699 argument, otherwise it should be set to 0 or 1.
6701 '``llvm.memmove``' Intrinsic
6702 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6707 This is an overloaded intrinsic. You can use llvm.memmove on any integer
6708 bit width and for different address space. Not all targets support all
6713 declare void @llvm.memmove.p0i8.p0i8.i32(i8* <dest>, i8* <src>,
6714 i32 <len>, i32 <align>, i1 <isvolatile>)
6715 declare void @llvm.memmove.p0i8.p0i8.i64(i8* <dest>, i8* <src>,
6716 i64 <len>, i32 <align>, i1 <isvolatile>)
6721 The '``llvm.memmove.*``' intrinsics move a block of memory from the
6722 source location to the destination location. It is similar to the
6723 '``llvm.memcpy``' intrinsic but allows the two memory locations to
6726 Note that, unlike the standard libc function, the ``llvm.memmove.*``
6727 intrinsics do not return a value, takes extra alignment/isvolatile
6728 arguments and the pointers can be in specified address spaces.
6733 The first argument is a pointer to the destination, the second is a
6734 pointer to the source. The third argument is an integer argument
6735 specifying the number of bytes to copy, the fourth argument is the
6736 alignment of the source and destination locations, and the fifth is a
6737 boolean indicating a volatile access.
6739 If the call to this intrinsic has an alignment value that is not 0 or 1,
6740 then the caller guarantees that the source and destination pointers are
6741 aligned to that boundary.
6743 If the ``isvolatile`` parameter is ``true``, the ``llvm.memmove`` call
6744 is a :ref:`volatile operation <volatile>`. The detailed access behavior is
6745 not very cleanly specified and it is unwise to depend on it.
6750 The '``llvm.memmove.*``' intrinsics copy a block of memory from the
6751 source location to the destination location, which may overlap. It
6752 copies "len" bytes of memory over. If the argument is known to be
6753 aligned to some boundary, this can be specified as the fourth argument,
6754 otherwise it should be set to 0 or 1.
6756 '``llvm.memset.*``' Intrinsics
6757 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
6762 This is an overloaded intrinsic. You can use llvm.memset on any integer
6763 bit width and for different address spaces. However, not all targets
6764 support all bit widths.
6768 declare void @llvm.memset.p0i8.i32(i8* <dest>, i8 <val>,
6769 i32 <len>, i32 <align>, i1 <isvolatile>)
6770 declare void @llvm.memset.p0i8.i64(i8* <dest>, i8 <val>,
6771 i64 <len>, i32 <align>, i1 <isvolatile>)
6776 The '``llvm.memset.*``' intrinsics fill a block of memory with a
6777 particular byte value.
6779 Note that, unlike the standard libc function, the ``llvm.memset``
6780 intrinsic does not return a value and takes extra alignment/volatile
6781 arguments. Also, the destination can be in an arbitrary address space.
6786 The first argument is a pointer to the destination to fill, the second
6787 is the byte value with which to fill it, the third argument is an
6788 integer argument specifying the number of bytes to fill, and the fourth
6789 argument is the known alignment of the destination location.
6791 If the call to this intrinsic has an alignment value that is not 0 or 1,
6792 then the caller guarantees that the destination pointer is aligned to
6795 If the ``isvolatile`` parameter is ``true``, the ``llvm.memset`` call is
6796 a :ref:`volatile operation <volatile>`. The detailed access behavior is not
6797 very cleanly specified and it is unwise to depend on it.
6802 The '``llvm.memset.*``' intrinsics fill "len" bytes of memory starting
6803 at the destination location. If the argument is known to be aligned to
6804 some boundary, this can be specified as the fourth argument, otherwise
6805 it should be set to 0 or 1.
6807 '``llvm.sqrt.*``' Intrinsic
6808 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6813 This is an overloaded intrinsic. You can use ``llvm.sqrt`` on any
6814 floating point or vector of floating point type. Not all targets support
6819 declare float @llvm.sqrt.f32(float %Val)
6820 declare double @llvm.sqrt.f64(double %Val)
6821 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6822 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6823 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
6828 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
6829 returning the same value as the libm '``sqrt``' functions would. Unlike
6830 ``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
6831 negative numbers other than -0.0 (which allows for better optimization,
6832 because there is no need to worry about errno being set).
6833 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.
6838 The argument and return value are floating point numbers of the same
6844 This function returns the sqrt of the specified operand if it is a
6845 nonnegative floating point number.
6847 '``llvm.powi.*``' Intrinsic
6848 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
6853 This is an overloaded intrinsic. You can use ``llvm.powi`` on any
6854 floating point or vector of floating point type. Not all targets support
6859 declare float @llvm.powi.f32(float %Val, i32 %power)
6860 declare double @llvm.powi.f64(double %Val, i32 %power)
6861 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6862 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6863 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
6868 The '``llvm.powi.*``' intrinsics return the first operand raised to the
6869 specified (positive or negative) power. The order of evaluation of
6870 multiplications is not defined. When a vector of floating point type is
6871 used, the second argument remains a scalar integer value.
6876 The second argument is an integer power, and the first is a value to
6877 raise to that power.
6882 This function returns the first value raised to the second power with an
6883 unspecified sequence of rounding operations.
