__has_builtin¶

This function-like macro takes a single identifier argument that is the name of a builtin function. It evaluates to 1 if the builtin is supported or 0 if not. It can be used like this:

#ifndef __has_builtin         // Optional of course.
  #define __has_builtin(x) 0  // Compatibility with non-clang compilers.
#endif

...
#if __has_builtin(__builtin_trap)
  __builtin_trap();
#else
  abort();
#endif
...

__has_feature and __has_extension¶

These function-like macros take a single identifier argument that is the name of a feature. __has_feature evaluates to 1 if the feature is both supported by Clang and standardized in the current language standard or 0 if not (but see below), while __has_extension evaluates to 1 if the feature is supported by Clang in the current language (either as a language extension or a standard language feature) or 0 if not. They can be used like this:

#ifndef __has_feature         // Optional of course.
  #define __has_feature(x) 0  // Compatibility with non-clang compilers.
#endif
#ifndef __has_extension
  #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
#endif

...
#if __has_feature(cxx_rvalue_references)
// This code will only be compiled with the -std=c++11 and -std=gnu++11
// options, because rvalue references are only standardized in C++11.
#endif

#if __has_extension(cxx_rvalue_references)
// This code will be compiled with the -std=c++11, -std=gnu++11, -std=c++98
// and -std=gnu++98 options, because rvalue references are supported as a
// language extension in C++98.
#endif

For backward compatibility, __has_feature can also be used to test for support for non-standardized features, i.e. features not prefixed c_, cxx_ or objc_.

Another use of __has_feature is to check for compiler features not related to the language standard, such as e.g. AddressSanitizer.

If the -pedantic-errors option is given, __has_extension is equivalent to __has_feature.

The feature tag is described along with the language feature below.

The feature name or extension name can also be specified with a preceding and following __ (double underscore) to avoid interference from a macro with the same name. For instance, __cxx_rvalue_references__ can be used instead of cxx_rvalue_references.

__has_cpp_attribute¶

This function-like macro is available in C++2a by default, and is provided as an extension in earlier language standards. It takes a single argument that is the name of a double-square-bracket-style attribute. The argument can either be a single identifier or a scoped identifier. If the attribute is supported, a nonzero value is returned. If the attribute is a standards-based attribute, this macro returns a nonzero value based on the year and month in which the attribute was voted into the working draft. See WG21 SD-6 for the list of values returned for standards-based attributes. If the attribute is not supported by the current compliation target, this macro evaluates to 0. It can be used like this:

#ifndef __has_cpp_attribute         // For backwards compatibility
  #define __has_cpp_attribute(x) 0
#endif

...
#if __has_cpp_attribute(clang::fallthrough)
#define FALLTHROUGH [[clang::fallthrough]]
#else
#define FALLTHROUGH
#endif
...

The attribute scope tokens clang and _Clang are interchangeable, as are the attribute scope tokens gnu and __gnu__. Attribute tokens in either of these namespaces can be specified with a preceding and following __ (double underscore) to avoid interference from a macro with the same name. For instance, gnu::__const__ can be used instead of gnu::const.

__has_c_attribute¶

This function-like macro takes a single argument that is the name of an attribute exposed with the double square-bracket syntax in C mode. The argument can either be a single identifier or a scoped identifier. If the attribute is supported, a nonzero value is returned. If the attribute is not supported by the current compilation target, this macro evaluates to 0. It can be used like this:

#ifndef __has_c_attribute         // Optional of course.
  #define __has_c_attribute(x) 0  // Compatibility with non-clang compilers.
#endif

...
#if __has_c_attribute(fallthrough)
  #define FALLTHROUGH [[fallthrough]]
#else
  #define FALLTHROUGH
#endif
...

The attribute scope tokens clang and _Clang are interchangeable, as are the attribute scope tokens gnu and __gnu__. Attribute tokens in either of these namespaces can be specified with a preceding and following __ (double underscore) to avoid interference from a macro with the same name. For instance, gnu::__const__ can be used instead of gnu::const.

__has_attribute¶

This function-like macro takes a single identifier argument that is the name of a GNU-style attribute. It evaluates to 1 if the attribute is supported by the current compilation target, or 0 if not. It can be used like this:

#ifndef __has_attribute         // Optional of course.
  #define __has_attribute(x) 0  // Compatibility with non-clang compilers.
#endif

...
#if __has_attribute(always_inline)
#define ALWAYS_INLINE __attribute__((always_inline))
#else
#define ALWAYS_INLINE
#endif
...

The attribute name can also be specified with a preceding and following __ (double underscore) to avoid interference from a macro with the same name. For instance, __always_inline__ can be used instead of always_inline.

__has_declspec_attribute¶

This function-like macro takes a single identifier argument that is the name of an attribute implemented as a Microsoft-style __declspec attribute. It evaluates to 1 if the attribute is supported by the current compilation target, or 0 if not. It can be used like this:

#ifndef __has_declspec_attribute         // Optional of course.
  #define __has_declspec_attribute(x) 0  // Compatibility with non-clang compilers.
#endif

...
#if __has_declspec_attribute(dllexport)
#define DLLEXPORT __declspec(dllexport)
#else
#define DLLEXPORT
#endif
...

The attribute name can also be specified with a preceding and following __ (double underscore) to avoid interference from a macro with the same name. For instance, __dllexport__ can be used instead of dllexport.

__is_identifier¶

This function-like macro takes a single identifier argument that might be either a reserved word or a regular identifier. It evaluates to 1 if the argument is just a regular identifier and not a reserved word, in the sense that it can then be used as the name of a user-defined function or variable. Otherwise it evaluates to 0. It can be used like this:

...
#ifdef __is_identifier          // Compatibility with non-clang compilers.
  #if __is_identifier(__wchar_t)
    typedef wchar_t __wchar_t;
  #endif
#endif

__wchar_t WideCharacter;
...

__has_include¶

This function-like macro takes a single file name string argument that is the name of an include file. It evaluates to 1 if the file can be found using the include paths, or 0 otherwise:

// Note the two possible file name string formats.
#if __has_include("myinclude.h") && __has_include(<stdint.h>)
# include "myinclude.h"
#endif

To test for this feature, use #if defined(__has_include):

// To avoid problem with non-clang compilers not having this macro.
#if defined(__has_include)
#if __has_include("myinclude.h")
# include "myinclude.h"
#endif
#endif

__has_include_next¶

This function-like macro takes a single file name string argument that is the name of an include file. It is like __has_include except that it looks for the second instance of the given file found in the include paths. It evaluates to 1 if the second instance of the file can be found using the include paths, or 0 otherwise:

// Note the two possible file name string formats.
#if __has_include_next("myinclude.h") && __has_include_next(<stdint.h>)
# include_next "myinclude.h"
#endif

// To avoid problem with non-clang compilers not having this macro.
#if defined(__has_include_next)
#if __has_include_next("myinclude.h")
# include_next "myinclude.h"
#endif
#endif

Note that __has_include_next, like the GNU extension #include_next directive, is intended for use in headers only, and will issue a warning if used in the top-level compilation file. A warning will also be issued if an absolute path is used in the file argument.

