1 @c Copyright (C) 1988,1989,1992,1993,1994,1996,1998,1999,2000,2001,2002, 2003
2 @c Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
7 @chapter C Implementation-defined behavior
8 @cindex implementation-defined behavior, C language
10 A conforming implementation of ISO C is required to document its
11 choice of behavior in each of the areas that are designated
12 ``implementation defined.'' The following lists all such areas,
13 along with the section number from the ISO/IEC 9899:1999 standard.
16 * Translation implementation::
17 * Environment implementation::
18 * Identifiers implementation::
19 * Characters implementation::
20 * Integers implementation::
21 * Floating point implementation::
22 * Arrays and pointers implementation::
23 * Hints implementation::
24 * Structures unions enumerations and bit-fields implementation::
25 * Qualifiers implementation::
26 * Preprocessing directives implementation::
27 * Library functions implementation::
28 * Architecture implementation::
29 * Locale-specific behavior implementation::
32 @node Translation implementation
37 @cite{How a diagnostic is identified (3.10, 5.1.1.3).}
39 Diagnostics consist of all the output sent to stderr by GCC.
42 @cite{Whether each nonempty sequence of white-space characters other than
43 new-line is retained or replaced by one space character in translation
47 @node Environment implementation
50 The behavior of these points are dependent on the implementation
51 of the C library, and are not defined by GCC itself.
53 @node Identifiers implementation
58 @cite{Which additional multibyte characters may appear in identifiers
59 and their correspondence to universal character names (6.4.2).}
62 @cite{The number of significant initial characters in an identifier
65 For internal names, all characters are significant. For external names,
66 the number of significant characters are defined by the linker; for
67 almost all targets, all characters are significant.
71 @node Characters implementation
76 @cite{The number of bits in a byte (3.6).}
79 @cite{The values of the members of the execution character set (5.2.1).}
82 @cite{The unique value of the member of the execution character set produced
83 for each of the standard alphabetic escape sequences (5.2.2).}
86 @cite{The value of a @code{char} object into which has been stored any
87 character other than a member of the basic execution character set (6.2.5).}
90 @cite{Which of @code{signed char} or @code{unsigned char} has the same range,
91 representation, and behavior as ``plain'' @code{char} (6.2.5, 6.3.1.1).}
94 @cite{The mapping of members of the source character set (in character
95 constants and string literals) to members of the execution character
96 set (6.4.4.4, 5.1.1.2).}
99 @cite{The value of an integer character constant containing more than one
100 character or containing a character or escape sequence that does not map
101 to a single-byte execution character (6.4.4.4).}
104 @cite{The value of a wide character constant containing more than one
105 multibyte character, or containing a multibyte character or escape
106 sequence not represented in the extended execution character set (6.4.4.4).}
109 @cite{The current locale used to convert a wide character constant consisting
110 of a single multibyte character that maps to a member of the extended
111 execution character set into a corresponding wide character code (6.4.4.4).}
114 @cite{The current locale used to convert a wide string literal into
115 corresponding wide character codes (6.4.5).}
118 @cite{The value of a string literal containing a multibyte character or escape
119 sequence not represented in the execution character set (6.4.5).}
122 @node Integers implementation
127 @cite{Any extended integer types that exist in the implementation (6.2.5).}
130 @cite{Whether signed integer types are represented using sign and magnitude,
131 two's complement, or one's complement, and whether the extraordinary value
132 is a trap representation or an ordinary value (6.2.6.2).}
134 GCC supports only two's complement integer types, and all bit patterns
138 @cite{The rank of any extended integer type relative to another extended
139 integer type with the same precision (6.3.1.1).}
142 @cite{The result of, or the signal raised by, converting an integer to a
143 signed integer type when the value cannot be represented in an object of
144 that type (6.3.1.3).}
147 @cite{The results of some bitwise operations on signed integers (6.5).}
150 @node Floating point implementation
151 @section Floating point
155 @cite{The accuracy of the floating-point operations and of the library
156 functions in @code{<math.h>} and @code{<complex.h>} that return floating-point
157 results (5.2.4.2.2).}
160 @cite{The rounding behaviors characterized by non-standard values
161 of @code{FLT_ROUNDS} @gol
165 @cite{The evaluation methods characterized by non-standard negative
166 values of @code{FLT_EVAL_METHOD} (5.2.4.2.2).}
169 @cite{The direction of rounding when an integer is converted to a
170 floating-point number that cannot exactly represent the original
174 @cite{The direction of rounding when a floating-point number is
175 converted to a narrower floating-point number (6.3.1.5).}
178 @cite{How the nearest representable value or the larger or smaller
179 representable value immediately adjacent to the nearest representable
180 value is chosen for certain floating constants (6.4.4.2).}
183 @cite{Whether and how floating expressions are contracted when not
184 disallowed by the @code{FP_CONTRACT} pragma (6.5).}
187 @cite{The default state for the @code{FENV_ACCESS} pragma (7.6.1).}
190 @cite{Additional floating-point exceptions, rounding modes, environments,
191 and classifications, and their macro names (7.6, 7.12).}
194 @cite{The default state for the @code{FP_CONTRACT} pragma (7.12.2).}
197 @cite{Whether the ``inexact'' floating-point exception can be raised
198 when the rounded result actually does equal the mathematical result
199 in an IEC 60559 conformant implementation (F.9).}
202 @cite{Whether the ``underflow'' (and ``inexact'') floating-point
203 exception can be raised when a result is tiny but not inexact in an
204 IEC 60559 conformant implementation (F.9).}
208 @node Arrays and pointers implementation
209 @section Arrays and pointers
213 @cite{The result of converting a pointer to an integer or
214 vice versa (6.3.2.3).}
216 A cast from pointer to integer discards most-significant bits if the
217 pointer representation is larger than the integer type,
218 sign-extends@footnote{Future versions of GCC may zero-extend, or use
219 a target-defined @code{ptr_extend} pattern. Do not rely on sign extension.}
220 if the pointer representation is smaller than the integer type, otherwise
221 the bits are unchanged.
222 @c ??? We've always claimed that pointers were unsigned entities.
223 @c Shouldn't we therefore be doing zero-extension? If so, the bug
224 @c is in convert_to_integer, where we call type_for_size and request
225 @c a signed integral type. On the other hand, it might be most useful
226 @c for the target if we extend according to POINTERS_EXTEND_UNSIGNED.
228 A cast from integer to pointer discards most-significant bits if the
229 pointer representation is smaller than the integer type, extends according
230 to the signedness of the integer type if the pointer representation
231 is larger than the integer type, otherwise the bits are unchanged.
233 When casting from pointer to integer and back again, the resulting
234 pointer must reference the same object as the original pointer, otherwise
235 the behavior is undefined. That is, one may not use integer arithmetic to
236 avoid the undefined behavior of pointer arithmetic as proscribed in 6.5.6/8.
239 @cite{The size of the result of subtracting two pointers to elements
240 of the same array (6.5.6).}
244 @node Hints implementation
249 @cite{The extent to which suggestions made by using the @code{register}
250 storage-class specifier are effective (6.7.1).}
252 The @code{register} specifier affects code generation only in these ways:
256 When used as part of the register variable extension, see
257 @ref{Explicit Reg Vars}.
260 When @option{-O0} is in use, the compiler allocates distinct stack
261 memory for all variables that do not have the @code{register}
262 storage-class specifier; if @code{register} is specified, the variable
263 may have a shorter lifespan than the code would indicate and may never
267 On some rare x86 targets, @code{setjmp} doesn't save the registers in
268 all circumstances. In those cases, GCC doesn't allocate any variables
269 in registers unless they are marked @code{register}.
274 @cite{The extent to which suggestions made by using the inline function
275 specifier are effective (6.7.4).}
277 GCC will not inline any functions if the @option{-fno-inline} option is
278 used or if @option{-O0} is used. Otherwise, GCC may still be unable to
279 inline a function for many reasons; the @option{-Winline} option may be
280 used to determine if a function has not been inlined and why not.
284 @node Structures unions enumerations and bit-fields implementation
285 @section Structures, unions, enumerations, and bit-fields
289 @cite{Whether a ``plain'' int bit-field is treated as a @code{signed int}
290 bit-field or as an @code{unsigned int} bit-field (6.7.2, 6.7.2.1).}
293 @cite{Allowable bit-field types other than @code{_Bool}, @code{signed int},
294 and @code{unsigned int} (6.7.2.1).}
297 @cite{Whether a bit-field can straddle a storage-unit boundary (6.7.2.1).}
300 @cite{The order of allocation of bit-fields within a unit (6.7.2.1).}
303 @cite{The alignment of non-bit-field members of structures (6.7.2.1).}
306 @cite{The integer type compatible with each enumerated type (6.7.2.2).}
310 @node Qualifiers implementation
315 @cite{What constitutes an access to an object that has volatile-qualified
320 @node Preprocessing directives implementation
321 @section Preprocessing directives
325 @cite{How sequences in both forms of header names are mapped to headers
326 or external source file names (6.4.7).}
329 @cite{Whether the value of a character constant in a constant expression
330 that controls conditional inclusion matches the value of the same character
331 constant in the execution character set (6.10.1).}
334 @cite{Whether the value of a single-character character constant in a
335 constant expression that controls conditional inclusion may have a
336 negative value (6.10.1).}
339 @cite{The places that are searched for an included @samp{<>} delimited
340 header, and how the places are specified or the header is
341 identified (6.10.2).}
344 @cite{How the named source file is searched for in an included @samp{""}
345 delimited header (6.10.2).}
348 @cite{The method by which preprocessing tokens (possibly resulting from
349 macro expansion) in a @code{#include} directive are combined into a header
353 @cite{The nesting limit for @code{#include} processing (6.10.2).}
355 GCC imposes a limit of 200 nested @code{#include}s.
358 @cite{Whether the @samp{#} operator inserts a @samp{\} character before
359 the @samp{\} character that begins a universal character name in a
360 character constant or string literal (6.10.3.2).}
363 @cite{The behavior on each recognized non-@code{STDC #pragma}
367 @cite{The definitions for @code{__DATE__} and @code{__TIME__} when
368 respectively, the date and time of translation are not available (6.10.8).}
370 If the date and time are not available, @code{__DATE__} expands to
371 @code{@w{"??? ?? ????"}} and @code{__TIME__} expands to
376 @node Library functions implementation
377 @section Library functions
379 The behavior of these points are dependent on the implementation
380 of the C library, and are not defined by GCC itself.