6885 '``llvm.sin.*``' Intrinsic
6886 ^^^^^^^^^^^^^^^^^^^^^^^^^^
6891 This is an overloaded intrinsic. You can use ``llvm.sin`` on any
6892 floating point or vector of floating point type. Not all targets support
6897 declare float @llvm.sin.f32(float %Val)
6898 declare double @llvm.sin.f64(double %Val)
6899 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6900 declare fp128 @llvm.sin.f128(fp128 %Val)
6901 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6906 The '``llvm.sin.*``' intrinsics return the sine of the operand.
6911 The argument and return value are floating point numbers of the same
6917 This function returns the sine of the specified operand, returning the
6918 same values as the libm ``sin`` functions would, and handles error
6919 conditions in the same way.
6921 '``llvm.cos.*``' Intrinsic
6922 ^^^^^^^^^^^^^^^^^^^^^^^^^^
6927 This is an overloaded intrinsic. You can use ``llvm.cos`` on any
6928 floating point or vector of floating point type. Not all targets support
6933 declare float @llvm.cos.f32(float %Val)
6934 declare double @llvm.cos.f64(double %Val)
6935 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6936 declare fp128 @llvm.cos.f128(fp128 %Val)
6937 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6942 The '``llvm.cos.*``' intrinsics return the cosine of the operand.
6947 The argument and return value are floating point numbers of the same
6953 This function returns the cosine of the specified operand, returning the
6954 same values as the libm ``cos`` functions would, and handles error
6955 conditions in the same way.
6957 '``llvm.pow.*``' Intrinsic
6958 ^^^^^^^^^^^^^^^^^^^^^^^^^^
6963 This is an overloaded intrinsic. You can use ``llvm.pow`` on any
6964 floating point or vector of floating point type. Not all targets support
6969 declare float @llvm.pow.f32(float %Val, float %Power)
6970 declare double @llvm.pow.f64(double %Val, double %Power)
6971 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6972 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6973 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6978 The '``llvm.pow.*``' intrinsics return the first operand raised to the
6979 specified (positive or negative) power.
6984 The second argument is a floating point power, and the first is a value
6985 to raise to that power.
6990 This function returns the first value raised to the second power,
6991 returning the same values as the libm ``pow`` functions would, and
6992 handles error conditions in the same way.
6994 '``llvm.exp.*``' Intrinsic
6995 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7000 This is an overloaded intrinsic. You can use ``llvm.exp`` on any
7001 floating point or vector of floating point type. Not all targets support
7006 declare float @llvm.exp.f32(float %Val)
7007 declare double @llvm.exp.f64(double %Val)
7008 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7009 declare fp128 @llvm.exp.f128(fp128 %Val)
7010 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7015 The '``llvm.exp.*``' intrinsics perform the exp function.
7020 The argument and return value are floating point numbers of the same
7026 This function returns the same values as the libm ``exp`` functions
7027 would, and handles error conditions in the same way.
7029 '``llvm.exp2.*``' Intrinsic
7030 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7035 This is an overloaded intrinsic. You can use ``llvm.exp2`` on any
7036 floating point or vector of floating point type. Not all targets support
7041 declare float @llvm.exp2.f32(float %Val)
7042 declare double @llvm.exp2.f64(double %Val)
7043 declare x86_fp80 @llvm.exp2.f80(x86_fp80 %Val)
7044 declare fp128 @llvm.exp2.f128(fp128 %Val)
7045 declare ppc_fp128 @llvm.exp2.ppcf128(ppc_fp128 %Val)
7050 The '``llvm.exp2.*``' intrinsics perform the exp2 function.
7055 The argument and return value are floating point numbers of the same
7061 This function returns the same values as the libm ``exp2`` functions
7062 would, and handles error conditions in the same way.
7064 '``llvm.log.*``' Intrinsic
7065 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7070 This is an overloaded intrinsic. You can use ``llvm.log`` on any
7071 floating point or vector of floating point type. Not all targets support
7076 declare float @llvm.log.f32(float %Val)
7077 declare double @llvm.log.f64(double %Val)
7078 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7079 declare fp128 @llvm.log.f128(fp128 %Val)
7080 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7085 The '``llvm.log.*``' intrinsics perform the log function.
7090 The argument and return value are floating point numbers of the same
7096 This function returns the same values as the libm ``log`` functions
7097 would, and handles error conditions in the same way.
7099 '``llvm.log10.*``' Intrinsic
7100 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7105 This is an overloaded intrinsic. You can use ``llvm.log10`` on any
7106 floating point or vector of floating point type. Not all targets support
7111 declare float @llvm.log10.f32(float %Val)
7112 declare double @llvm.log10.f64(double %Val)
7113 declare x86_fp80 @llvm.log10.f80(x86_fp80 %Val)
7114 declare fp128 @llvm.log10.f128(fp128 %Val)
7115 declare ppc_fp128 @llvm.log10.ppcf128(ppc_fp128 %Val)
7120 The '``llvm.log10.*``' intrinsics perform the log10 function.
7125 The argument and return value are floating point numbers of the same
7131 This function returns the same values as the libm ``log10`` functions
7132 would, and handles error conditions in the same way.