__has_warning¶

This function-like macro takes a string literal that represents a command line option for a warning and returns true if that is a valid warning option.

#if __has_warning("-Wformat")
...
#endif

Vector Literals¶

Vector literals can be used to create vectors from a set of scalars, or vectors. Either parentheses or braces form can be used. In the parentheses form the number of literal values specified must be one, i.e. referring to a scalar value, or must match the size of the vector type being created. If a single scalar literal value is specified, the scalar literal value will be replicated to all the components of the vector type. In the brackets form any number of literals can be specified. For example:

typedef int v4si __attribute__((__vector_size__(16)));
typedef float float4 __attribute__((ext_vector_type(4)));
typedef float float2 __attribute__((ext_vector_type(2)));

v4si vsi = (v4si){1, 2, 3, 4};
float4 vf = (float4)(1.0f, 2.0f, 3.0f, 4.0f);
vector int vi1 = (vector int)(1);    // vi1 will be (1, 1, 1, 1).
vector int vi2 = (vector int){1};    // vi2 will be (1, 0, 0, 0).
vector int vi3 = (vector int)(1, 2); // error
vector int vi4 = (vector int){1, 2}; // vi4 will be (1, 2, 0, 0).
vector int vi5 = (vector int)(1, 2, 3, 4);
float4 vf = (float4)((float2)(1.0f, 2.0f), (float2)(3.0f, 4.0f));

Vector Operations¶

The table below shows the support for each operation by vector extension. A dash indicates that an operation is not accepted according to a corresponding specification.

See also __builtin_shufflevector, __builtin_convertvector.

C++98¶

The features listed below are part of the C++98 standard. These features are enabled by default when compiling C++ code.

C++11¶

The features listed below are part of the C++11 standard. As a result, all these features are enabled with the -std=c++11 or -std=gnu++11 option when compiling C++ code.

C++14¶

The features listed below are part of the C++14 standard. As a result, all these features are enabled with the -std=C++14 or -std=gnu++14 option when compiling C++ code.

C11¶

The features listed below are part of the C11 standard. As a result, all these features are enabled with the -std=c11 or -std=gnu11 option when compiling C code. Additionally, because these features are all backward-compatible, they are available as extensions in all language modes.

Modules¶

Use __has_feature(modules) to determine if Modules have been enabled. For example, compiling code with -fmodules enables the use of Modules.

More information could be found here.

Related result types¶

According to Cocoa conventions, Objective-C methods with certain names (“init”, “alloc”, etc.) always return objects that are an instance of the receiving class’s type. Such methods are said to have a “related result type”, meaning that a message send to one of these methods will have the same static type as an instance of the receiver class. For example, given the following classes:

@interface NSObject
+ (id)alloc;
- (id)init;
@end

@interface NSArray : NSObject
@end

and this common initialization pattern

NSArray *array = [[NSArray alloc] init];

the type of the expression [NSArray alloc] is NSArray* because alloc implicitly has a related result type. Similarly, the type of the expression [[NSArray alloc] init] is NSArray*, since init has a related result type and its receiver is known to have the type NSArray *. If neither alloc nor init had a related result type, the expressions would have had type id, as declared in the method signature.

A method with a related result type can be declared by using the type instancetype as its result type. instancetype is a contextual keyword that is only permitted in the result type of an Objective-C method, e.g.

@interface A
+ (instancetype)constructAnA;
@end

The related result type can also be inferred for some methods. To determine whether a method has an inferred related result type, the first word in the camel-case selector (e.g., “init” in “initWithObjects”) is considered, and the method will have a related result type if its return type is compatible with the type of its class and if:

• the first word is “alloc” or “new”, and the method is a class method, or
• the first word is “autorelease”, “init”, “retain”, or “self”, and the method is an instance method.

If a method with a related result type is overridden by a subclass method, the subclass method must also return a type that is compatible with the subclass type. For example:

@interface NSString : NSObject
- (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString
@end

Related result types only affect the type of a message send or property access via the given method. In all other respects, a method with a related result type is treated the same way as method that returns id.

Use __has_feature(objc_instancetype) to determine whether the instancetype contextual keyword is available.

Automatic reference counting¶

Clang provides support for automated reference counting in Objective-C, which eliminates the need for manual retain/release/autorelease message sends. There are three feature macros associated with automatic reference counting: __has_feature(objc_arc) indicates the availability of automated reference counting in general, while __has_feature(objc_arc_weak) indicates that automated reference counting also includes support for __weak pointers to Objective-C objects. __has_feature(objc_arc_fields) indicates that C structs are allowed to have fields that are pointers to Objective-C objects managed by automatic reference counting.

Weak references¶

Clang supports ARC-style weak and unsafe references in Objective-C even outside of ARC mode. Weak references must be explicitly enabled with the -fobjc-weak option; use __has_feature((objc_arc_weak)) to test whether they are enabled. Unsafe references are enabled unconditionally. ARC-style weak and unsafe references cannot be used when Objective-C garbage collection is enabled.

Except as noted below, the language rules for the __weak and __unsafe_unretained qualifiers (and the weak and unsafe_unretained property attributes) are just as laid out in the ARC specification. In particular, note that some classes do not support forming weak references to their instances, and note that special care must be taken when storing weak references in memory where initialization and deinitialization are outside the responsibility of the compiler (such as in malloc-ed memory).

Loading from a __weak variable always implicitly retains the loaded value. In non-ARC modes, this retain is normally balanced by an implicit autorelease. This autorelease can be suppressed by performing the load in the receiver position of a -retain message send (e.g. [weakReference retain]); note that this performs only a single retain (the retain done when primitively loading from the weak reference).

For the most part, __unsafe_unretained in non-ARC modes is just the default behavior of variables and therefore is not needed. However, it does have an effect on the semantics of block captures: normally, copying a block which captures an Objective-C object or block pointer causes the captured pointer to be retained or copied, respectively, but that behavior is suppressed when the captured variable is qualified with __unsafe_unretained.

Note that the __weak qualifier formerly meant the GC qualifier in all non-ARC modes and was silently ignored outside of GC modes. It now means the ARC-style qualifier in all non-GC modes and is no longer allowed if not enabled by either -fobjc-arc or -fobjc-weak. It is expected that -fobjc-weak will eventually be enabled by default in all non-GC Objective-C modes.

Enumerations with a fixed underlying type¶

Clang provides support for C++11 enumerations with a fixed underlying type within Objective-C. For example, one can write an enumeration type as:

typedef enum : unsigned char { Red, Green, Blue } Color;

This specifies that the underlying type, which is used to store the enumeration value, is unsigned char.

Use __has_feature(objc_fixed_enum) to determine whether support for fixed underlying types is available in Objective-C.