382 @node Architecture implementation
383 @section Architecture
387 @cite{The values or expressions assigned to the macros specified in the
388 headers @code{<float.h>}, @code{<limits.h>}, and @code{<stdint.h>}
389 (5.2.4.2, 7.18.2, 7.18.3).}
392 @cite{The number, order, and encoding of bytes in any object
393 (when not explicitly specified in this International Standard) (6.2.6.1).}
396 @cite{The value of the result of the sizeof operator (6.5.3.4).}
400 @node Locale-specific behavior implementation
401 @section Locale-specific behavior
403 The behavior of these points are dependent on the implementation
404 of the C library, and are not defined by GCC itself.
407 @chapter Extensions to the C Language Family
408 @cindex extensions, C language
409 @cindex C language extensions
412 GNU C provides several language features not found in ISO standard C@.
413 (The @option{-pedantic} option directs GCC to print a warning message if
414 any of these features is used.) To test for the availability of these
415 features in conditional compilation, check for a predefined macro
416 @code{__GNUC__}, which is always defined under GCC@.
418 These extensions are available in C and Objective-C@. Most of them are
419 also available in C++. @xref{C++ Extensions,,Extensions to the
420 C++ Language}, for extensions that apply @emph{only} to C++.
422 Some features that are in ISO C99 but not C89 or C++ are also, as
423 extensions, accepted by GCC in C89 mode and in C++.
426 * Statement Exprs:: Putting statements and declarations inside expressions.
427 * Local Labels:: Labels local to a statement-expression.
428 * Labels as Values:: Getting pointers to labels, and computed gotos.
429 * Nested Functions:: As in Algol and Pascal, lexical scoping of functions.
430 * Constructing Calls:: Dispatching a call to another function.
431 * Typeof:: @code{typeof}: referring to the type of an expression.
432 * Lvalues:: Using @samp{?:}, @samp{,} and casts in lvalues.
433 * Conditionals:: Omitting the middle operand of a @samp{?:} expression.
434 * Long Long:: Double-word integers---@code{long long int}.
435 * Complex:: Data types for complex numbers.
436 * Hex Floats:: Hexadecimal floating-point constants.
437 * Zero Length:: Zero-length arrays.
438 * Variable Length:: Arrays whose length is computed at run time.
439 * Empty Structures:: Structures with no members.
440 * Variadic Macros:: Macros with a variable number of arguments.
441 * Escaped Newlines:: Slightly looser rules for escaped newlines.
442 * Multi-line Strings:: String literals with embedded newlines.
443 * Subscripting:: Any array can be subscripted, even if not an lvalue.
444 * Pointer Arith:: Arithmetic on @code{void}-pointers and function pointers.
445 * Initializers:: Non-constant initializers.
446 * Compound Literals:: Compound literals give structures, unions
448 * Designated Inits:: Labeling elements of initializers.
449 * Cast to Union:: Casting to union type from any member of the union.
450 * Case Ranges:: `case 1 ... 9' and such.
451 * Mixed Declarations:: Mixing declarations and code.
452 * Function Attributes:: Declaring that functions have no side effects,
453 or that they can never return.
454 * Attribute Syntax:: Formal syntax for attributes.
455 * Function Prototypes:: Prototype declarations and old-style definitions.
456 * C++ Comments:: C++ comments are recognized.
457 * Dollar Signs:: Dollar sign is allowed in identifiers.
458 * Character Escapes:: @samp{\e} stands for the character @key{ESC}.
459 * Variable Attributes:: Specifying attributes of variables.
460 * Type Attributes:: Specifying attributes of types.
461 * Alignment:: Inquiring about the alignment of a type or variable.
462 * Inline:: Defining inline functions (as fast as macros).
463 * Extended Asm:: Assembler instructions with C expressions as operands.
464 (With them you can define ``built-in'' functions.)
465 * Constraints:: Constraints for asm operands
466 * Asm Labels:: Specifying the assembler name to use for a C symbol.
467 * Explicit Reg Vars:: Defining variables residing in specified registers.
468 * Alternate Keywords:: @code{__const__}, @code{__asm__}, etc., for header files.
469 * Incomplete Enums:: @code{enum foo;}, with details to follow.
470 * Function Names:: Printable strings which are the name of the current
472 * Return Address:: Getting the return or frame address of a function.
473 * Vector Extensions:: Using vector instructions through built-in functions.
474 * Other Builtins:: Other built-in functions.
475 * Target Builtins:: Built-in functions specific to particular targets.
476 * Pragmas:: Pragmas accepted by GCC.
477 * Unnamed Fields:: Unnamed struct/union fields within structs/unions.
478 * Thread-Local:: Per-thread variables.
481 @node Statement Exprs
482 @section Statements and Declarations in Expressions
483 @cindex statements inside expressions
484 @cindex declarations inside expressions
485 @cindex expressions containing statements
486 @cindex macros, statements in expressions
488 @c the above section title wrapped and causes an underfull hbox.. i
489 @c changed it from "within" to "in". --mew 4feb93
490 A compound statement enclosed in parentheses may appear as an expression
491 in GNU C@. This allows you to use loops, switches, and local variables
492 within an expression.
494 Recall that a compound statement is a sequence of statements surrounded
495 by braces; in this construct, parentheses go around the braces. For
499 (@{ int y = foo (); int z;
506 is a valid (though slightly more complex than necessary) expression
507 for the absolute value of @code{foo ()}.
509 The last thing in the compound statement should be an expression
510 followed by a semicolon; the value of this subexpression serves as the
511 value of the entire construct. (If you use some other kind of statement
512 last within the braces, the construct has type @code{void}, and thus
513 effectively no value.)
515 This feature is especially useful in making macro definitions ``safe'' (so
516 that they evaluate each operand exactly once). For example, the
517 ``maximum'' function is commonly defined as a macro in standard C as
521 #define max(a,b) ((a) > (b) ? (a) : (b))
525 @cindex side effects, macro argument
526 But this definition computes either @var{a} or @var{b} twice, with bad
527 results if the operand has side effects. In GNU C, if you know the
528 type of the operands (here let's assume @code{int}), you can define
529 the macro safely as follows:
532 #define maxint(a,b) \
533 (@{int _a = (a), _b = (b); _a > _b ? _a : _b; @})
536 Embedded statements are not allowed in constant expressions, such as
537 the value of an enumeration constant, the width of a bit-field, or
538 the initial value of a static variable.
540 If you don't know the type of the operand, you can still do this, but you
541 must use @code{typeof} (@pxref{Typeof}).
543 Statement expressions are not supported fully in G++, and their fate
544 there is unclear. (It is possible that they will become fully supported
545 at some point, or that they will be deprecated, or that the bugs that
546 are present will continue to exist indefinitely.) Presently, statement
547 expressions do not work well as default arguments.
549 In addition, there are semantic issues with statement-expressions in
550 C++. If you try to use statement-expressions instead of inline
551 functions in C++, you may be surprised at the way object destruction is
552 handled. For example:
555 #define foo(a) (@{int b = (a); b + 3; @})
559 does not work the same way as:
562 inline int foo(int a) @{ int b = a; return b + 3; @}
566 In particular, if the expression passed into @code{foo} involves the
567 creation of temporaries, the destructors for those temporaries will be
568 run earlier in the case of the macro than in the case of the function.
570 These considerations mean that it is probably a bad idea to use
571 statement-expressions of this form in header files that are designed to
572 work with C++. (Note that some versions of the GNU C Library contained
573 header files using statement-expression that lead to precisely this
577 @section Locally Declared Labels
579 @cindex macros, local labels
581 Each statement expression is a scope in which @dfn{local labels} can be
582 declared. A local label is simply an identifier; you can jump to it
583 with an ordinary @code{goto} statement, but only from within the
584 statement expression it belongs to.
586 A local label declaration looks like this:
589 __label__ @var{label};
596 __label__ @var{label1}, @var{label2}, /* @r{@dots{}} */;
599 Local label declarations must come at the beginning of the statement
600 expression, right after the @samp{(@{}, before any ordinary
603 The label declaration defines the label @emph{name}, but does not define
604 the label itself. You must do this in the usual way, with
605 @code{@var{label}:}, within the statements of the statement expression.