7134 '``llvm.log2.*``' Intrinsic
7135 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7140 This is an overloaded intrinsic. You can use ``llvm.log2`` on any
7141 floating point or vector of floating point type. Not all targets support
7146 declare float @llvm.log2.f32(float %Val)
7147 declare double @llvm.log2.f64(double %Val)
7148 declare x86_fp80 @llvm.log2.f80(x86_fp80 %Val)
7149 declare fp128 @llvm.log2.f128(fp128 %Val)
7150 declare ppc_fp128 @llvm.log2.ppcf128(ppc_fp128 %Val)
7155 The '``llvm.log2.*``' intrinsics perform the log2 function.
7160 The argument and return value are floating point numbers of the same
7166 This function returns the same values as the libm ``log2`` functions
7167 would, and handles error conditions in the same way.
7169 '``llvm.fma.*``' Intrinsic
7170 ^^^^^^^^^^^^^^^^^^^^^^^^^^
7175 This is an overloaded intrinsic. You can use ``llvm.fma`` on any
7176 floating point or vector of floating point type. Not all targets support
7181 declare float @llvm.fma.f32(float %a, float %b, float %c)
7182 declare double @llvm.fma.f64(double %a, double %b, double %c)
7183 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7184 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7185 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7190 The '``llvm.fma.*``' intrinsics perform the fused multiply-add
7196 The argument and return value are floating point numbers of the same
7202 This function returns the same values as the libm ``fma`` functions
7205 '``llvm.fabs.*``' Intrinsic
7206 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7211 This is an overloaded intrinsic. You can use ``llvm.fabs`` on any
7212 floating point or vector of floating point type. Not all targets support
7217 declare float @llvm.fabs.f32(float %Val)
7218 declare double @llvm.fabs.f64(double %Val)
7219 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
7220 declare fp128 @llvm.fabs.f128(fp128 %Val)
7221 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
7226 The '``llvm.fabs.*``' intrinsics return the absolute value of the
7232 The argument and return value are floating point numbers of the same
7238 This function returns the same values as the libm ``fabs`` functions
7239 would, and handles error conditions in the same way.
7241 '``llvm.floor.*``' Intrinsic
7242 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7247 This is an overloaded intrinsic. You can use ``llvm.floor`` on any
7248 floating point or vector of floating point type. Not all targets support
7253 declare float @llvm.floor.f32(float %Val)
7254 declare double @llvm.floor.f64(double %Val)
7255 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
7256 declare fp128 @llvm.floor.f128(fp128 %Val)
7257 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
7262 The '``llvm.floor.*``' intrinsics return the floor of the operand.
7267 The argument and return value are floating point numbers of the same
7273 This function returns the same values as the libm ``floor`` functions
7274 would, and handles error conditions in the same way.
7276 '``llvm.ceil.*``' Intrinsic
7277 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7282 This is an overloaded intrinsic. You can use ``llvm.ceil`` on any
7283 floating point or vector of floating point type. Not all targets support
7288 declare float @llvm.ceil.f32(float %Val)
7289 declare double @llvm.ceil.f64(double %Val)
7290 declare x86_fp80 @llvm.ceil.f80(x86_fp80 %Val)
7291 declare fp128 @llvm.ceil.f128(fp128 %Val)
7292 declare ppc_fp128 @llvm.ceil.ppcf128(ppc_fp128 %Val)
7297 The '``llvm.ceil.*``' intrinsics return the ceiling of the operand.
7302 The argument and return value are floating point numbers of the same
7308 This function returns the same values as the libm ``ceil`` functions
7309 would, and handles error conditions in the same way.
7311 '``llvm.trunc.*``' Intrinsic
7312 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7317 This is an overloaded intrinsic. You can use ``llvm.trunc`` on any
7318 floating point or vector of floating point type. Not all targets support
7323 declare float @llvm.trunc.f32(float %Val)
7324 declare double @llvm.trunc.f64(double %Val)
7325 declare x86_fp80 @llvm.trunc.f80(x86_fp80 %Val)
7326 declare fp128 @llvm.trunc.f128(fp128 %Val)
7327 declare ppc_fp128 @llvm.trunc.ppcf128(ppc_fp128 %Val)
7332 The '``llvm.trunc.*``' intrinsics returns the operand rounded to the
7333 nearest integer not larger in magnitude than 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 ``trunc`` functions
7345 would, and handles error conditions in the same way.
7347 '``llvm.rint.*``' Intrinsic
7348 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7353 This is an overloaded intrinsic. You can use ``llvm.rint`` on any
7354 floating point or vector of floating point type. Not all targets support
7359 declare float @llvm.rint.f32(float %Val)
7360 declare double @llvm.rint.f64(double %Val)
7361 declare x86_fp80 @llvm.rint.f80(x86_fp80 %Val)
7362 declare fp128 @llvm.rint.f128(fp128 %Val)
7363 declare ppc_fp128 @llvm.rint.ppcf128(ppc_fp128 %Val)
7368 The '``llvm.rint.*``' intrinsics returns the operand rounded to the
7369 nearest integer. It may raise an inexact floating-point exception if the
7370 operand isn't an integer.
7375 The argument and return value are floating point numbers of the same
7381 This function returns the same values as the libm ``rint`` functions
7382 would, and handles error conditions in the same way.
7384 '``llvm.nearbyint.*``' Intrinsic
7385 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7390 This is an overloaded intrinsic. You can use ``llvm.nearbyint`` on any
7391 floating point or vector of floating point type. Not all targets support
7396 declare float @llvm.nearbyint.f32(float %Val)
7397 declare double @llvm.nearbyint.f64(double %Val)
7398 declare x86_fp80 @llvm.nearbyint.f80(x86_fp80 %Val)
7399 declare fp128 @llvm.nearbyint.f128(fp128 %Val)
7400 declare ppc_fp128 @llvm.nearbyint.ppcf128(ppc_fp128 %Val)
7405 The '``llvm.nearbyint.*``' intrinsics returns the operand rounded to the
7411 The argument and return value are floating point numbers of the same
7417 This function returns the same values as the libm ``nearbyint``
7418 functions would, and handles error conditions in the same way.