Interoperability with C++11 lambdas¶

Clang provides interoperability between C++11 lambdas and blocks-based APIs, by permitting a lambda to be implicitly converted to a block pointer with the corresponding signature. For example, consider an API such as NSArray’s array-sorting method:

- (NSArray *)sortedArrayUsingComparator:(NSComparator)cmptr;

NSComparator is simply a typedef for the block pointer NSComparisonResult (^)(id, id), and parameters of this type are generally provided with block literals as arguments. However, one can also use a C++11 lambda so long as it provides the same signature (in this case, accepting two parameters of type id and returning an NSComparisonResult):

NSArray *array = @[@"string 1", @"string 21", @"string 12", @"String 11",
                   @"String 02"];
const NSStringCompareOptions comparisonOptions
  = NSCaseInsensitiveSearch | NSNumericSearch |
    NSWidthInsensitiveSearch | NSForcedOrderingSearch;
NSLocale *currentLocale = [NSLocale currentLocale];
NSArray *sorted
  = [array sortedArrayUsingComparator:[=](id s1, id s2) -> NSComparisonResult {
             NSRange string1Range = NSMakeRange(0, [s1 length]);
             return [s1 compare:s2 options:comparisonOptions
             range:string1Range locale:currentLocale];
     }];
NSLog(@"sorted: %@", sorted);

This code relies on an implicit conversion from the type of the lambda expression (an unnamed, local class type called the closure type) to the corresponding block pointer type. The conversion itself is expressed by a conversion operator in that closure type that produces a block pointer with the same signature as the lambda itself, e.g.,

operator NSComparisonResult (^)(id, id)() const;

This conversion function returns a new block that simply forwards the two parameters to the lambda object (which it captures by copy), then returns the result. The returned block is first copied (with Block_copy) and then autoreleased. As an optimization, if a lambda expression is immediately converted to a block pointer (as in the first example, above), then the block is not copied and autoreleased: rather, it is given the same lifetime as a block literal written at that point in the program, which avoids the overhead of copying a block to the heap in the common case.

The conversion from a lambda to a block pointer is only available in Objective-C++, and not in C++ with blocks, due to its use of Objective-C memory management (autorelease).

Object Literals and Subscripting¶

Clang provides support for Object Literals and Subscripting in Objective-C, which simplifies common Objective-C programming patterns, makes programs more concise, and improves the safety of container creation. There are several feature macros associated with object literals and subscripting: __has_feature(objc_array_literals) tests the availability of array literals; __has_feature(objc_dictionary_literals) tests the availability of dictionary literals; __has_feature(objc_subscripting) tests the availability of object subscripting.

Objective-C Autosynthesis of Properties¶

Clang provides support for autosynthesis of declared properties. Using this feature, clang provides default synthesis of those properties not declared @dynamic and not having user provided backing getter and setter methods. __has_feature(objc_default_synthesize_properties) checks for availability of this feature in version of clang being used.

Objective-C retaining behavior attributes¶

In Objective-C, functions and methods are generally assumed to follow the Cocoa Memory Management conventions for ownership of object arguments and return values. However, there are exceptions, and so Clang provides attributes to allow these exceptions to be documented. This are used by ARC and the static analyzer Some exceptions may be better described using the objc_method_family attribute instead.

Usage: The ns_returns_retained, ns_returns_not_retained, ns_returns_autoreleased, cf_returns_retained, and cf_returns_not_retained attributes can be placed on methods and functions that return Objective-C or CoreFoundation objects. They are commonly placed at the end of a function prototype or method declaration:

id foo() __attribute__((ns_returns_retained));

- (NSString *)bar:(int)x __attribute__((ns_returns_retained));

The *_returns_retained attributes specify that the returned object has a +1 retain count. The *_returns_not_retained attributes specify that the return object has a +0 retain count, even if the normal convention for its selector would be +1. ns_returns_autoreleased specifies that the returned object is +0, but is guaranteed to live at least as long as the next flush of an autorelease pool.

Usage: The ns_consumed and cf_consumed attributes can be placed on an parameter declaration; they specify that the argument is expected to have a +1 retain count, which will be balanced in some way by the function or method. The ns_consumes_self attribute can only be placed on an Objective-C method; it specifies that the method expects its self parameter to have a +1 retain count, which it will balance in some way.

void foo(__attribute__((ns_consumed)) NSString *string);

- (void) bar __attribute__((ns_consumes_self));
- (void) baz:(id) __attribute__((ns_consumed)) x;

Further examples of these attributes are available in the static analyzer’s list of annotations for analysis.

Query for these features with __has_attribute(ns_consumed), __has_attribute(ns_returns_retained), etc.

Objective-C @available¶

It is possible to use the newest SDK but still build a program that can run on older versions of macOS and iOS by passing -mmacosx-version-min= / -miphoneos-version-min=.

Before LLVM 5.0, when calling a function that exists only in the OS that’s newer than the target OS (as determined by the minimum deployment version), programmers had to carefully check if the function exists at runtime, using null checks for weakly-linked C functions, +class for Objective-C classes, and -respondsToSelector: or +instancesRespondToSelector: for Objective-C methods. If such a check was missed, the program would compile fine, run fine on newer systems, but crash on older systems.

As of LLVM 5.0, -Wunguarded-availability uses the availability attributes together with the new @available() keyword to assist with this issue. When a method that’s introduced in the OS newer than the target OS is called, a -Wunguarded-availability warning is emitted if that call is not guarded:

void my_fun(NSSomeClass* var) {
  // If fancyNewMethod was added in e.g. macOS 10.12, but the code is
  // built with -mmacosx-version-min=10.11, then this unconditional call
  // will emit a -Wunguarded-availability warning:
  [var fancyNewMethod];
}

To fix the warning and to avoid the crash on macOS 10.11, wrap it in if(@available()):

void my_fun(NSSomeClass* var) {
  if (@available(macOS 10.12, *)) {
    [var fancyNewMethod];
  } else {
    // Put fallback behavior for old macOS versions (and for non-mac
    // platforms) here.
  }
}

The * is required and means that platforms not explicitly listed will take the true branch, and the compiler will emit -Wunguarded-availability warnings for unlisted platforms based on those platform’s deployment target. More than one platform can be listed in @available():

void my_fun(NSSomeClass* var) {
  if (@available(macOS 10.12, iOS 10, *)) {
    [var fancyNewMethod];
  }
}

If the caller of my_fun() already checks that my_fun() is only called on 10.12, then add an availability attribute to it, which will also suppress the warning and require that calls to my_fun() are checked:

API_AVAILABLE(macos(10.12)) void my_fun(NSSomeClass* var) {
  [var fancyNewMethod];  // Now ok.
}

@available() is only available in Objective-C code. To use the feature in C and C++ code, use the __builtin_available() spelling instead.

If existing code uses null checks or -respondsToSelector:, it should be changed to use @available() (or __builtin_available) instead.

-Wunguarded-availability is disabled by default, but -Wunguarded-availability-new, which only emits this warning for APIs that have been introduced in macOS >= 10.13, iOS >= 11, watchOS >= 4 and tvOS >= 11, is enabled by default.

Objective-C++ ABI: protocol-qualifier mangling of parameters¶

Starting with LLVM 3.4, Clang produces a new mangling for parameters whose type is a qualified-id (e.g., id<Foo>). This mangling allows such parameters to be differentiated from those with the regular unqualified id type.

This was a non-backward compatible mangling change to the ABI. This change allows proper overloading, and also prevents mangling conflicts with template parameters of protocol-qualified type.