607 The local label feature is useful because statement expressions are
608 often used in macros. If the macro contains nested loops, a @code{goto}
609 can be useful for breaking out of them. However, an ordinary label
610 whose scope is the whole function cannot be used: if the macro can be
611 expanded several times in one function, the label will be multiply
612 defined in that function. A local label avoids this problem. For
616 #define SEARCH(array, target) \
619 typeof (target) _SEARCH_target = (target); \
620 typeof (*(array)) *_SEARCH_array = (array); \
623 for (i = 0; i < max; i++) \
624 for (j = 0; j < max; j++) \
625 if (_SEARCH_array[i][j] == _SEARCH_target) \
626 @{ value = i; goto found; @} \
633 @node Labels as Values
634 @section Labels as Values
635 @cindex labels as values
636 @cindex computed gotos
637 @cindex goto with computed label
638 @cindex address of a label
640 You can get the address of a label defined in the current function
641 (or a containing function) with the unary operator @samp{&&}. The
642 value has type @code{void *}. This value is a constant and can be used
643 wherever a constant of that type is valid. For example:
651 To use these values, you need to be able to jump to one. This is done
652 with the computed goto statement@footnote{The analogous feature in
653 Fortran is called an assigned goto, but that name seems inappropriate in
654 C, where one can do more than simply store label addresses in label
655 variables.}, @code{goto *@var{exp};}. For example,
662 Any expression of type @code{void *} is allowed.
664 One way of using these constants is in initializing a static array that
665 will serve as a jump table:
668 static void *array[] = @{ &&foo, &&bar, &&hack @};
671 Then you can select a label with indexing, like this:
678 Note that this does not check whether the subscript is in bounds---array
679 indexing in C never does that.
681 Such an array of label values serves a purpose much like that of the
682 @code{switch} statement. The @code{switch} statement is cleaner, so
683 use that rather than an array unless the problem does not fit a
684 @code{switch} statement very well.
686 Another use of label values is in an interpreter for threaded code.
687 The labels within the interpreter function can be stored in the
688 threaded code for super-fast dispatching.
690 You may not use this mechanism to jump to code in a different function.
691 If you do that, totally unpredictable things will happen. The best way to
692 avoid this is to store the label address only in automatic variables and
693 never pass it as an argument.
695 An alternate way to write the above example is
698 static const int array[] = @{ &&foo - &&foo, &&bar - &&foo,
700 goto *(&&foo + array[i]);
704 This is more friendly to code living in shared libraries, as it reduces
705 the number of dynamic relocations that are needed, and by consequence,
706 allows the data to be read-only.
708 @node Nested Functions
709 @section Nested Functions
710 @cindex nested functions
711 @cindex downward funargs
714 A @dfn{nested function} is a function defined inside another function.
715 (Nested functions are not supported for GNU C++.) The nested function's
716 name is local to the block where it is defined. For example, here we
717 define a nested function named @code{square}, and call it twice:
721 foo (double a, double b)
723 double square (double z) @{ return z * z; @}
725 return square (a) + square (b);
730 The nested function can access all the variables of the containing
731 function that are visible at the point of its definition. This is
732 called @dfn{lexical scoping}. For example, here we show a nested
733 function which uses an inherited variable named @code{offset}:
737 bar (int *array, int offset, int size)
739 int access (int *array, int index)
740 @{ return array[index + offset]; @}
743 for (i = 0; i < size; i++)
744 /* @r{@dots{}} */ access (array, i) /* @r{@dots{}} */
749 Nested function definitions are permitted within functions in the places
750 where variable definitions are allowed; that is, in any block, before
751 the first statement in the block.
753 It is possible to call the nested function from outside the scope of its
754 name by storing its address or passing the address to another function:
757 hack (int *array, int size)
759 void store (int index, int value)
760 @{ array[index] = value; @}
762 intermediate (store, size);
766 Here, the function @code{intermediate} receives the address of
767 @code{store} as an argument. If @code{intermediate} calls @code{store},
768 the arguments given to @code{store} are used to store into @code{array}.
769 But this technique works only so long as the containing function
770 (@code{hack}, in this example) does not exit.
772 If you try to call the nested function through its address after the
773 containing function has exited, all hell will break loose. If you try
774 to call it after a containing scope level has exited, and if it refers
775 to some of the variables that are no longer in scope, you may be lucky,
776 but it's not wise to take the risk. If, however, the nested function
777 does not refer to anything that has gone out of scope, you should be
780 GCC implements taking the address of a nested function using a technique
781 called @dfn{trampolines}. A paper describing them is available as
784 @uref{http://people.debian.org/~aaronl/Usenix88-lexic.pdf}.
786 A nested function can jump to a label inherited from a containing
787 function, provided the label was explicitly declared in the containing
788 function (@pxref{Local Labels}). Such a jump returns instantly to the
789 containing function, exiting the nested function which did the
790 @code{goto} and any intermediate functions as well. Here is an example:
794 bar (int *array, int offset, int size)
797 int access (int *array, int index)
801 return array[index + offset];
805 for (i = 0; i < size; i++)
806 /* @r{@dots{}} */ access (array, i) /* @r{@dots{}} */
810 /* @r{Control comes here from @code{access}
811 if it detects an error.} */
818 A nested function always has internal linkage. Declaring one with
819 @code{extern} is erroneous. If you need to declare the nested function
820 before its definition, use @code{auto} (which is otherwise meaningless
821 for function declarations).
824 bar (int *array, int offset, int size)
827 auto int access (int *, int);
829 int access (int *array, int index)
833 return array[index + offset];
839 @node Constructing Calls
840 @section Constructing Function Calls
841 @cindex constructing calls
842 @cindex forwarding calls
844 Using the built-in functions described below, you can record
845 the arguments a function received, and call another function
846 with the same arguments, without knowing the number or types
849 You can also record the return value of that function call,
850 and later return that value, without knowing what data type
851 the function tried to return (as long as your caller expects
854 @deftypefn {Built-in Function} {void *} __builtin_apply_args ()
855 This built-in function returns a pointer to data
856 describing how to perform a call with the same arguments as were passed
857 to the current function.
859 The function saves the arg pointer register, structure value address,
860 and all registers that might be used to pass arguments to a function
861 into a block of memory allocated on the stack. Then it returns the
862 address of that block.
865 @deftypefn {Built-in Function} {void *} __builtin_apply (void (*@var{function})(), void *@var{arguments}, size_t @var{size})
866 This built-in function invokes @var{function}
867 with a copy of the parameters described by @var{arguments}
870 The value of @var{arguments} should be the value returned by
871 @code{__builtin_apply_args}. The argument @var{size} specifies the size
872 of the stack argument data, in bytes.
874 This function returns a pointer to data describing
875 how to return whatever value was returned by @var{function}. The data
876 is saved in a block of memory allocated on the stack.
878 It is not always simple to compute the proper value for @var{size}. The
879 value is used by @code{__builtin_apply} to compute the amount of data
880 that should be pushed on the stack and copied from the incoming argument
884 @deftypefn {Built-in Function} {void} __builtin_return (void *@var{result})
885 This built-in function returns the value described by @var{result} from
886 the containing function. You should specify, for @var{result}, a value
887 returned by @code{__builtin_apply}.
891 @section Referring to a Type with @code{typeof}
894 @cindex macros, types of arguments
896 Another way to refer to the type of an expression is with @code{typeof}.
897 The syntax of using of this keyword looks like @code{sizeof}, but the
898 construct acts semantically like a type name defined with @code{typedef}.
900 There are two ways of writing the argument to @code{typeof}: with an
901 expression or with a type. Here is an example with an expression:
908 This assumes that @code{x} is an array of pointers to functions;
909 the type described is that of the values of the functions.
911 Here is an example with a typename as the argument:
918 Here the type described is that of pointers to @code{int}.
920 If you are writing a header file that must work when included in ISO C
921 programs, write @code{__typeof__} instead of @code{typeof}.
922 @xref{Alternate Keywords}.
924 A @code{typeof}-construct can be used anywhere a typedef name could be
925 used. For example, you can use it in a declaration, in a cast, or inside
926 of @code{sizeof} or @code{typeof}.
928 @code{typeof} is often useful in conjunction with the
929 statements-within-expressions feature. Here is how the two together can
930 be used to define a safe ``maximum'' macro that operates on any
931 arithmetic type and evaluates each of its arguments exactly once:
935 (@{ typeof (a) _a = (a); \
936 typeof (b) _b = (b); \
937 _a > _b ? _a : _b; @})
940 @cindex underscores in variables in macros
941 @cindex @samp{_} in variables in macros
942 @cindex local variables in macros
943 @cindex variables, local, in macros
944 @cindex macros, local variables in
946 The reason for using names that start with underscores for the local
947 variables is to avoid conflicts with variable names that occur within the
948 expressions that are substituted for @code{a} and @code{b}. Eventually we
949 hope to design a new form of declaration syntax that allows you to declare
950 variables whose scopes start only after their initializers; this will be a
951 more reliable way to prevent such conflicts.
954 Some more examples of the use of @code{typeof}:
958 This declares @code{y} with the type of what @code{x} points to.
965 This declares @code{y} as an array of such values.
972 This declares @code{y} as an array of pointers to characters:
975 typeof (typeof (char *)[4]) y;
979 It is equivalent to the following traditional C declaration:
985 To see the meaning of the declaration using @code{typeof}, and why it
986 might be a useful way to write, let's rewrite it with these macros:
989 #define pointer(T) typeof(T *)
990 #define array(T, N) typeof(T [N])
994 Now the declaration can be rewritten this way:
997 array (pointer (char), 4) y;
1001 Thus, @code{array (pointer (char), 4)} is the type of arrays of 4
1002 pointers to @code{char}.