7420 Bit Manipulation Intrinsics
7421 ---------------------------
7423 LLVM provides intrinsics for a few important bit manipulation
7424 operations. These allow efficient code generation for some algorithms.
7426 '``llvm.bswap.*``' Intrinsics
7427 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7432 This is an overloaded intrinsic function. You can use bswap on any
7433 integer type that is an even number of bytes (i.e. BitWidth % 16 == 0).
7437 declare i16 @llvm.bswap.i16(i16 <id>)
7438 declare i32 @llvm.bswap.i32(i32 <id>)
7439 declare i64 @llvm.bswap.i64(i64 <id>)
7444 The '``llvm.bswap``' family of intrinsics is used to byte swap integer
7445 values with an even number of bytes (positive multiple of 16 bits).
7446 These are useful for performing operations on data that is not in the
7447 target's native byte order.
7452 The ``llvm.bswap.i16`` intrinsic returns an i16 value that has the high
7453 and low byte of the input i16 swapped. Similarly, the ``llvm.bswap.i32``
7454 intrinsic returns an i32 value that has the four bytes of the input i32
7455 swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the
7456 returned i32 will have its bytes in 3, 2, 1, 0 order. The
7457 ``llvm.bswap.i48``, ``llvm.bswap.i64`` and other intrinsics extend this
7458 concept to additional even-byte lengths (6 bytes, 8 bytes and more,
7461 '``llvm.ctpop.*``' Intrinsic
7462 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7467 This is an overloaded intrinsic. You can use llvm.ctpop on any integer
7468 bit width, or on any vector with integer elements. Not all targets
7469 support all bit widths or vector types, however.
7473 declare i8 @llvm.ctpop.i8(i8 <src>)
7474 declare i16 @llvm.ctpop.i16(i16 <src>)
7475 declare i32 @llvm.ctpop.i32(i32 <src>)
7476 declare i64 @llvm.ctpop.i64(i64 <src>)
7477 declare i256 @llvm.ctpop.i256(i256 <src>)
7478 declare <2 x i32> @llvm.ctpop.v2i32(<2 x i32> <src>)
7483 The '``llvm.ctpop``' family of intrinsics counts the number of bits set
7489 The only argument is the value to be counted. The argument may be of any
7490 integer type, or a vector with integer elements. The return type must
7491 match the argument type.
7496 The '``llvm.ctpop``' intrinsic counts the 1's in a variable, or within
7497 each element of a vector.
7499 '``llvm.ctlz.*``' Intrinsic
7500 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7505 This is an overloaded intrinsic. You can use ``llvm.ctlz`` on any
7506 integer bit width, or any vector whose elements are integers. Not all
7507 targets support all bit widths or vector types, however.
7511 declare i8 @llvm.ctlz.i8 (i8 <src>, i1 <is_zero_undef>)
7512 declare i16 @llvm.ctlz.i16 (i16 <src>, i1 <is_zero_undef>)
7513 declare i32 @llvm.ctlz.i32 (i32 <src>, i1 <is_zero_undef>)
7514 declare i64 @llvm.ctlz.i64 (i64 <src>, i1 <is_zero_undef>)
7515 declare i256 @llvm.ctlz.i256(i256 <src>, i1 <is_zero_undef>)
7516 declase <2 x i32> @llvm.ctlz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
7521 The '``llvm.ctlz``' family of intrinsic functions counts the number of
7522 leading zeros in a variable.
7527 The first argument is the value to be counted. This argument may be of
7528 any integer type, or a vectory with integer element type. The return
7529 type must match the first argument type.
7531 The second argument must be a constant and is a flag to indicate whether
7532 the intrinsic should ensure that a zero as the first argument produces a
7533 defined result. Historically some architectures did not provide a
7534 defined result for zero values as efficiently, and many algorithms are
7535 now predicated on avoiding zero-value inputs.
7540 The '``llvm.ctlz``' intrinsic counts the leading (most significant)
7541 zeros in a variable, or within each element of the vector. If
7542 ``src == 0`` then the result is the size in bits of the type of ``src``
7543 if ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
7544 ``llvm.ctlz(i32 2) = 30``.
7546 '``llvm.cttz.*``' Intrinsic
7547 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
7552 This is an overloaded intrinsic. You can use ``llvm.cttz`` on any
7553 integer bit width, or any vector of integer elements. Not all targets
7554 support all bit widths or vector types, however.
7558 declare i8 @llvm.cttz.i8 (i8 <src>, i1 <is_zero_undef>)
7559 declare i16 @llvm.cttz.i16 (i16 <src>, i1 <is_zero_undef>)
7560 declare i32 @llvm.cttz.i32 (i32 <src>, i1 <is_zero_undef>)
7561 declare i64 @llvm.cttz.i64 (i64 <src>, i1 <is_zero_undef>)
7562 declare i256 @llvm.cttz.i256(i256 <src>, i1 <is_zero_undef>)
7563 declase <2 x i32> @llvm.cttz.v2i32(<2 x i32> <src>, i1 <is_zero_undef>)
7568 The '``llvm.cttz``' family of intrinsic functions counts the number of
7574 The first argument is the value to be counted. This argument may be of
7575 any integer type, or a vectory with integer element type. The return
7576 type must match the first argument type.