Query the presence of this new mangling with __has_feature(objc_protocol_qualifier_mangling).

C++ for OpenCL¶

This functionality is built on top of OpenCL C v2.0 and C++17 enabling most of regular C++ features in OpenCL kernel code. Most functionality from OpenCL C is inherited. This section describes minor differences to OpenCL C and any limitations related to C++ support as well as interactions between OpenCL and C++ features that are not documented elsewhere.

template <typename T>
void foo() {
  T m; // address space of m will be known at template instantiation time.
  T * ptr; // ptr points to __generic address space object.
  T & ref = ...; // ref references an object in __generic address space.
};

template <int N>
struct S {
  int i; // i has no address space
  static int ii; // ii is in global address space
  int * ptr; // ptr points to __generic address space int.
  int & ref = ...; // ref references int in __generic address space.
};

template <int N>
void bar()
{
  S<N> s; // s is in __private address space
}

__private int & ref = ...; // references int in __private address space

struct C {
   void foo() __local;
   void foo();
};

__kernel void bar() {
  __local C c1;
  C c2;
  __constant C c3;
  c1.foo(); // will resolve to the first foo
  c2.foo(); // will resolve to the second foo
  c3.foo(); // error due to mismatching address spaces - can't convert to
            // __local or __generic
}

class C {
  // Has the following implicit definition
  // void C() __generic;
  // void C(const __generic C &) __generic;
  // void C(__generic C &&) __generic;
  // operator= '__generic C &(__generic C &&)'
  // operator= '__generic C &(const __generic C &) __generic
}

 1 template<typename T>
 2 void foo(T* i){
 3   T var;
 4 }
 5
 6 __global int g;
 7 void bar(){
 8   foo(&g); // error: template instantiation failed as function scope variable
 9            // appears to be declared in __global address space (see line 3)
10 }

template <typename T>
void foo() {
  __private T var;
}

void bar() {
  foo<__global int>(); // error: conflicting address space qualifiers are provided
                       // __global and __private
}

template<typename T>
void foo(){
  T var;
}

void bar(){
  foo<__global int>(); // error: function scope variable cannot be declared in
                       // __global address space
}

int bar(const unsigned int &i);

void foo() {
  bar(1); // temporary is created in __private address space but converted
          // to __generic address space of parameter reference
}

__global const int& f(__global float &ref) {
  return ref; // error: address space mismatch between temporary object
              // created to hold value converted float->int and return
              // value type (can't convert from __private to __global)
}

clang -cl-std=clc++ test.cl

__builtin_assume¶

__builtin_assume is used to provide the optimizer with a boolean invariant that is defined to be true.

Syntax:

__builtin_assume(bool)

Example of Use:

int foo(int x) {
  __builtin_assume(x != 0);

  // The optimizer may short-circuit this check using the invariant.
  if (x == 0)
    return do_something();

  return do_something_else();
}

Description:

The boolean argument to this function is defined to be true. The optimizer may analyze the form of the expression provided as the argument and deduce from that information used to optimize the program. If the condition is violated during execution, the behavior is undefined. The argument itself is never evaluated, so any side effects of the expression will be discarded.

Query for this feature with __has_builtin(__builtin_assume).

__builtin_readcyclecounter¶

__builtin_readcyclecounter is used to access the cycle counter register (or a similar low-latency, high-accuracy clock) on those targets that support it.

Syntax:

__builtin_readcyclecounter()

Example of Use:

unsigned long long t0 = __builtin_readcyclecounter();
do_something();
unsigned long long t1 = __builtin_readcyclecounter();
unsigned long long cycles_to_do_something = t1 - t0; // assuming no overflow

Description:

The __builtin_readcyclecounter() builtin returns the cycle counter value, which may be either global or process/thread-specific depending on the target. As the backing counters often overflow quickly (on the order of seconds) this should only be used for timing small intervals. When not supported by the target, the return value is always zero. This builtin takes no arguments and produces an unsigned long long result.

Query for this feature with __has_builtin(__builtin_readcyclecounter). Note that even if present, its use may depend on run-time privilege or other OS controlled state.

__builtin_shufflevector¶

__builtin_shufflevector is used to express generic vector permutation/shuffle/swizzle operations. This builtin is also very important for the implementation of various target-specific header files like <xmmintrin.h>.

Syntax:

__builtin_shufflevector(vec1, vec2, index1, index2, ...)

Examples:

// identity operation - return 4-element vector v1.
__builtin_shufflevector(v1, v1, 0, 1, 2, 3)

// "Splat" element 0 of V1 into a 4-element result.
__builtin_shufflevector(V1, V1, 0, 0, 0, 0)

// Reverse 4-element vector V1.
__builtin_shufflevector(V1, V1, 3, 2, 1, 0)

// Concatenate every other element of 4-element vectors V1 and V2.
__builtin_shufflevector(V1, V2, 0, 2, 4, 6)

// Concatenate every other element of 8-element vectors V1 and V2.
__builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14)

// Shuffle v1 with some elements being undefined
__builtin_shufflevector(v1, v1, 3, -1, 1, -1)

Description:

The first two arguments to __builtin_shufflevector are vectors that have the same element type. The remaining arguments are a list of integers that specify the elements indices of the first two vectors that should be extracted and returned in a new vector. These element indices are numbered sequentially starting with the first vector, continuing into the second vector. Thus, if vec1 is a 4-element vector, index 5 would refer to the second element of vec2. An index of -1 can be used to indicate that the corresponding element in the returned vector is a don’t care and can be optimized by the backend.

The result of __builtin_shufflevector is a vector with the same element type as vec1/vec2 but that has an element count equal to the number of indices specified.

Query for this feature with __has_builtin(__builtin_shufflevector).

__builtin_convertvector¶

__builtin_convertvector is used to express generic vector type-conversion operations. The input vector and the output vector type must have the same number of elements.

Syntax:

__builtin_convertvector(src_vec, dst_vec_type)

Examples:

typedef double vector4double __attribute__((__vector_size__(32)));
typedef float  vector4float  __attribute__((__vector_size__(16)));
typedef short  vector4short  __attribute__((__vector_size__(8)));
vector4float vf; vector4short vs;

// convert from a vector of 4 floats to a vector of 4 doubles.
__builtin_convertvector(vf, vector4double)
// equivalent to:
(vector4double) { (double) vf[0], (double) vf[1], (double) vf[2], (double) vf[3] }

// convert from a vector of 4 shorts to a vector of 4 floats.
__builtin_convertvector(vs, vector4float)
// equivalent to:
(vector4float) { (float) vs[0], (float) vs[1], (float) vs[2], (float) vs[3] }

Description:

The first argument to __builtin_convertvector is a vector, and the second argument is a vector type with the same number of elements as the first argument.

The result of __builtin_convertvector is a vector with the same element type as the second argument, with a value defined in terms of the action of a C-style cast applied to each element of the first argument.

Query for this feature with __has_builtin(__builtin_convertvector).