1005 @emph{Compatibility Note:} In addition to @code{typeof}, GCC 2 supported
1006 a more limited extension which permitted one to write
1009 typedef @var{T} = @var{expr};
1013 with the effect of declaring @var{T} to have the type of the expression
1014 @var{expr}. This extension does not work with GCC 3 (versions between
1015 3.0 and 3.2 will crash; 3.2.1 and later give an error). Code which
1016 relies on it should be rewritten to use @code{typeof}:
1019 typedef typeof(@var{expr}) @var{T};
1023 This will work with all versions of GCC@.
1026 @section Generalized Lvalues
1027 @cindex compound expressions as lvalues
1028 @cindex expressions, compound, as lvalues
1029 @cindex conditional expressions as lvalues
1030 @cindex expressions, conditional, as lvalues
1031 @cindex casts as lvalues
1032 @cindex generalized lvalues
1033 @cindex lvalues, generalized
1034 @cindex extensions, @code{?:}
1035 @cindex @code{?:} extensions
1037 Compound expressions, conditional expressions and casts are allowed as
1038 lvalues provided their operands are lvalues. This means that you can take
1039 their addresses or store values into them.
1041 Standard C++ allows compound expressions and conditional expressions as
1042 lvalues, and permits casts to reference type, so use of this extension
1043 is deprecated for C++ code.
1045 For example, a compound expression can be assigned, provided the last
1046 expression in the sequence is an lvalue. These two expressions are
1054 Similarly, the address of the compound expression can be taken. These two
1055 expressions are equivalent:
1062 A conditional expression is a valid lvalue if its type is not void and the
1063 true and false branches are both valid lvalues. For example, these two
1064 expressions are equivalent:
1068 (a ? b = 5 : (c = 5))
1071 A cast is a valid lvalue if its operand is an lvalue. A simple
1072 assignment whose left-hand side is a cast works by converting the
1073 right-hand side first to the specified type, then to the type of the
1074 inner left-hand side expression. After this is stored, the value is
1075 converted back to the specified type to become the value of the
1076 assignment. Thus, if @code{a} has type @code{char *}, the following two
1077 expressions are equivalent:
1081 (int)(a = (char *)(int)5)
1084 An assignment-with-arithmetic operation such as @samp{+=} applied to a cast
1085 performs the arithmetic using the type resulting from the cast, and then
1086 continues as in the previous case. Therefore, these two expressions are
1091 (int)(a = (char *)(int) ((int)a + 5))
1094 You cannot take the address of an lvalue cast, because the use of its
1095 address would not work out coherently. Suppose that @code{&(int)f} were
1096 permitted, where @code{f} has type @code{float}. Then the following
1097 statement would try to store an integer bit-pattern where a floating
1098 point number belongs:
1104 This is quite different from what @code{(int)f = 1} would do---that
1105 would convert 1 to floating point and store it. Rather than cause this
1106 inconsistency, we think it is better to prohibit use of @samp{&} on a cast.
1108 If you really do want an @code{int *} pointer with the address of
1109 @code{f}, you can simply write @code{(int *)&f}.
1112 @section Conditionals with Omitted Operands
1113 @cindex conditional expressions, extensions
1114 @cindex omitted middle-operands
1115 @cindex middle-operands, omitted
1116 @cindex extensions, @code{?:}
1117 @cindex @code{?:} extensions
1119 The middle operand in a conditional expression may be omitted. Then
1120 if the first operand is nonzero, its value is the value of the conditional
1123 Therefore, the expression
1130 has the value of @code{x} if that is nonzero; otherwise, the value of
1133 This example is perfectly equivalent to
1139 @cindex side effect in ?:
1140 @cindex ?: side effect
1142 In this simple case, the ability to omit the middle operand is not
1143 especially useful. When it becomes useful is when the first operand does,
1144 or may (if it is a macro argument), contain a side effect. Then repeating
1145 the operand in the middle would perform the side effect twice. Omitting
1146 the middle operand uses the value already computed without the undesirable
1147 effects of recomputing it.
1150 @section Double-Word Integers
1151 @cindex @code{long long} data types
1152 @cindex double-word arithmetic
1153 @cindex multiprecision arithmetic
1154 @cindex @code{LL} integer suffix
1155 @cindex @code{ULL} integer suffix
1157 ISO C99 supports data types for integers that are at least 64 bits wide,
1158 and as an extension GCC supports them in C89 mode and in C++.
1159 Simply write @code{long long int} for a signed integer, or
1160 @code{unsigned long long int} for an unsigned integer. To make an
1161 integer constant of type @code{long long int}, add the suffix @samp{LL}
1162 to the integer. To make an integer constant of type @code{unsigned long
1163 long int}, add the suffix @samp{ULL} to the integer.
1165 You can use these types in arithmetic like any other integer types.
1166 Addition, subtraction, and bitwise boolean operations on these types
1167 are open-coded on all types of machines. Multiplication is open-coded
1168 if the machine supports fullword-to-doubleword a widening multiply
1169 instruction. Division and shifts are open-coded only on machines that
1170 provide special support. The operations that are not open-coded use
1171 special library routines that come with GCC@.
1173 There may be pitfalls when you use @code{long long} types for function
1174 arguments, unless you declare function prototypes. If a function
1175 expects type @code{int} for its argument, and you pass a value of type
1176 @code{long long int}, confusion will result because the caller and the
1177 subroutine will disagree about the number of bytes for the argument.
1178 Likewise, if the function expects @code{long long int} and you pass
1179 @code{int}. The best way to avoid such problems is to use prototypes.
1182 @section Complex Numbers
1183 @cindex complex numbers
1184 @cindex @code{_Complex} keyword
1185 @cindex @code{__complex__} keyword
1187 ISO C99 supports complex floating data types, and as an extension GCC
1188 supports them in C89 mode and in C++, and supports complex integer data
1189 types which are not part of ISO C99. You can declare complex types
1190 using the keyword @code{_Complex}. As an extension, the older GNU
1191 keyword @code{__complex__} is also supported.
1193 For example, @samp{_Complex double x;} declares @code{x} as a
1194 variable whose real part and imaginary part are both of type
1195 @code{double}. @samp{_Complex short int y;} declares @code{y} to
1196 have real and imaginary parts of type @code{short int}; this is not
1197 likely to be useful, but it shows that the set of complex types is
1200 To write a constant with a complex data type, use the suffix @samp{i} or
1201 @samp{j} (either one; they are equivalent). For example, @code{2.5fi}
1202 has type @code{_Complex float} and @code{3i} has type
1203 @code{_Complex int}. Such a constant always has a pure imaginary
1204 value, but you can form any complex value you like by adding one to a
1205 real constant. This is a GNU extension; if you have an ISO C99
1206 conforming C library (such as GNU libc), and want to construct complex
1207 constants of floating type, you should include @code{<complex.h>} and
1208 use the macros @code{I} or @code{_Complex_I} instead.
1210 @cindex @code{__real__} keyword
1211 @cindex @code{__imag__} keyword
1212 To extract the real part of a complex-valued expression @var{exp}, write
1213 @code{__real__ @var{exp}}. Likewise, use @code{__imag__} to
1214 extract the imaginary part. This is a GNU extension; for values of
1215 floating type, you should use the ISO C99 functions @code{crealf},
1216 @code{creal}, @code{creall}, @code{cimagf}, @code{cimag} and
1217 @code{cimagl}, declared in @code{<complex.h>} and also provided as
1218 built-in functions by GCC@.
1220 @cindex complex conjugation
1221 The operator @samp{~} performs complex conjugation when used on a value
1222 with a complex type. This is a GNU extension; for values of
1223 floating type, you should use the ISO C99 functions @code{conjf},
1224 @code{conj} and @code{conjl}, declared in @code{<complex.h>} and also
1225 provided as built-in functions by GCC@.
1227 GCC can allocate complex automatic variables in a noncontiguous
1228 fashion; it's even possible for the real part to be in a register while
1229 the imaginary part is on the stack (or vice-versa). Only the DWARF2
1230 debug info format can represent this, so use of DWARF2 is recommended.
1231 If you are using the stabs debug info format, GCC describes a noncontiguous
1232 complex variable as if it were two separate variables of noncomplex type.
1233 If the variable's actual name is @code{foo}, the two fictitious
1234 variables are named @code{foo$real} and @code{foo$imag}. You can
1235 examine and set these two fictitious variables with your debugger.
1241 ISO C99 supports floating-point numbers written not only in the usual
1242 decimal notation, such as @code{1.55e1}, but also numbers such as
1243 @code{0x1.fp3} written in hexadecimal format. As a GNU extension, GCC
1244 supports this in C89 mode (except in some cases when strictly
1245 conforming) and in C++. In that format the
1246 @samp{0x} hex introducer and the @samp{p} or @samp{P} exponent field are
1247 mandatory. The exponent is a decimal number that indicates the power of
1248 2 by which the significant part will be multiplied. Thus @samp{0x1.f} is
1255 @samp{p3} multiplies it by 8, and the value of @code{0x1.fp3}
1256 is the same as @code{1.55e1}.
1258 Unlike for floating-point numbers in the decimal notation the exponent
1259 is always required in the hexadecimal notation. Otherwise the compiler
1260 would not be able to resolve the ambiguity of, e.g., @code{0x1.f}. This
1261 could mean @code{1.0f} or @code{1.9375} since @samp{f} is also the
1262 extension for floating-point constants of type @code{float}.
1265 @section Arrays of Length Zero
1266 @cindex arrays of length zero
1267 @cindex zero-length arrays
1268 @cindex length-zero arrays
1269 @cindex flexible array members
1271 Zero-length arrays are allowed in GNU C@. They are very useful as the
1272 last element of a structure which is really a header for a variable-length
1281 struct line *thisline = (struct line *)
1282 malloc (sizeof (struct line) + this_length);
1283 thisline->length = this_length;
1286 In ISO C90, you would have to give @code{contents} a length of 1, which
1287 means either you waste space or complicate the argument to @code{malloc}.