7578 The second argument must be a constant and is a flag to indicate whether
7579 the intrinsic should ensure that a zero as the first argument produces a
7580 defined result. Historically some architectures did not provide a
7581 defined result for zero values as efficiently, and many algorithms are
7582 now predicated on avoiding zero-value inputs.
7587 The '``llvm.cttz``' intrinsic counts the trailing (least significant)
7588 zeros in a variable, or within each element of a vector. If ``src == 0``
7589 then the result is the size in bits of the type of ``src`` if
7590 ``is_zero_undef == 0`` and ``undef`` otherwise. For example,
7591 ``llvm.cttz(2) = 1``.
7593 Arithmetic with Overflow Intrinsics
7594 -----------------------------------
7596 LLVM provides intrinsics for some arithmetic with overflow operations.
7598 '``llvm.sadd.with.overflow.*``' Intrinsics
7599 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7604 This is an overloaded intrinsic. You can use ``llvm.sadd.with.overflow``
7605 on any integer bit width.
7609 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7610 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7611 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7616 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
7617 a signed addition of the two arguments, and indicate whether an overflow
7618 occurred during the signed summation.
7623 The arguments (%a and %b) and the first element of the result structure
7624 may be of integer types of any bit width, but they must have the same
7625 bit width. The second element of the result structure must be of type
7626 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
7632 The '``llvm.sadd.with.overflow``' family of intrinsic functions perform
7633 a signed addition of the two variables. They return a structure --- the
7634 first element of which is the signed summation, and the second element
7635 of which is a bit specifying if the signed summation resulted in an
7641 .. code-block:: llvm
7643 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7644 %sum = extractvalue {i32, i1} %res, 0
7645 %obit = extractvalue {i32, i1} %res, 1
7646 br i1 %obit, label %overflow, label %normal
7648 '``llvm.uadd.with.overflow.*``' Intrinsics
7649 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7654 This is an overloaded intrinsic. You can use ``llvm.uadd.with.overflow``
7655 on any integer bit width.
7659 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7660 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7661 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7666 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
7667 an unsigned addition of the two arguments, and indicate whether a carry
7668 occurred during the unsigned summation.
7673 The arguments (%a and %b) and the first element of the result structure
7674 may be of integer types of any bit width, but they must have the same
7675 bit width. The second element of the result structure must be of type
7676 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
7682 The '``llvm.uadd.with.overflow``' family of intrinsic functions perform
7683 an unsigned addition of the two arguments. They return a structure --- the
7684 first element of which is the sum, and the second element of which is a
7685 bit specifying if the unsigned summation resulted in a carry.
7690 .. code-block:: llvm
7692 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7693 %sum = extractvalue {i32, i1} %res, 0
7694 %obit = extractvalue {i32, i1} %res, 1
7695 br i1 %obit, label %carry, label %normal
7697 '``llvm.ssub.with.overflow.*``' Intrinsics
7698 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7703 This is an overloaded intrinsic. You can use ``llvm.ssub.with.overflow``
7704 on any integer bit width.
7708 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7709 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7710 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7715 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
7716 a signed subtraction of the two arguments, and indicate whether an
7717 overflow occurred during the signed subtraction.
7722 The arguments (%a and %b) and the first element of the result structure
7723 may be of integer types of any bit width, but they must have the same
7724 bit width. The second element of the result structure must be of type
7725 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
7731 The '``llvm.ssub.with.overflow``' family of intrinsic functions perform
7732 a signed subtraction of the two arguments. They return a structure --- the
7733 first element of which is the subtraction, and the second element of
7734 which is a bit specifying if the signed subtraction resulted in an
7740 .. code-block:: llvm
7742 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7743 %sum = extractvalue {i32, i1} %res, 0
7744 %obit = extractvalue {i32, i1} %res, 1
7745 br i1 %obit, label %overflow, label %normal
7747 '``llvm.usub.with.overflow.*``' Intrinsics
7748 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7753 This is an overloaded intrinsic. You can use ``llvm.usub.with.overflow``
7754 on any integer bit width.
7758 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7759 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7760 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7765 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
7766 an unsigned subtraction of the two arguments, and indicate whether an
7767 overflow occurred during the unsigned subtraction.
7772 The arguments (%a and %b) and the first element of the result structure
7773 may be of integer types of any bit width, but they must have the same
7774 bit width. The second element of the result structure must be of type
7775 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
7781 The '``llvm.usub.with.overflow``' family of intrinsic functions perform
7782 an unsigned subtraction of the two arguments. They return a structure ---
7783 the first element of which is the subtraction, and the second element of
7784 which is a bit specifying if the unsigned subtraction resulted in an
7790 .. code-block:: llvm
7792 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7793 %sum = extractvalue {i32, i1} %res, 0
7794 %obit = extractvalue {i32, i1} %res, 1
7795 br i1 %obit, label %overflow, label %normal
7797 '``llvm.smul.with.overflow.*``' Intrinsics
7798 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7803 This is an overloaded intrinsic. You can use ``llvm.smul.with.overflow``
7804 on any integer bit width.
7808 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7809 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7810 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7815 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
7816 a signed multiplication of the two arguments, and indicate whether an
7817 overflow occurred during the signed multiplication.