__builtin_bitreverse¶

• __builtin_bitreverse8
• __builtin_bitreverse16
• __builtin_bitreverse32
• __builtin_bitreverse64

Syntax:

__builtin_bitreverse32(x)

Examples:

uint8_t rev_x = __builtin_bitreverse8(x);
uint16_t rev_x = __builtin_bitreverse16(x);
uint32_t rev_y = __builtin_bitreverse32(y);
uint64_t rev_z = __builtin_bitreverse64(z);

Description:

The ‘__builtin_bitreverse’ family of builtins is used to reverse the bitpattern of an integer value; for example 0b10110110 becomes 0b01101101.

__builtin_rotateleft¶

• __builtin_rotateleft8
• __builtin_rotateleft16
• __builtin_rotateleft32
• __builtin_rotateleft64

Syntax:

__builtin_rotateleft32(x, y)

Examples:

uint8_t rot_x = __builtin_rotateleft8(x, y);
uint16_t rot_x = __builtin_rotateleft16(x, y);
uint32_t rot_x = __builtin_rotateleft32(x, y);
uint64_t rot_x = __builtin_rotateleft64(x, y);

Description:

The ‘__builtin_rotateleft’ family of builtins is used to rotate the bits in the first argument by the amount in the second argument. For example, 0b10000110 rotated left by 11 becomes 0b00110100. The shift value is treated as an unsigned amount modulo the size of the arguments. Both arguments and the result have the bitwidth specified by the name of the builtin.

__builtin_rotateright¶

• __builtin_rotateright8
• __builtin_rotateright16
• __builtin_rotateright32
• __builtin_rotateright64

Syntax:

__builtin_rotateright32(x, y)

Examples:

uint8_t rot_x = __builtin_rotateright8(x, y);
uint16_t rot_x = __builtin_rotateright16(x, y);
uint32_t rot_x = __builtin_rotateright32(x, y);
uint64_t rot_x = __builtin_rotateright64(x, y);

Description:

The ‘__builtin_rotateright’ family of builtins is used to rotate the bits in the first argument by the amount in the second argument. For example, 0b10000110 rotated right by 3 becomes 0b11010000. The shift value is treated as an unsigned amount modulo the size of the arguments. Both arguments and the result have the bitwidth specified by the name of the builtin.

__builtin_unreachable¶

__builtin_unreachable is used to indicate that a specific point in the program cannot be reached, even if the compiler might otherwise think it can. This is useful to improve optimization and eliminates certain warnings. For example, without the __builtin_unreachable in the example below, the compiler assumes that the inline asm can fall through and prints a “function declared ‘noreturn’ should not return” warning.

Syntax:

__builtin_unreachable()

Example of use:

void myabort(void) __attribute__((noreturn));
void myabort(void) {
  asm("int3");
  __builtin_unreachable();
}

Description:

The __builtin_unreachable() builtin has completely undefined behavior. Since it has undefined behavior, it is a statement that it is never reached and the optimizer can take advantage of this to produce better code. This builtin takes no arguments and produces a void result.

Query for this feature with __has_builtin(__builtin_unreachable).

__builtin_unpredictable¶

__builtin_unpredictable is used to indicate that a branch condition is unpredictable by hardware mechanisms such as branch prediction logic.

Syntax:

__builtin_unpredictable(long long)

Example of use:

if (__builtin_unpredictable(x > 0)) {
   foo();
}

Description:

The __builtin_unpredictable() builtin is expected to be used with control flow conditions such as in if and switch statements.

Query for this feature with __has_builtin(__builtin_unpredictable).

__sync_swap¶

__sync_swap is used to atomically swap integers or pointers in memory.

Syntax:

type __sync_swap(type *ptr, type value, ...)

Example of Use:

int old_value = __sync_swap(&value, new_value);

Description:

The __sync_swap() builtin extends the existing __sync_*() family of atomic intrinsics to allow code to atomically swap the current value with the new value. More importantly, it helps developers write more efficient and correct code by avoiding expensive loops around __sync_bool_compare_and_swap() or relying on the platform specific implementation details of __sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier.

__builtin_addressof¶

__builtin_addressof performs the functionality of the built-in & operator, ignoring any operator& overload. This is useful in constant expressions in C++11, where there is no other way to take the address of an object that overloads operator&.

Example of use:

template<typename T> constexpr T *addressof(T &value) {
  return __builtin_addressof(value);
}

__builtin_operator_new and __builtin_operator_delete¶

__builtin_operator_new allocates memory just like a non-placement non-class new-expression. This is exactly like directly calling the normal non-placement ::operator new, except that it allows certain optimizations that the C++ standard does not permit for a direct function call to ::operator new (in particular, removing new / delete pairs and merging allocations).

Likewise, __builtin_operator_delete deallocates memory just like a non-class delete-expression, and is exactly like directly calling the normal ::operator delete, except that it permits optimizations. Only the unsized form of __builtin_operator_delete is currently available.

These builtins are intended for use in the implementation of std::allocator and other similar allocation libraries, and are only available in C++.

__builtin_preserve_access_index¶

__builtin_preserve_access_index specifies a code section where array subscript access and structure/union member access are relocatable under bpf compile-once run-everywhere framework. Debuginfo (typically with -g) is needed, otherwise, the compiler will exit with an error.

Syntax:

const void * __builtin_preserve_access_index(const void * ptr)

Example of Use:

struct t {
  int i;
  int j;
  union {
    int a;
    int b;
  } c[4];
};
struct t *v = ...;
const void *pb =__builtin_preserve_access_index(&v->c[3].b);

Multiprecision Arithmetic Builtins¶

Clang provides a set of builtins which expose multiprecision arithmetic in a manner amenable to C. They all have the following form:

unsigned x = ..., y = ..., carryin = ..., carryout;
unsigned sum = __builtin_addc(x, y, carryin, &carryout);

Thus one can form a multiprecision addition chain in the following manner:

unsigned *x, *y, *z, carryin=0, carryout;
z[0] = __builtin_addc(x[0], y[0], carryin, &carryout);
carryin = carryout;
z[1] = __builtin_addc(x[1], y[1], carryin, &carryout);
carryin = carryout;
z[2] = __builtin_addc(x[2], y[2], carryin, &carryout);
carryin = carryout;
z[3] = __builtin_addc(x[3], y[3], carryin, &carryout);

The complete list of builtins are:

unsigned char      __builtin_addcb (unsigned char x, unsigned char y, unsigned char carryin, unsigned char *carryout);
unsigned short     __builtin_addcs (unsigned short x, unsigned short y, unsigned short carryin, unsigned short *carryout);
unsigned           __builtin_addc  (unsigned x, unsigned y, unsigned carryin, unsigned *carryout);
unsigned long      __builtin_addcl (unsigned long x, unsigned long y, unsigned long carryin, unsigned long *carryout);
unsigned long long __builtin_addcll(unsigned long long x, unsigned long long y, unsigned long long carryin, unsigned long long *carryout);
unsigned char      __builtin_subcb (unsigned char x, unsigned char y, unsigned char carryin, unsigned char *carryout);
unsigned short     __builtin_subcs (unsigned short x, unsigned short y, unsigned short carryin, unsigned short *carryout);
unsigned           __builtin_subc  (unsigned x, unsigned y, unsigned carryin, unsigned *carryout);
unsigned long      __builtin_subcl (unsigned long x, unsigned long y, unsigned long carryin, unsigned long *carryout);
unsigned long long __builtin_subcll(unsigned long long x, unsigned long long y, unsigned long long carryin, unsigned long long *carryout);