1289 In ISO C99, you would use a @dfn{flexible array member}, which is
1290 slightly different in syntax and semantics:
1294 Flexible array members are written as @code{contents[]} without
1298 Flexible array members have incomplete type, and so the @code{sizeof}
1299 operator may not be applied. As a quirk of the original implementation
1300 of zero-length arrays, @code{sizeof} evaluates to zero.
1303 Flexible array members may only appear as the last member of a
1304 @code{struct} that is otherwise non-empty.
1307 A structure containing a flexible array member, or a union containing
1308 such a structure (possibly recursively), may not be a member of a
1309 structure or an element of an array. (However, these uses are
1310 permitted by GCC as extensions.)
1313 GCC versions before 3.0 allowed zero-length arrays to be statically
1314 initialized, as if they were flexible arrays. In addition to those
1315 cases that were useful, it also allowed initializations in situations
1316 that would corrupt later data. Non-empty initialization of zero-length
1317 arrays is now treated like any case where there are more initializer
1318 elements than the array holds, in that a suitable warning about "excess
1319 elements in array" is given, and the excess elements (all of them, in
1320 this case) are ignored.
1322 Instead GCC allows static initialization of flexible array members.
1323 This is equivalent to defining a new structure containing the original
1324 structure followed by an array of sufficient size to contain the data.
1325 I.e.@: in the following, @code{f1} is constructed as if it were declared
1331 @} f1 = @{ 1, @{ 2, 3, 4 @} @};
1334 struct f1 f1; int data[3];
1335 @} f2 = @{ @{ 1 @}, @{ 2, 3, 4 @} @};
1339 The convenience of this extension is that @code{f1} has the desired
1340 type, eliminating the need to consistently refer to @code{f2.f1}.
1342 This has symmetry with normal static arrays, in that an array of
1343 unknown size is also written with @code{[]}.
1345 Of course, this extension only makes sense if the extra data comes at
1346 the end of a top-level object, as otherwise we would be overwriting
1347 data at subsequent offsets. To avoid undue complication and confusion
1348 with initialization of deeply nested arrays, we simply disallow any
1349 non-empty initialization except when the structure is the top-level
1350 object. For example:
1353 struct foo @{ int x; int y[]; @};
1354 struct bar @{ struct foo z; @};
1356 struct foo a = @{ 1, @{ 2, 3, 4 @} @}; // @r{Valid.}
1357 struct bar b = @{ @{ 1, @{ 2, 3, 4 @} @} @}; // @r{Invalid.}
1358 struct bar c = @{ @{ 1, @{ @} @} @}; // @r{Valid.}
1359 struct foo d[1] = @{ @{ 1 @{ 2, 3, 4 @} @} @}; // @r{Invalid.}
1362 @node Empty Structures
1363 @section Structures With No Members
1364 @cindex empty structures
1365 @cindex zero-size structures
1367 GCC permits a C structure to have no members:
1374 The structure will have size zero. In C++, empty structures are part
1375 of the language. G++ treats empty structures as if they had a single
1376 member of type @code{char}.
1378 @node Variable Length
1379 @section Arrays of Variable Length
1380 @cindex variable-length arrays
1381 @cindex arrays of variable length
1384 Variable-length automatic arrays are allowed in ISO C99, and as an
1385 extension GCC accepts them in C89 mode and in C++. (However, GCC's
1386 implementation of variable-length arrays does not yet conform in detail
1387 to the ISO C99 standard.) These arrays are
1388 declared like any other automatic arrays, but with a length that is not
1389 a constant expression. The storage is allocated at the point of
1390 declaration and deallocated when the brace-level is exited. For
1395 concat_fopen (char *s1, char *s2, char *mode)
1397 char str[strlen (s1) + strlen (s2) + 1];
1400 return fopen (str, mode);
1404 @cindex scope of a variable length array
1405 @cindex variable-length array scope
1406 @cindex deallocating variable length arrays
1407 Jumping or breaking out of the scope of the array name deallocates the
1408 storage. Jumping into the scope is not allowed; you get an error
1411 @cindex @code{alloca} vs variable-length arrays
1412 You can use the function @code{alloca} to get an effect much like
1413 variable-length arrays. The function @code{alloca} is available in
1414 many other C implementations (but not in all). On the other hand,
1415 variable-length arrays are more elegant.
1417 There are other differences between these two methods. Space allocated
1418 with @code{alloca} exists until the containing @emph{function} returns.
1419 The space for a variable-length array is deallocated as soon as the array
1420 name's scope ends. (If you use both variable-length arrays and
1421 @code{alloca} in the same function, deallocation of a variable-length array
1422 will also deallocate anything more recently allocated with @code{alloca}.)
1424 You can also use variable-length arrays as arguments to functions:
1428 tester (int len, char data[len][len])
1434 The length of an array is computed once when the storage is allocated
1435 and is remembered for the scope of the array in case you access it with
1438 If you want to pass the array first and the length afterward, you can
1439 use a forward declaration in the parameter list---another GNU extension.
1443 tester (int len; char data[len][len], int len)
1449 @cindex parameter forward declaration
1450 The @samp{int len} before the semicolon is a @dfn{parameter forward
1451 declaration}, and it serves the purpose of making the name @code{len}
1452 known when the declaration of @code{data} is parsed.
1454 You can write any number of such parameter forward declarations in the
1455 parameter list. They can be separated by commas or semicolons, but the
1456 last one must end with a semicolon, which is followed by the ``real''
1457 parameter declarations. Each forward declaration must match a ``real''
1458 declaration in parameter name and data type. ISO C99 does not support
1459 parameter forward declarations.
1461 @node Variadic Macros
1462 @section Macros with a Variable Number of Arguments.
1463 @cindex variable number of arguments
1464 @cindex macro with variable arguments
1465 @cindex rest argument (in macro)
1466 @cindex variadic macros
1468 In the ISO C standard of 1999, a macro can be declared to accept a
1469 variable number of arguments much as a function can. The syntax for
1470 defining the macro is similar to that of a function. Here is an
1474 #define debug(format, ...) fprintf (stderr, format, __VA_ARGS__)
1477 Here @samp{@dots{}} is a @dfn{variable argument}. In the invocation of
1478 such a macro, it represents the zero or more tokens until the closing
1479 parenthesis that ends the invocation, including any commas. This set of
1480 tokens replaces the identifier @code{__VA_ARGS__} in the macro body
1481 wherever it appears. See the CPP manual for more information.
1483 GCC has long supported variadic macros, and used a different syntax that
1484 allowed you to give a name to the variable arguments just like any other
1485 argument. Here is an example:
1488 #define debug(format, args...) fprintf (stderr, format, args)
1491 This is in all ways equivalent to the ISO C example above, but arguably
1492 more readable and descriptive.
1494 GNU CPP has two further variadic macro extensions, and permits them to
1495 be used with either of the above forms of macro definition.
1497 In standard C, you are not allowed to leave the variable argument out
1498 entirely; but you are allowed to pass an empty argument. For example,
1499 this invocation is invalid in ISO C, because there is no comma after
1506 GNU CPP permits you to completely omit the variable arguments in this
1507 way. In the above examples, the compiler would complain, though since
1508 the expansion of the macro still has the extra comma after the format
1511 To help solve this problem, CPP behaves specially for variable arguments
1512 used with the token paste operator, @samp{##}. If instead you write
1515 #define debug(format, ...) fprintf (stderr, format, ## __VA_ARGS__)
1518 and if the variable arguments are omitted or empty, the @samp{##}
1519 operator causes the preprocessor to remove the comma before it. If you
1520 do provide some variable arguments in your macro invocation, GNU CPP
1521 does not complain about the paste operation and instead places the
1522 variable arguments after the comma. Just like any other pasted macro
1523 argument, these arguments are not macro expanded.
1525 @node Escaped Newlines
1526 @section Slightly Looser Rules for Escaped Newlines
1527 @cindex escaped newlines
1528 @cindex newlines (escaped)
1530 Recently, the preprocessor has relaxed its treatment of escaped
1531 newlines. Previously, the newline had to immediately follow a
1532 backslash. The current implementation allows whitespace in the form of
1533 spaces, horizontal and vertical tabs, and form feeds between the
1534 backslash and the subsequent newline. The preprocessor issues a
1535 warning, but treats it as a valid escaped newline and combines the two
1536 lines to form a single logical line. This works within comments and
1537 tokens, including multi-line strings, as well as between tokens.
1538 Comments are @emph{not} treated as whitespace for the purposes of this
1539 relaxation, since they have not yet been replaced with spaces.
1541 @node Multi-line Strings
1542 @section String Literals with Embedded Newlines
1543 @cindex multi-line string literals
1545 As an extension, GNU CPP permits string literals to cross multiple lines
1546 without escaping the embedded newlines. Each embedded newline is
1547 replaced with a single @samp{\n} character in the resulting string
1548 literal, regardless of what form the newline took originally.
1550 CPP currently allows such strings in directives as well (other than the
1551 @samp{#include} family). This is deprecated and will eventually be
1555 @section Non-Lvalue Arrays May Have Subscripts
1556 @cindex subscripting
1557 @cindex arrays, non-lvalue
1559 @cindex subscripting and function values
1560 In ISO C99, arrays that are not lvalues still decay to pointers, and
1561 may be subscripted, although they may not be modified or used after
1562 the next sequence point and the unary @samp{&} operator may not be
1563 applied to them. As an extension, GCC allows such arrays to be
1564 subscripted in C89 mode, though otherwise they do not decay to
1565 pointers outside C99 mode. For example,
1566 this is valid in GNU C though not valid in C89:
1570 struct foo @{int a[4];@};
1576 return f().a[index];
1582 @section Arithmetic on @code{void}- and Function-Pointers
1583 @cindex void pointers, arithmetic
1584 @cindex void, size of pointer to
1585 @cindex function pointers, arithmetic
1586 @cindex function, size of pointer to
1588 In GNU C, addition and subtraction operations are supported on pointers to
1589 @code{void} and on pointers to functions. This is done by treating the
1590 size of a @code{void} or of a function as 1.