7822 The arguments (%a and %b) and the first element of the result structure
7823 may be of integer types of any bit width, but they must have the same
7824 bit width. The second element of the result structure must be of type
7825 ``i1``. ``%a`` and ``%b`` are the two values that will undergo signed
7831 The '``llvm.smul.with.overflow``' family of intrinsic functions perform
7832 a signed multiplication of the two arguments. They return a structure ---
7833 the first element of which is the multiplication, and the second element
7834 of which is a bit specifying if the signed multiplication resulted in an
7840 .. code-block:: llvm
7842 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7843 %sum = extractvalue {i32, i1} %res, 0
7844 %obit = extractvalue {i32, i1} %res, 1
7845 br i1 %obit, label %overflow, label %normal
7847 '``llvm.umul.with.overflow.*``' Intrinsics
7848 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7853 This is an overloaded intrinsic. You can use ``llvm.umul.with.overflow``
7854 on any integer bit width.
7858 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7859 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7860 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7865 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
7866 a unsigned multiplication of the two arguments, and indicate whether an
7867 overflow occurred during the unsigned multiplication.
7872 The arguments (%a and %b) and the first element of the result structure
7873 may be of integer types of any bit width, but they must have the same
7874 bit width. The second element of the result structure must be of type
7875 ``i1``. ``%a`` and ``%b`` are the two values that will undergo unsigned
7881 The '``llvm.umul.with.overflow``' family of intrinsic functions perform
7882 an unsigned multiplication of the two arguments. They return a structure ---
7883 the first element of which is the multiplication, and the second
7884 element of which is a bit specifying if the unsigned multiplication
7885 resulted in an overflow.
7890 .. code-block:: llvm
7892 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7893 %sum = extractvalue {i32, i1} %res, 0
7894 %obit = extractvalue {i32, i1} %res, 1
7895 br i1 %obit, label %overflow, label %normal
7897 Specialised Arithmetic Intrinsics
7898 ---------------------------------
7900 '``llvm.fmuladd.*``' Intrinsic
7901 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7908 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
7909 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
7914 The '``llvm.fmuladd.*``' intrinsic functions represent multiply-add
7915 expressions that can be fused if the code generator determines that (a) the
7916 target instruction set has support for a fused operation, and (b) that the
7917 fused operation is more efficient than the equivalent, separate pair of mul
7918 and add instructions.
7923 The '``llvm.fmuladd.*``' intrinsics each take three arguments: two
7924 multiplicands, a and b, and an addend c.
7933 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
7935 is equivalent to the expression a \* b + c, except that rounding will
7936 not be performed between the multiplication and addition steps if the
7937 code generator fuses the operations. Fusion is not guaranteed, even if
7938 the target platform supports it. If a fused multiply-add is required the
7939 corresponding llvm.fma.\* intrinsic function should be used instead.
7944 .. code-block:: llvm
7946 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields {float}:r2 = (a * b) + c
7948 Half Precision Floating Point Intrinsics
7949 ----------------------------------------
7951 For most target platforms, half precision floating point is a
7952 storage-only format. This means that it is a dense encoding (in memory)
7953 but does not support computation in the format.
7955 This means that code must first load the half-precision floating point
7956 value as an i16, then convert it to float with
7957 :ref:`llvm.convert.from.fp16 <int_convert_from_fp16>`. Computation can
7958 then be performed on the float value (including extending to double
7959 etc). To store the value back to memory, it is first converted to float
7960 if needed, then converted to i16 with
7961 :ref:`llvm.convert.to.fp16 <int_convert_to_fp16>`, then storing as an
7964 .. _int_convert_to_fp16:
7966 '``llvm.convert.to.fp16``' Intrinsic
7967 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
7974 declare i16 @llvm.convert.to.fp16(f32 %a)
7979 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion
7980 from single precision floating point format to half precision floating
7986 The intrinsic function contains single argument - the value to be
7992 The '``llvm.convert.to.fp16``' intrinsic function performs a conversion
7993 from single precision floating point format to half precision floating
7994 point format. The return value is an ``i16`` which contains the
8000 .. code-block:: llvm
8002 %res = call i16 @llvm.convert.to.fp16(f32 %a)
8003 store i16 %res, i16* @x, align 2
8005 .. _int_convert_from_fp16:
8007 '``llvm.convert.from.fp16``' Intrinsic
8008 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8015 declare f32 @llvm.convert.from.fp16(i16 %a)
8020 The '``llvm.convert.from.fp16``' intrinsic function performs a
8021 conversion from half precision floating point format to single precision
8022 floating point format.
8027 The intrinsic function contains single argument - the value to be
8033 The '``llvm.convert.from.fp16``' intrinsic function performs a
8034 conversion from half single precision floating point format to single
8035 precision floating point format. The input half-float value is
8036 represented by an ``i16`` value.