Checked Arithmetic Builtins¶

Clang provides a set of builtins that implement checked arithmetic for security critical applications in a manner that is fast and easily expressable in C. As an example of their usage:

errorcode_t security_critical_application(...) {
  unsigned x, y, result;
  ...
  if (__builtin_mul_overflow(x, y, &result))
    return kErrorCodeHackers;
  ...
  use_multiply(result);
  ...
}

Clang provides the following checked arithmetic builtins:

bool __builtin_add_overflow   (type1 x, type2 y, type3 *sum);
bool __builtin_sub_overflow   (type1 x, type2 y, type3 *diff);
bool __builtin_mul_overflow   (type1 x, type2 y, type3 *prod);
bool __builtin_uadd_overflow  (unsigned x, unsigned y, unsigned *sum);
bool __builtin_uaddl_overflow (unsigned long x, unsigned long y, unsigned long *sum);
bool __builtin_uaddll_overflow(unsigned long long x, unsigned long long y, unsigned long long *sum);
bool __builtin_usub_overflow  (unsigned x, unsigned y, unsigned *diff);
bool __builtin_usubl_overflow (unsigned long x, unsigned long y, unsigned long *diff);
bool __builtin_usubll_overflow(unsigned long long x, unsigned long long y, unsigned long long *diff);
bool __builtin_umul_overflow  (unsigned x, unsigned y, unsigned *prod);
bool __builtin_umull_overflow (unsigned long x, unsigned long y, unsigned long *prod);
bool __builtin_umulll_overflow(unsigned long long x, unsigned long long y, unsigned long long *prod);
bool __builtin_sadd_overflow  (int x, int y, int *sum);
bool __builtin_saddl_overflow (long x, long y, long *sum);
bool __builtin_saddll_overflow(long long x, long long y, long long *sum);
bool __builtin_ssub_overflow  (int x, int y, int *diff);
bool __builtin_ssubl_overflow (long x, long y, long *diff);
bool __builtin_ssubll_overflow(long long x, long long y, long long *diff);
bool __builtin_smul_overflow  (int x, int y, int *prod);
bool __builtin_smull_overflow (long x, long y, long *prod);
bool __builtin_smulll_overflow(long long x, long long y, long long *prod);

Each builtin performs the specified mathematical operation on the first two arguments and stores the result in the third argument. If possible, the result will be equal to mathematically-correct result and the builtin will return 0. Otherwise, the builtin will return 1 and the result will be equal to the unique value that is equivalent to the mathematically-correct result modulo two raised to the k power, where k is the number of bits in the result type. The behavior of these builtins is well-defined for all argument values.

The first three builtins work generically for operands of any integer type, including boolean types. The operands need not have the same type as each other, or as the result. The other builtins may implicitly promote or convert their operands before performing the operation.

Query for this feature with __has_builtin(__builtin_add_overflow), etc.

Floating point builtins¶

__builtin_canonicalize¶

double __builtin_canonicalize(double);
float __builtin_canonicalizef(float);
long double__builtin_canonicalizel(long double);

Returns the platform specific canonical encoding of a floating point number. This canonicalization is useful for implementing certain numeric primitives such as frexp. See LLVM canonicalize intrinsic for more information on the semantics.

String builtins¶

Clang provides constant expression evaluation support for builtins forms of the following functions from the C standard library <string.h> header:

• memchr
• memcmp
• strchr
• strcmp
• strlen
• strncmp
• wcschr
• wcscmp
• wcslen
• wcsncmp
• wmemchr
• wmemcmp

In each case, the builtin form has the name of the C library function prefixed by __builtin_. Example:

void *p = __builtin_memchr("foobar", 'b', 5);

In addition to the above, one further builtin is provided:

char *__builtin_char_memchr(const char *haystack, int needle, size_t size);

__builtin_char_memchr(a, b, c) is identical to (char*)__builtin_memchr(a, b, c) except that its use is permitted within constant expressions in C++11 onwards (where a cast from void* to char* is disallowed in general).

Support for constant expression evaluation for the above builtins be detected with __has_feature(cxx_constexpr_string_builtins).

Atomic Min/Max builtins with memory ordering¶

There are two atomic builtins with min/max in-memory comparison and swap. The syntax and semantics are similar to GCC-compatible __atomic_* builtins.

• __atomic_fetch_min
• __atomic_fetch_max

The builtins work with signed and unsigned integers and require to specify memory ordering. The return value is the original value that was stored in memory before comparison.

Example:

unsigned int val = __atomic_fetch_min(unsigned int *pi, unsigned int ui, __ATOMIC_RELAXED);

The third argument is one of the memory ordering specifiers __ATOMIC_RELAXED, __ATOMIC_CONSUME, __ATOMIC_ACQUIRE, __ATOMIC_RELEASE, __ATOMIC_ACQ_REL, or __ATOMIC_SEQ_CST following C++11 memory model semantics.

In terms or aquire-release ordering barriers these two operations are always considered as operations with load-store semantics, even when the original value is not actually modified after comparison.

__c11_atomic builtins¶

Clang provides a set of builtins which are intended to be used to implement C11’s <stdatomic.h> header. These builtins provide the semantics of the _explicit form of the corresponding C11 operation, and are named with a __c11_ prefix. The supported operations, and the differences from the corresponding C11 operations, are:

• __c11_atomic_init
• __c11_atomic_thread_fence
• __c11_atomic_signal_fence
• __c11_atomic_is_lock_free (The argument is the size of the _Atomic(...) object, instead of its address)
• __c11_atomic_store
• __c11_atomic_load
• __c11_atomic_exchange
• __c11_atomic_compare_exchange_strong
• __c11_atomic_compare_exchange_weak
• __c11_atomic_fetch_add
• __c11_atomic_fetch_sub
• __c11_atomic_fetch_and
• __c11_atomic_fetch_or
• __c11_atomic_fetch_xor

The macros __ATOMIC_RELAXED, __ATOMIC_CONSUME, __ATOMIC_ACQUIRE, __ATOMIC_RELEASE, __ATOMIC_ACQ_REL, and __ATOMIC_SEQ_CST are provided, with values corresponding to the enumerators of C11’s memory_order enumeration.

(Note that Clang additionally provides GCC-compatible __atomic_* builtins and OpenCL 2.0 __opencl_atomic_* builtins. The OpenCL 2.0 atomic builtins are an explicit form of the corresponding OpenCL 2.0 builtin function, and are named with a __opencl_ prefix. The macros __OPENCL_MEMORY_SCOPE_WORK_ITEM, __OPENCL_MEMORY_SCOPE_WORK_GROUP, __OPENCL_MEMORY_SCOPE_DEVICE, __OPENCL_MEMORY_SCOPE_ALL_SVM_DEVICES, and __OPENCL_MEMORY_SCOPE_SUB_GROUP are provided, with values corresponding to the enumerators of OpenCL’s memory_scope enumeration.)