1592 A consequence of this is that @code{sizeof} is also allowed on @code{void}
1593 and on function types, and returns 1.
1595 @opindex Wpointer-arith
1596 The option @option{-Wpointer-arith} requests a warning if these extensions
1600 @section Non-Constant Initializers
1601 @cindex initializers, non-constant
1602 @cindex non-constant initializers
1604 As in standard C++ and ISO C99, the elements of an aggregate initializer for an
1605 automatic variable are not required to be constant expressions in GNU C@.
1606 Here is an example of an initializer with run-time varying elements:
1609 foo (float f, float g)
1611 float beat_freqs[2] = @{ f-g, f+g @};
1616 @node Compound Literals
1617 @section Compound Literals
1618 @cindex constructor expressions
1619 @cindex initializations in expressions
1620 @cindex structures, constructor expression
1621 @cindex expressions, constructor
1622 @cindex compound literals
1623 @c The GNU C name for what C99 calls compound literals was "constructor expressions".
1625 ISO C99 supports compound literals. A compound literal looks like
1626 a cast containing an initializer. Its value is an object of the
1627 type specified in the cast, containing the elements specified in
1628 the initializer; it is an lvalue. As an extension, GCC supports
1629 compound literals in C89 mode and in C++.
1631 Usually, the specified type is a structure. Assume that
1632 @code{struct foo} and @code{structure} are declared as shown:
1635 struct foo @{int a; char b[2];@} structure;
1639 Here is an example of constructing a @code{struct foo} with a compound literal:
1642 structure = ((struct foo) @{x + y, 'a', 0@});
1646 This is equivalent to writing the following:
1650 struct foo temp = @{x + y, 'a', 0@};
1655 You can also construct an array. If all the elements of the compound literal
1656 are (made up of) simple constant expressions, suitable for use in
1657 initializers of objects of static storage duration, then the compound
1658 literal can be coerced to a pointer to its first element and used in
1659 such an initializer, as shown here:
1662 char **foo = (char *[]) @{ "x", "y", "z" @};
1665 Compound literals for scalar types and union types are is
1666 also allowed, but then the compound literal is equivalent
1669 As a GNU extension, GCC allows initialization of objects with static storage
1670 duration by compound literals (which is not possible in ISO C99, because
1671 the initializer is not a constant).
1672 It is handled as if the object was initialized only with the bracket
1673 enclosed list if compound literal's and object types match.
1674 The initializer list of the compound literal must be constant.
1675 If the object being initialized has array type of unknown size, the size is
1676 determined by compound literal size.
1679 static struct foo x = (struct foo) @{1, 'a', 'b'@};
1680 static int y[] = (int []) @{1, 2, 3@};
1681 static int z[] = (int [3]) @{1@};
1685 The above lines are equivalent to the following:
1687 static struct foo x = @{1, 'a', 'b'@};
1688 static int y[] = @{1, 2, 3@};
1689 static int z[] = @{1, 0, 0@};
1692 @node Designated Inits
1693 @section Designated Initializers
1694 @cindex initializers with labeled elements
1695 @cindex labeled elements in initializers
1696 @cindex case labels in initializers
1697 @cindex designated initializers
1699 Standard C89 requires the elements of an initializer to appear in a fixed
1700 order, the same as the order of the elements in the array or structure
1703 In ISO C99 you can give the elements in any order, specifying the array
1704 indices or structure field names they apply to, and GNU C allows this as
1705 an extension in C89 mode as well. This extension is not
1706 implemented in GNU C++.
1708 To specify an array index, write
1709 @samp{[@var{index}] =} before the element value. For example,
1712 int a[6] = @{ [4] = 29, [2] = 15 @};
1719 int a[6] = @{ 0, 0, 15, 0, 29, 0 @};
1723 The index values must be constant expressions, even if the array being
1724 initialized is automatic.
1726 An alternative syntax for this which has been obsolete since GCC 2.5 but
1727 GCC still accepts is to write @samp{[@var{index}]} before the element
1728 value, with no @samp{=}.
1730 To initialize a range of elements to the same value, write
1731 @samp{[@var{first} ... @var{last}] = @var{value}}. This is a GNU
1732 extension. For example,
1735 int widths[] = @{ [0 ... 9] = 1, [10 ... 99] = 2, [100] = 3 @};
1739 If the value in it has side-effects, the side-effects will happen only once,
1740 not for each initialized field by the range initializer.
1743 Note that the length of the array is the highest value specified
1746 In a structure initializer, specify the name of a field to initialize
1747 with @samp{.@var{fieldname} =} before the element value. For example,
1748 given the following structure,
1751 struct point @{ int x, y; @};
1755 the following initialization
1758 struct point p = @{ .y = yvalue, .x = xvalue @};
1765 struct point p = @{ xvalue, yvalue @};
1768 Another syntax which has the same meaning, obsolete since GCC 2.5, is
1769 @samp{@var{fieldname}:}, as shown here:
1772 struct point p = @{ y: yvalue, x: xvalue @};
1776 The @samp{[@var{index}]} or @samp{.@var{fieldname}} is known as a
1777 @dfn{designator}. You can also use a designator (or the obsolete colon
1778 syntax) when initializing a union, to specify which element of the union
1779 should be used. For example,
1782 union foo @{ int i; double d; @};
1784 union foo f = @{ .d = 4 @};
1788 will convert 4 to a @code{double} to store it in the union using
1789 the second element. By contrast, casting 4 to type @code{union foo}
1790 would store it into the union as the integer @code{i}, since it is
1791 an integer. (@xref{Cast to Union}.)
1793 You can combine this technique of naming elements with ordinary C
1794 initialization of successive elements. Each initializer element that
1795 does not have a designator applies to the next consecutive element of the
1796 array or structure. For example,
1799 int a[6] = @{ [1] = v1, v2, [4] = v4 @};
1806 int a[6] = @{ 0, v1, v2, 0, v4, 0 @};
1809 Labeling the elements of an array initializer is especially useful
1810 when the indices are characters or belong to an @code{enum} type.
1815 = @{ [' '] = 1, ['\t'] = 1, ['\h'] = 1,
1816 ['\f'] = 1, ['\n'] = 1, ['\r'] = 1 @};
1819 @cindex designator lists
1820 You can also write a series of @samp{.@var{fieldname}} and
1821 @samp{[@var{index}]} designators before an @samp{=} to specify a
1822 nested subobject to initialize; the list is taken relative to the
1823 subobject corresponding to the closest surrounding brace pair. For
1824 example, with the @samp{struct point} declaration above:
1827 struct point ptarray[10] = @{ [2].y = yv2, [2].x = xv2, [0].x = xv0 @};
1831 If the same field is initialized multiple times, it will have value from
1832 the last initialization. If any such overridden initialization has
1833 side-effect, it is unspecified whether the side-effect happens or not.
1834 Currently, gcc will discard them and issue a warning.
1837 @section Case Ranges
1839 @cindex ranges in case statements
1841 You can specify a range of consecutive values in a single @code{case} label,
1845 case @var{low} ... @var{high}:
1849 This has the same effect as the proper number of individual @code{case}
1850 labels, one for each integer value from @var{low} to @var{high}, inclusive.
1852 This feature is especially useful for ranges of ASCII character codes:
1858 @strong{Be careful:} Write spaces around the @code{...}, for otherwise
1859 it may be parsed wrong when you use it with integer values. For example,
1874 @section Cast to a Union Type
1875 @cindex cast to a union
1876 @cindex union, casting to a
1878 A cast to union type is similar to other casts, except that the type
1879 specified is a union type. You can specify the type either with
1880 @code{union @var{tag}} or with a typedef name. A cast to union is actually
1881 a constructor though, not a cast, and hence does not yield an lvalue like
1882 normal casts. (@xref{Compound Literals}.)
1884 The types that may be cast to the union type are those of the members
1885 of the union. Thus, given the following union and variables:
1888 union foo @{ int i; double d; @};
1894 both @code{x} and @code{y} can be cast to type @code{union foo}.
1896 Using the cast as the right-hand side of an assignment to a variable of
1897 union type is equivalent to storing in a member of the union:
1902 u = (union foo) x @equiv{} u.i = x
1903 u = (union foo) y @equiv{} u.d = y
1906 You can also use the union cast as a function argument:
1909 void hack (union foo);
1911 hack ((union foo) x);
1914 @node Mixed Declarations
1915 @section Mixed Declarations and Code
1916 @cindex mixed declarations and code
1917 @cindex declarations, mixed with code
1918 @cindex code, mixed with declarations
1920 ISO C99 and ISO C++ allow declarations and code to be freely mixed
1921 within compound statements. As an extension, GCC also allows this in
1922 C89 mode. For example, you could do:
1931 Each identifier is visible from where it is declared until the end of
1932 the enclosing block.