8041 .. code-block:: llvm
8043 %a = load i16* @x, align 2
8044 %res = call f32 @llvm.convert.from.fp16(i16 %a)
8049 The LLVM debugger intrinsics (which all start with ``llvm.dbg.``
8050 prefix), are described in the `LLVM Source Level
8051 Debugging <SourceLevelDebugging.html#format_common_intrinsics>`_
8054 Exception Handling Intrinsics
8055 -----------------------------
8057 The LLVM exception handling intrinsics (which all start with
8058 ``llvm.eh.`` prefix), are described in the `LLVM Exception
8059 Handling <ExceptionHandling.html#format_common_intrinsics>`_ document.
8063 Trampoline Intrinsics
8064 ---------------------
8066 These intrinsics make it possible to excise one parameter, marked with
8067 the :ref:`nest <nest>` attribute, from a function. The result is a
8068 callable function pointer lacking the nest parameter - the caller does
8069 not need to provide a value for it. Instead, the value to use is stored
8070 in advance in a "trampoline", a block of memory usually allocated on the
8071 stack, which also contains code to splice the nest value into the
8072 argument list. This is used to implement the GCC nested function address
8075 For example, if the function is ``i32 f(i8* nest %c, i32 %x, i32 %y)``
8076 then the resulting function pointer has signature ``i32 (i32, i32)*``.
8077 It can be created as follows:
8079 .. code-block:: llvm
8081 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
8082 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
8083 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
8084 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
8085 %fp = bitcast i8* %p to i32 (i32, i32)*
8087 The call ``%val = call i32 %fp(i32 %x, i32 %y)`` is then equivalent to
8088 ``%val = call i32 %f(i8* %nval, i32 %x, i32 %y)``.
8092 '``llvm.init.trampoline``' Intrinsic
8093 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8100 declare void @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
8105 This fills the memory pointed to by ``tramp`` with executable code,
8106 turning it into a trampoline.
8111 The ``llvm.init.trampoline`` intrinsic takes three arguments, all
8112 pointers. The ``tramp`` argument must point to a sufficiently large and
8113 sufficiently aligned block of memory; this memory is written to by the
8114 intrinsic. Note that the size and the alignment are target-specific -
8115 LLVM currently provides no portable way of determining them, so a
8116 front-end that generates this intrinsic needs to have some
8117 target-specific knowledge. The ``func`` argument must hold a function
8118 bitcast to an ``i8*``.
8123 The block of memory pointed to by ``tramp`` is filled with target
8124 dependent code, turning it into a function. Then ``tramp`` needs to be
8125 passed to :ref:`llvm.adjust.trampoline <int_at>` to get a pointer which can
8126 be :ref:`bitcast (to a new function) and called <int_trampoline>`. The new
8127 function's signature is the same as that of ``func`` with any arguments
8128 marked with the ``nest`` attribute removed. At most one such ``nest``
8129 argument is allowed, and it must be of pointer type. Calling the new
8130 function is equivalent to calling ``func`` with the same argument list,
8131 but with ``nval`` used for the missing ``nest`` argument. If, after
8132 calling ``llvm.init.trampoline``, the memory pointed to by ``tramp`` is
8133 modified, then the effect of any later call to the returned function
8134 pointer is undefined.
8138 '``llvm.adjust.trampoline``' Intrinsic
8139 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8146 declare i8* @llvm.adjust.trampoline(i8* <tramp>)
8151 This performs any required machine-specific adjustment to the address of
8152 a trampoline (passed as ``tramp``).
8157 ``tramp`` must point to a block of memory which already has trampoline
8158 code filled in by a previous call to
8159 :ref:`llvm.init.trampoline <int_it>`.
8164 On some architectures the address of the code to be executed needs to be
8165 different to the address where the trampoline is actually stored. This
8166 intrinsic returns the executable address corresponding to ``tramp``
8167 after performing the required machine specific adjustments. The pointer
8168 returned can then be :ref:`bitcast and executed <int_trampoline>`.
8173 This class of intrinsics exists to information about the lifetime of
8174 memory objects and ranges where variables are immutable.
8176 '``llvm.lifetime.start``' Intrinsic
8177 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8184 declare void @llvm.lifetime.start(i64 <size>, i8* nocapture <ptr>)
8189 The '``llvm.lifetime.start``' intrinsic specifies the start of a memory
8195 The first argument is a constant integer representing the size of the
8196 object, or -1 if it is variable sized. The second argument is a pointer
8202 This intrinsic indicates that before this point in the code, the value
8203 of the memory pointed to by ``ptr`` is dead. This means that it is known
8204 to never be used and has an undefined value. A load from the pointer
8205 that precedes this intrinsic can be replaced with ``'undef'``.
8207 '``llvm.lifetime.end``' Intrinsic
8208 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8215 declare void @llvm.lifetime.end(i64 <size>, i8* nocapture <ptr>)
8220 The '``llvm.lifetime.end``' intrinsic specifies the end of a memory
8226 The first argument is a constant integer representing the size of the
8227 object, or -1 if it is variable sized. The second argument is a pointer
8233 This intrinsic indicates that after this point in the code, the value of
8234 the memory pointed to by ``ptr`` is dead. This means that it is known to
8235 never be used and has an undefined value. Any stores into the memory
8236 object following this intrinsic may be removed as dead.
8238 '``llvm.invariant.start``' Intrinsic
8239 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8246 declare {}* @llvm.invariant.start(i64 <size>, i8* nocapture <ptr>)
8251 The '``llvm.invariant.start``' intrinsic specifies that the contents of
8252 a memory object will not change.
8257 The first argument is a constant integer representing the size of the
8258 object, or -1 if it is variable sized. The second argument is a pointer
8264 This intrinsic indicates that until an ``llvm.invariant.end`` that uses
8265 the return value, the referenced memory location is constant and
8268 '``llvm.invariant.end``' Intrinsic
8269 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8276 declare void @llvm.invariant.end({}* <start>, i64 <size>, i8* nocapture <ptr>)
8281 The '``llvm.invariant.end``' intrinsic specifies that the contents of a
8282 memory object are mutable.