Low-level ARM exclusive memory builtins¶

Clang provides overloaded builtins giving direct access to the three key ARM instructions for implementing atomic operations.

T __builtin_arm_ldrex(const volatile T *addr);
T __builtin_arm_ldaex(const volatile T *addr);
int __builtin_arm_strex(T val, volatile T *addr);
int __builtin_arm_stlex(T val, volatile T *addr);
void __builtin_arm_clrex(void);

The types T currently supported are:

• Integer types with width at most 64 bits (or 128 bits on AArch64).
• Floating-point types
• Pointer types.

Note that the compiler does not guarantee it will not insert stores which clear the exclusive monitor in between an ldrex type operation and its paired strex. In practice this is only usually a risk when the extra store is on the same cache line as the variable being modified and Clang will only insert stack stores on its own, so it is best not to use these operations on variables with automatic storage duration.

Also, loads and stores may be implicit in code written between the ldrex and strex. Clang will not necessarily mitigate the effects of these either, so care should be exercised.

For these reasons the higher level atomic primitives should be preferred where possible.

Non-temporal load/store builtins¶

Clang provides overloaded builtins allowing generation of non-temporal memory accesses.

T __builtin_nontemporal_load(T *addr);
void __builtin_nontemporal_store(T value, T *addr);

The types T currently supported are:

• Integer types.
• Floating-point types.
• Vector types.

Note that the compiler does not guarantee that non-temporal loads or stores will be used.

C++ Coroutines support builtins¶

Clang provides experimental builtins to support C++ Coroutines as defined by https://wg21.link/P0057. The following four are intended to be used by the standard library to implement std::experimental::coroutine_handle type.

Syntax:

void  __builtin_coro_resume(void *addr);
void  __builtin_coro_destroy(void *addr);
bool  __builtin_coro_done(void *addr);
void *__builtin_coro_promise(void *addr, int alignment, bool from_promise)

Example of use:

template <> struct coroutine_handle<void> {
  void resume() const { __builtin_coro_resume(ptr); }
  void destroy() const { __builtin_coro_destroy(ptr); }
  bool done() const { return __builtin_coro_done(ptr); }
  // ...
protected:
  void *ptr;
};

template <typename Promise> struct coroutine_handle : coroutine_handle<> {
  // ...
  Promise &promise() const {
    return *reinterpret_cast<Promise *>(
      __builtin_coro_promise(ptr, alignof(Promise), /*from-promise=*/false));
  }
  static coroutine_handle from_promise(Promise &promise) {
    coroutine_handle p;
    p.ptr = __builtin_coro_promise(&promise, alignof(Promise),
                                                    /*from-promise=*/true);
    return p;
  }
};

Other coroutine builtins are either for internal clang use or for use during development of the coroutine feature. See Coroutines in LLVM for more information on their semantics. Note that builtins matching the intrinsics that take token as the first parameter (llvm.coro.begin, llvm.coro.alloc, llvm.coro.free and llvm.coro.suspend) omit the token parameter and fill it to an appropriate value during the emission.

Syntax:

size_t __builtin_coro_size()
void  *__builtin_coro_frame()
void  *__builtin_coro_free(void *coro_frame)

void  *__builtin_coro_id(int align, void *promise, void *fnaddr, void *parts)
bool   __builtin_coro_alloc()
void  *__builtin_coro_begin(void *memory)
void   __builtin_coro_end(void *coro_frame, bool unwind)
char   __builtin_coro_suspend(bool final)
bool   __builtin_coro_param(void *original, void *copy)

Note that there is no builtin matching the llvm.coro.save intrinsic. LLVM automatically will insert one if the first argument to llvm.coro.suspend is token none. If a user calls __builin_suspend, clang will insert token none as the first argument to the intrinsic.

Source location builtins¶

Clang provides experimental builtins to support C++ standard library implementation of std::experimental::source_location as specified in http://wg21.link/N4600. With the exception of __builtin_COLUMN, these builtins are also implemented by GCC.

Syntax:

const char *__builtin_FILE();
const char *__builtin_FUNCTION();
unsigned    __builtin_LINE();
unsigned    __builtin_COLUMN(); // Clang only

Example of use:

void my_assert(bool pred, int line = __builtin_LINE(), // Captures line of caller
               const char* file = __builtin_FILE(),
               const char* function = __builtin_FUNCTION()) {
  if (pred) return;
  printf("%s:%d assertion failed in function %s\n", file, line, function);
  std::abort();
}

struct MyAggregateType {
  int x;
  int line = __builtin_LINE(); // captures line where aggregate initialization occurs
};
static_assert(MyAggregateType{42}.line == __LINE__);

struct MyClassType {
  int line = __builtin_LINE(); // captures line of the constructor used during initialization
  constexpr MyClassType(int) { assert(line == __LINE__); }
};

Description:

The builtins __builtin_LINE, __builtin_FUNCTION, and __builtin_FILE return the values, at the “invocation point”, for __LINE__, __FUNCTION__, and __FILE__ respectively. These builtins are constant expressions.

When the builtins appear as part of a default function argument the invocation point is the location of the caller. When the builtins appear as part of a default member initializer, the invocation point is the location of the constructor or aggregate initialization used to create the object. Otherwise the invocation point is the same as the location of the builtin.

When the invocation point of __builtin_FUNCTION is not a function scope the empty string is returned.

ARM/AArch64 Language Extensions¶

X86/X86-64 Language Extensions¶

The X86 backend has these language extensions:

#define GS_RELATIVE __attribute__((address_space(256)))
int foo(int GS_RELATIVE *P) {
  return *P;
}

_foo:
        movl    4(%esp), %eax
        movl    %gs:(%eax), %eax
        ret

PowerPC Language Extensions¶

double __builtin_setrnd(int mode);

PowerPC Language Extensions¶

double __builtin_setrnd(int mode);

PowerPC Language Extensions¶

double __builtin_setrnd(int mode);

void __dcbf(const void* addr); /* Data Cache Block Flush */

int a = 1;
__builtin_dcbf (&a);

Vectorization and Interleaving¶

A vectorized loop performs multiple iterations of the original loop in parallel using vector instructions. The instruction set of the target processor determines which vector instructions are available and their vector widths. This restricts the types of loops that can be vectorized. The vectorizer automatically determines if the loop is safe and profitable to vectorize. A vector instruction cost model is used to select the vector width.

Interleaving multiple loop iterations allows modern processors to further improve instruction-level parallelism (ILP) using advanced hardware features, such as multiple execution units and out-of-order execution. The vectorizer uses a cost model that depends on the register pressure and generated code size to select the interleaving count.

Vectorization is enabled by vectorize(enable) and interleaving is enabled by interleave(enable). This is useful when compiling with -Os to manually enable vectorization or interleaving.

#pragma clang loop vectorize(enable)
#pragma clang loop interleave(enable)
for(...) {
  ...
}

The vector width is specified by vectorize_width(_value_) and the interleave count is specified by interleave_count(_value_), where _value_ is a positive integer. This is useful for specifying the optimal width/count of the set of target architectures supported by your application.