1934 @node Function Attributes
1935 @section Declaring Attributes of Functions
1936 @cindex function attributes
1937 @cindex declaring attributes of functions
1938 @cindex functions that never return
1939 @cindex functions that have no side effects
1940 @cindex functions in arbitrary sections
1941 @cindex functions that behave like malloc
1942 @cindex @code{volatile} applied to function
1943 @cindex @code{const} applied to function
1944 @cindex functions with @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon} style arguments
1945 @cindex functions with non-null pointer arguments
1946 @cindex functions that are passed arguments in registers on the 386
1947 @cindex functions that pop the argument stack on the 386
1948 @cindex functions that do not pop the argument stack on the 386
1950 In GNU C, you declare certain things about functions called in your program
1951 which help the compiler optimize function calls and check your code more
1954 The keyword @code{__attribute__} allows you to specify special
1955 attributes when making a declaration. This keyword is followed by an
1956 attribute specification inside double parentheses. The following
1957 attributes are currently defined for functions on all targets:
1958 @code{noreturn}, @code{noinline}, @code{always_inline},
1959 @code{pure}, @code{const}, @code{nothrow},
1960 @code{format}, @code{format_arg}, @code{no_instrument_function},
1961 @code{section}, @code{constructor}, @code{destructor}, @code{used},
1962 @code{unused}, @code{deprecated}, @code{weak}, @code{malloc},
1963 @code{alias}, and @code{nonnull}. Several other attributes are defined
1964 for functions on particular target systems. Other attributes, including
1965 @code{section} are supported for variables declarations
1966 (@pxref{Variable Attributes}) and for types (@pxref{Type Attributes}).
1968 You may also specify attributes with @samp{__} preceding and following
1969 each keyword. This allows you to use them in header files without
1970 being concerned about a possible macro of the same name. For example,
1971 you may use @code{__noreturn__} instead of @code{noreturn}.
1973 @xref{Attribute Syntax}, for details of the exact syntax for using
1977 @cindex @code{noreturn} function attribute
1979 A few standard library functions, such as @code{abort} and @code{exit},
1980 cannot return. GCC knows this automatically. Some programs define
1981 their own functions that never return. You can declare them
1982 @code{noreturn} to tell the compiler this fact. For example,
1986 void fatal () __attribute__ ((noreturn));
1989 fatal (/* @r{@dots{}} */)
1991 /* @r{@dots{}} */ /* @r{Print error message.} */ /* @r{@dots{}} */
1997 The @code{noreturn} keyword tells the compiler to assume that
1998 @code{fatal} cannot return. It can then optimize without regard to what
1999 would happen if @code{fatal} ever did return. This makes slightly
2000 better code. More importantly, it helps avoid spurious warnings of
2001 uninitialized variables.
2003 Do not assume that registers saved by the calling function are
2004 restored before calling the @code{noreturn} function.
2006 It does not make sense for a @code{noreturn} function to have a return
2007 type other than @code{void}.
2009 The attribute @code{noreturn} is not implemented in GCC versions
2010 earlier than 2.5. An alternative way to declare that a function does
2011 not return, which works in the current version and in some older
2012 versions, is as follows:
2015 typedef void voidfn ();
2017 volatile voidfn fatal;
2020 @cindex @code{noinline} function attribute
2022 This function attribute prevents a function from being considered for
2025 @cindex @code{always_inline} function attribute
2027 Generally, functions are not inlined unless optimization is specified.
2028 For functions declared inline, this attribute inlines the function even
2029 if no optimization level was specified.
2031 @cindex @code{pure} function attribute
2033 Many functions have no effects except the return value and their
2034 return value depends only on the parameters and/or global variables.
2035 Such a function can be subject
2036 to common subexpression elimination and loop optimization just as an
2037 arithmetic operator would be. These functions should be declared
2038 with the attribute @code{pure}. For example,
2041 int square (int) __attribute__ ((pure));
2045 says that the hypothetical function @code{square} is safe to call
2046 fewer times than the program says.
2048 Some of common examples of pure functions are @code{strlen} or @code{memcmp}.
2049 Interesting non-pure functions are functions with infinite loops or those
2050 depending on volatile memory or other system resource, that may change between
2051 two consecutive calls (such as @code{feof} in a multithreading environment).
2053 The attribute @code{pure} is not implemented in GCC versions earlier
2055 @cindex @code{const} function attribute
2057 Many functions do not examine any values except their arguments, and
2058 have no effects except the return value. Basically this is just slightly
2059 more strict class than the @code{pure} attribute above, since function is not
2060 allowed to read global memory.
2062 @cindex pointer arguments
2063 Note that a function that has pointer arguments and examines the data
2064 pointed to must @emph{not} be declared @code{const}. Likewise, a
2065 function that calls a non-@code{const} function usually must not be
2066 @code{const}. It does not make sense for a @code{const} function to
2069 The attribute @code{const} is not implemented in GCC versions earlier
2070 than 2.5. An alternative way to declare that a function has no side
2071 effects, which works in the current version and in some older versions,
2075 typedef int intfn ();
2077 extern const intfn square;
2080 This approach does not work in GNU C++ from 2.6.0 on, since the language
2081 specifies that the @samp{const} must be attached to the return value.
2083 @cindex @code{nothrow} function attribute
2085 The @code{nothrow} attribute is used to inform the compiler that a
2086 function cannot throw an exception. For example, most functions in
2087 the standard C library can be guaranteed not to throw an exception
2088 with the notable exceptions of @code{qsort} and @code{bsearch} that
2089 take function pointer arguments. The @code{nothrow} attribute is not
2090 implemented in GCC versions earlier than 3.2.
2092 @item format (@var{archetype}, @var{string-index}, @var{first-to-check})
2093 @cindex @code{format} function attribute
2095 The @code{format} attribute specifies that a function takes @code{printf},
2096 @code{scanf}, @code{strftime} or @code{strfmon} style arguments which
2097 should be type-checked against a format string. For example, the
2102 my_printf (void *my_object, const char *my_format, ...)
2103 __attribute__ ((format (printf, 2, 3)));
2107 causes the compiler to check the arguments in calls to @code{my_printf}
2108 for consistency with the @code{printf} style format string argument
2111 The parameter @var{archetype} determines how the format string is
2112 interpreted, and should be @code{printf}, @code{scanf}, @code{strftime}
2113 or @code{strfmon}. (You can also use @code{__printf__},
2114 @code{__scanf__}, @code{__strftime__} or @code{__strfmon__}.) The
2115 parameter @var{string-index} specifies which argument is the format
2116 string argument (starting from 1), while @var{first-to-check} is the
2117 number of the first argument to check against the format string. For
2118 functions where the arguments are not available to be checked (such as
2119 @code{vprintf}), specify the third parameter as zero. In this case the
2120 compiler only checks the format string for consistency. For
2121 @code{strftime} formats, the third parameter is required to be zero.
2122 Since non-static C++ methods have an implicit @code{this} argument, the
2123 arguments of such methods should be counted from two, not one, when
2124 giving values for @var{string-index} and @var{first-to-check}.
2126 In the example above, the format string (@code{my_format}) is the second
2127 argument of the function @code{my_print}, and the arguments to check
2128 start with the third argument, so the correct parameters for the format
2129 attribute are 2 and 3.
2131 @opindex ffreestanding
2132 The @code{format} attribute allows you to identify your own functions
2133 which take format strings as arguments, so that GCC can check the
2134 calls to these functions for errors. The compiler always (unless
2135 @option{-ffreestanding} is used) checks formats
2136 for the standard library functions @code{printf}, @code{fprintf},
2137 @code{sprintf}, @code{scanf}, @code{fscanf}, @code{sscanf}, @code{strftime},
2138 @code{vprintf}, @code{vfprintf} and @code{vsprintf} whenever such
2139 warnings are requested (using @option{-Wformat}), so there is no need to
2140 modify the header file @file{stdio.h}. In C99 mode, the functions
2141 @code{snprintf}, @code{vsnprintf}, @code{vscanf}, @code{vfscanf} and
2142 @code{vsscanf} are also checked. Except in strictly conforming C
2143 standard modes, the X/Open function @code{strfmon} is also checked as
2144 are @code{printf_unlocked} and @code{fprintf_unlocked}.
2145 @xref{C Dialect Options,,Options Controlling C Dialect}.
2147 @item format_arg (@var{string-index})
2148 @cindex @code{format_arg} function attribute
2149 @opindex Wformat-nonliteral
2150 The @code{format_arg} attribute specifies that a function takes a format
2151 string for a @code{printf}, @code{scanf}, @code{strftime} or
2152 @code{strfmon} style function and modifies it (for example, to translate
2153 it into another language), so the result can be passed to a
2154 @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon} style
2155 function (with the remaining arguments to the format function the same
2156 as they would have been for the unmodified string). For example, the
2161 my_dgettext (char *my_domain, const char *my_format)
2162 __attribute__ ((format_arg (2)));
2166 causes the compiler to check the arguments in calls to a @code{printf},
2167 @code{scanf}, @code{strftime} or @code{strfmon} type function, whose
2168 format string argument is a call to the @code{my_dgettext} function, for
2169 consistency with the format string argument @code{my_format}. If the
2170 @code{format_arg} attribute had not been specified, all the compiler
2171 could tell in such calls to format functions would be that the format
2172 string argument is not constant; this would generate a warning when
2173 @option{-Wformat-nonliteral} is used, but the calls could not be checked
2174 without the attribute.
2176 The parameter @var{string-index} specifies which argument is the format
2177 string argument (starting from one). Since non-static C++ methods have
2178 an implicit @code{this} argument, the arguments of such methods should
2179 be counted from two.
2181 The @code{format-arg} attribute allows you to identify your own
2182 functions which modify format strings, so that GCC can check the
2183 calls to @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon}
2184 type function whose operands are a call to one of your own function.