8287 The first argument is the matching ``llvm.invariant.start`` intrinsic.
8288 The second argument is a constant integer representing the size of the
8289 object, or -1 if it is variable sized and the third argument is a
8290 pointer to the object.
8295 This intrinsic indicates that the memory is mutable again.
8300 This class of intrinsics is designed to be generic and has no specific
8303 '``llvm.var.annotation``' Intrinsic
8304 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8311 declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int>)
8316 The '``llvm.var.annotation``' intrinsic.
8321 The first argument is a pointer to a value, the second is a pointer to a
8322 global string, the third is a pointer to a global string which is the
8323 source file name, and the last argument is the line number.
8328 This intrinsic allows annotation of local variables with arbitrary
8329 strings. This can be useful for special purpose optimizations that want
8330 to look for these annotations. These have no other defined use; they are
8331 ignored by code generation and optimization.
8333 '``llvm.annotation.*``' Intrinsic
8334 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8339 This is an overloaded intrinsic. You can use '``llvm.annotation``' on
8340 any integer bit width.
8344 declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int>)
8345 declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int>)
8346 declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int>)
8347 declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int>)
8348 declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int>)
8353 The '``llvm.annotation``' intrinsic.
8358 The first argument is an integer value (result of some expression), the
8359 second is a pointer to a global string, the third is a pointer to a
8360 global string which is the source file name, and the last argument is
8361 the line number. It returns the value of the first argument.
8366 This intrinsic allows annotations to be put on arbitrary expressions
8367 with arbitrary strings. This can be useful for special purpose
8368 optimizations that want to look for these annotations. These have no
8369 other defined use; they are ignored by code generation and optimization.
8371 '``llvm.trap``' Intrinsic
8372 ^^^^^^^^^^^^^^^^^^^^^^^^^
8379 declare void @llvm.trap() noreturn nounwind
8384 The '``llvm.trap``' intrinsic.
8394 This intrinsic is lowered to the target dependent trap instruction. If
8395 the target does not have a trap instruction, this intrinsic will be
8396 lowered to a call of the ``abort()`` function.
8398 '``llvm.debugtrap``' Intrinsic
8399 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8406 declare void @llvm.debugtrap() nounwind
8411 The '``llvm.debugtrap``' intrinsic.
8421 This intrinsic is lowered to code which is intended to cause an
8422 execution trap with the intention of requesting the attention of a
8425 '``llvm.stackprotector``' Intrinsic
8426 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8433 declare void @llvm.stackprotector(i8* <guard>, i8** <slot>)
8438 The ``llvm.stackprotector`` intrinsic takes the ``guard`` and stores it
8439 onto the stack at ``slot``. The stack slot is adjusted to ensure that it
8440 is placed on the stack before local variables.
8445 The ``llvm.stackprotector`` intrinsic requires two pointer arguments.
8446 The first argument is the value loaded from the stack guard
8447 ``@__stack_chk_guard``. The second variable is an ``alloca`` that has
8448 enough space to hold the value of the guard.
8453 This intrinsic causes the prologue/epilogue inserter to force the
8454 position of the ``AllocaInst`` stack slot to be before local variables
8455 on the stack. This is to ensure that if a local variable on the stack is
8456 overwritten, it will destroy the value of the guard. When the function
8457 exits, the guard on the stack is checked against the original guard. If
8458 they are different, then the program aborts by calling the
8459 ``__stack_chk_fail()`` function.
8461 '``llvm.objectsize``' Intrinsic
8462 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8469 declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
8470 declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
8475 The ``llvm.objectsize`` intrinsic is designed to provide information to
8476 the optimizers to determine at compile time whether a) an operation
8477 (like memcpy) will overflow a buffer that corresponds to an object, or
8478 b) that a runtime check for overflow isn't necessary. An object in this
8479 context means an allocation of a specific class, structure, array, or
8485 The ``llvm.objectsize`` intrinsic takes two arguments. The first
8486 argument is a pointer to or into the ``object``. The second argument is
8487 a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
8488 or -1 (if false) when the object size is unknown. The second argument
8489 only accepts constants.
8494 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
8495 the size of the object concerned. If the size cannot be determined at
8496 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
8497 on the ``min`` argument).
8499 '``llvm.expect``' Intrinsic
8500 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
8507 declare i32 @llvm.expect.i32(i32 <val>, i32 <expected_val>)
8508 declare i64 @llvm.expect.i64(i64 <val>, i64 <expected_val>)
8513 The ``llvm.expect`` intrinsic provides information about expected (the
8514 most probable) value of ``val``, which can be used by optimizers.
8519 The ``llvm.expect`` intrinsic takes two arguments. The first argument is
8520 a value. The second argument is an expected value, this needs to be a
8521 constant value, variables are not allowed.
8526 This intrinsic is lowered to the ``val``.
8528 '``llvm.donothing``' Intrinsic
8529 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
8536 declare void @llvm.donothing() nounwind readnone
8541 The ``llvm.donothing`` intrinsic doesn't perform any operation. It's the
8542 only intrinsic that can be called with an invoke instruction.
8552 This intrinsic does nothing, and it's removed by optimizers and ignored