#pragma clang loop vectorize_width(2)
#pragma clang loop interleave_count(2)
for(...) {
  ...
}

Specifying a width/count of 1 disables the optimization, and is equivalent to vectorize(disable) or interleave(disable).

Loop Unrolling¶

Unrolling a loop reduces the loop control overhead and exposes more opportunities for ILP. Loops can be fully or partially unrolled. Full unrolling eliminates the loop and replaces it with an enumerated sequence of loop iterations. Full unrolling is only possible if the loop trip count is known at compile time. Partial unrolling replicates the loop body within the loop and reduces the trip count.

If unroll(enable) is specified the unroller will attempt to fully unroll the loop if the trip count is known at compile time. If the fully unrolled code size is greater than an internal limit the loop will be partially unrolled up to this limit. If the trip count is not known at compile time the loop will be partially unrolled with a heuristically chosen unroll factor.

#pragma clang loop unroll(enable)
for(...) {
  ...
}

If unroll(full) is specified the unroller will attempt to fully unroll the loop if the trip count is known at compile time identically to unroll(enable). However, with unroll(full) the loop will not be unrolled if the loop count is not known at compile time.

#pragma clang loop unroll(full)
for(...) {
  ...
}

The unroll count can be specified explicitly with unroll_count(_value_) where _value_ is a positive integer. If this value is greater than the trip count the loop will be fully unrolled. Otherwise the loop is partially unrolled subject to the same code size limit as with unroll(enable).

#pragma clang loop unroll_count(8)
for(...) {
  ...
}

Unrolling of a loop can be prevented by specifying unroll(disable).

Loop Distribution¶

Loop Distribution allows splitting a loop into multiple loops. This is beneficial for example when the entire loop cannot be vectorized but some of the resulting loops can.

If distribute(enable)) is specified and the loop has memory dependencies that inhibit vectorization, the compiler will attempt to isolate the offending operations into a new loop. This optimization is not enabled by default, only loops marked with the pragma are considered.

#pragma clang loop distribute(enable)
for (i = 0; i < N; ++i) {
  S1: A[i + 1] = A[i] + B[i];
  S2: C[i] = D[i] * E[i];
}

This loop will be split into two loops between statements S1 and S2. The second loop containing S2 will be vectorized.

Loop Distribution is currently not enabled by default in the optimizer because it can hurt performance in some cases. For example, instruction-level parallelism could be reduced by sequentializing the execution of the statements S1 and S2 above.

If Loop Distribution is turned on globally with -mllvm -enable-loop-distribution, specifying distribute(disable) can be used the disable it on a per-loop basis.

Additional Information¶

For convenience multiple loop hints can be specified on a single line.

#pragma clang loop vectorize_width(4) interleave_count(8)
for(...) {
  ...
}

If an optimization cannot be applied any hints that apply to it will be ignored. For example, the hint vectorize_width(4) is ignored if the loop is not proven safe to vectorize. To identify and diagnose optimization issues use -Rpass, -Rpass-missed, and -Rpass-analysis command line options. See the user guide for details.

Subject Match Rules¶

The set of declarations that receive a single attribute from the attribute stack depends on the subject match rules that were specified in the pragma. Subject match rules are specified after the attribute. The compiler expects an identifier that corresponds to the subject set specifier. The apply_to specifier is currently the only supported subject set specifier. It allows you to specify match rules that form a subset of the attribute’s allowed subject set, i.e. the compiler doesn’t require all of the attribute’s subjects. For example, an attribute like [[nodiscard]] whose subject set includes enum, record and hasType(functionType), requires the presence of at least one of these rules after apply_to:

#pragma clang attribute push([[nodiscard]], apply_to = enum)

enum Enum1 { A1, B1 }; // The enum will receive [[nodiscard]]

struct Record1 { }; // The struct will *not* receive [[nodiscard]]

#pragma clang attribute pop

#pragma clang attribute push([[nodiscard]], apply_to = any(record, enum))

enum Enum2 { A2, B2 }; // The enum will receive [[nodiscard]]

struct Record2 { }; // The struct *will* receive [[nodiscard]]

#pragma clang attribute pop

// This is an error, since [[nodiscard]] can't be applied to namespaces:
#pragma clang attribute push([[nodiscard]], apply_to = any(record, namespace))

#pragma clang attribute pop

Multiple match rules can be specified using the any match rule, as shown in the example above. The any rule applies attributes to all declarations that are matched by at least one of the rules in the any. It doesn’t nest and can’t be used inside the other match rules. Redundant match rules or rules that conflict with one another should not be used inside of any.

Clang supports the following match rules:

• function: Can be used to apply attributes to functions. This includes C++ member functions, static functions, operators, and constructors/destructors.
• function(is_member): Can be used to apply attributes to C++ member functions. This includes members like static functions, operators, and constructors/destructors.
• hasType(functionType): Can be used to apply attributes to functions, C++ member functions, and variables/fields whose type is a function pointer. It does not apply attributes to Objective-C methods or blocks.
• type_alias: Can be used to apply attributes to typedef declarations and C++11 type aliases.
• record: Can be used to apply attributes to struct, class, and union declarations.
• record(unless(is_union)): Can be used to apply attributes only to struct and class declarations.
• enum: Can be be used to apply attributes to enumeration declarations.
• enum_constant: Can be used to apply attributes to enumerators.
• variable: Can be used to apply attributes to variables, including local variables, parameters, global variables, and static member variables. It does not apply attributes to instance member variables or Objective-C ivars.
• variable(is_thread_local): Can be used to apply attributes to thread-local variables only.
• variable(is_global): Can be used to apply attributes to global variables only.
• variable(is_parameter): Can be used to apply attributes to parameters only.
• variable(unless(is_parameter)): Can be used to apply attributes to all the variables that are not parameters.
• field: Can be used to apply attributes to non-static member variables in a record. This includes Objective-C ivars.
• namespace: Can be used to apply attributes to namespace declarations.
• objc_interface: Can be used to apply attributes to @interface declarations.
• objc_protocol: Can be used to apply attributes to @protocol declarations.
• objc_category: Can be used to apply attributes to category declarations, including class extensions.
• objc_method: Can be used to apply attributes to Objective-C methods, including instance and class methods. Implicit methods like implicit property getters and setters do not receive the attribute.
• objc_method(is_instance): Can be used to apply attributes to Objective-C instance methods.
• objc_property: Can be used to apply attributes to @property declarations.
• block: Can be used to apply attributes to block declarations. This does not include variables/fields of block pointer type.

The use of unless in match rules is currently restricted to a strict set of sub-rules that are used by the supported attributes. That means that even though variable(unless(is_parameter)) is a valid match rule, variable(unless(is_thread_local)) is not.

Supported Attributes¶

Not all attributes can be used with the #pragma clang attribute directive. Notably, statement attributes like [[fallthrough]] or type attributes like address_space aren’t supported by this directive. You can determine whether or not an attribute is supported by the pragma by referring to the individual documentation for that attribute.

The attributes are applied to all matching declarations individually, even when the attribute is semantically incorrect. The attributes that aren’t applied to any declaration are not verified semantically.