2185 The compiler always treats @code{gettext}, @code{dgettext}, and
2186 @code{dcgettext} in this manner except when strict ISO C support is
2187 requested by @option{-ansi} or an appropriate @option{-std} option, or
2188 @option{-ffreestanding} is used. @xref{C Dialect Options,,Options
2189 Controlling C Dialect}.
2191 @item nonnull (@var{arg-index}, @dots{})
2192 @cindex @code{nonnull} function attribute
2193 The @code{nonnull} attribute specifies that some function parameters should
2194 be non-null pointers. For instance, the declaration:
2198 my_memcpy (void *dest, const void *src, size_t len)
2199 __attribute__((nonnull (1, 2)));
2203 causes the compiler to check that, in calls to @code{my_memcpy},
2204 arguments @var{dest} and @var{src} are non-null. If the compiler
2205 determines that a null pointer is passed in an argument slot marked
2206 as non-null, and the @option{-Wnonnull} option is enabled, a warning
2207 is issued. The compiler may also choose to make optimizations based
2208 on the knowledge that certain function arguments will not be null.
2210 If no argument index list is given to the @code{nonnull} attribute,
2211 all pointer arguments are marked as non-null. To illustrate, the
2212 following declaration is equivalent to the previous example:
2216 my_memcpy (void *dest, const void *src, size_t len)
2217 __attribute__((nonnull));
2220 @item no_instrument_function
2221 @cindex @code{no_instrument_function} function attribute
2222 @opindex finstrument-functions
2223 If @option{-finstrument-functions} is given, profiling function calls will
2224 be generated at entry and exit of most user-compiled functions.
2225 Functions with this attribute will not be so instrumented.
2227 @item section ("@var{section-name}")
2228 @cindex @code{section} function attribute
2229 Normally, the compiler places the code it generates in the @code{text} section.
2230 Sometimes, however, you need additional sections, or you need certain
2231 particular functions to appear in special sections. The @code{section}
2232 attribute specifies that a function lives in a particular section.
2233 For example, the declaration:
2236 extern void foobar (void) __attribute__ ((section ("bar")));
2240 puts the function @code{foobar} in the @code{bar} section.
2242 Some file formats do not support arbitrary sections so the @code{section}
2243 attribute is not available on all platforms.
2244 If you need to map the entire contents of a module to a particular
2245 section, consider using the facilities of the linker instead.
2249 @cindex @code{constructor} function attribute
2250 @cindex @code{destructor} function attribute
2251 The @code{constructor} attribute causes the function to be called
2252 automatically before execution enters @code{main ()}. Similarly, the
2253 @code{destructor} attribute causes the function to be called
2254 automatically after @code{main ()} has completed or @code{exit ()} has
2255 been called. Functions with these attributes are useful for
2256 initializing data that will be used implicitly during the execution of
2259 These attributes are not currently implemented for Objective-C@.
2261 @cindex @code{unused} attribute.
2263 This attribute, attached to a function, means that the function is meant
2264 to be possibly unused. GCC will not produce a warning for this
2265 function. GNU C++ does not currently support this attribute as
2266 definitions without parameters are valid in C++.
2268 @cindex @code{used} attribute.
2270 This attribute, attached to a function, means that code must be emitted
2271 for the function even if it appears that the function is not referenced.
2272 This is useful, for example, when the function is referenced only in
2275 @cindex @code{deprecated} attribute.
2277 The @code{deprecated} attribute results in a warning if the function
2278 is used anywhere in the source file. This is useful when identifying
2279 functions that are expected to be removed in a future version of a
2280 program. The warning also includes the location of the declaration
2281 of the deprecated function, to enable users to easily find further
2282 information about why the function is deprecated, or what they should
2283 do instead. Note that the warnings only occurs for uses:
2286 int old_fn () __attribute__ ((deprecated));
2288 int (*fn_ptr)() = old_fn;
2291 results in a warning on line 3 but not line 2.
2293 The @code{deprecated} attribute can also be used for variables and
2294 types (@pxref{Variable Attributes}, @pxref{Type Attributes}.)
2297 @cindex @code{weak} attribute
2298 The @code{weak} attribute causes the declaration to be emitted as a weak
2299 symbol rather than a global. This is primarily useful in defining
2300 library functions which can be overridden in user code, though it can
2301 also be used with non-function declarations. Weak symbols are supported
2302 for ELF targets, and also for a.out targets when using the GNU assembler
2306 @cindex @code{malloc} attribute
2307 The @code{malloc} attribute is used to tell the compiler that a function
2308 may be treated as if it were the malloc function. The compiler assumes
2309 that calls to malloc result in a pointers that cannot alias anything.
2310 This will often improve optimization.
2312 @item alias ("@var{target}")
2313 @cindex @code{alias} attribute
2314 The @code{alias} attribute causes the declaration to be emitted as an
2315 alias for another symbol, which must be specified. For instance,
2318 void __f () @{ /* @r{Do something.} */; @}
2319 void f () __attribute__ ((weak, alias ("__f")));
2322 declares @samp{f} to be a weak alias for @samp{__f}. In C++, the
2323 mangled name for the target must be used.
2325 Not all target machines support this attribute.
2327 @item visibility ("@var{visibility_type}")
2328 @cindex @code{visibility} attribute
2329 The @code{visibility} attribute on ELF targets causes the declaration
2330 to be emitted with default, hidden, protected or internal visibility.
2333 void __attribute__ ((visibility ("protected")))
2334 f () @{ /* @r{Do something.} */; @}
2335 int i __attribute__ ((visibility ("hidden")));
2338 See the ELF gABI for complete details, but the short story is:
2342 Default visibility is the normal case for ELF. This value is
2343 available for the visibility attribute to override other options
2344 that may change the assumed visibility of symbols.
2347 Hidden visibility indicates that the symbol will not be placed into
2348 the dynamic symbol table, so no other @dfn{module} (executable or
2349 shared library) can reference it directly.
2352 Protected visibility indicates that the symbol will be placed in the
2353 dynamic symbol table, but that references within the defining module
2354 will bind to the local symbol. That is, the symbol cannot be overridden
2358 Internal visibility is like hidden visibility, but with additional
2359 processor specific semantics. Unless otherwise specified by the psABI,
2360 gcc defines internal visibility to mean that the function is @emph{never}
2361 called from another module. Note that hidden symbols, while then cannot
2362 be referenced directly by other modules, can be referenced indirectly via
2363 function pointers. By indicating that a symbol cannot be called from
2364 outside the module, gcc may for instance omit the load of a PIC register
2365 since it is known that the calling function loaded the correct value.
2368 Not all ELF targets support this attribute.
2370 @item regparm (@var{number})
2371 @cindex @code{regparm} attribute
2372 @cindex functions that are passed arguments in registers on the 386
2373 On the Intel 386, the @code{regparm} attribute causes the compiler to
2374 pass up to @var{number} integer arguments in registers EAX,
2375 EDX, and ECX instead of on the stack. Functions that take a
2376 variable number of arguments will continue to be passed all of their
2377 arguments on the stack.
2379 Beware that on some ELF systems this attribute is unsuitable for
2380 global functions in shared libraries with lazy binding (which is the
2381 default). Lazy binding will send the first call via resolving code in
2382 the loader, which might assume EAX, EDX and ECX can be clobbered, as
2383 per the standard calling conventions. Solaris 8 is affected by this.
2384 GNU systems with GLIBC 2.1 or higher, and FreeBSD, are believed to be
2385 safe since the loaders there save all registers. (Lazy binding can be
2386 disabled with the linker or the loader if desired, to avoid the
2390 @cindex functions that pop the argument stack on the 386
2391 On the Intel 386, the @code{stdcall} attribute causes the compiler to
2392 assume that the called function will pop off the stack space used to
2393 pass arguments, unless it takes a variable number of arguments.
2396 @cindex functions that pop the argument stack on the 386
2397 On the Intel 386, the @code{fastcall} attribute causes the compiler to
2398 pass the first two arguments in the registers ECX and EDX. Subsequent
2399 arguments are passed on the stack. The called function will pop the
2400 arguments off the stack. If the number of arguments is variable all
2401 arguments are pushed on the stack.
2404 @cindex functions that do pop the argument stack on the 386
2406 On the Intel 386, the @code{cdecl} attribute causes the compiler to
2407 assume that the calling function will pop off the stack space used to
2408 pass arguments. This is
2409 useful to override the effects of the @option{-mrtd} switch.
2411 @item longcall/shortcall
2412 @cindex functions called via pointer on the RS/6000 and PowerPC
2413 On the RS/6000 and PowerPC, the @code{longcall} attribute causes the
2414 compiler to always call this function via a pointer, just as it would if
2415 the @option{-mlongcall} option had been specified. The @code{shortcall}
2416 attribute causes the compiler not to do this. These attributes override
2417 both the @option{-mlongcall} switch and the @code{#pragma longcall}
2420 @xref{RS/6000 and PowerPC Options}, for more information on when long
2421 calls are and are not necessary.
2423 @item long_call/short_call
2424 @cindex indirect calls on ARM
2425 This attribute allows to specify how to call a particular function on
2426 ARM@. Both attributes override the @option{-mlong-calls} (@pxref{ARM Options})
2427 command line switch and @code{#pragma long_calls} settings. The
2428 @code{long_call} attribute causes the compiler to always call the
2429 function by first loading its address into a register and then using the
2430 contents of that register. The @code{short_call} attribute always places
2431 the offset to the function from the call site into the @samp{BL}
2432 instruction directly.
2434 @item function_vector
2435 @cindex calling functions through the function vector on the H8/300 processors
2436 Use this attribute on the H8/300 and H8/300H to indicate that the specified
2437 function should be called through the function vector. Calling a
2438 function through the function vector will reduce code size, however;
2439 the function vector has a limited size (maximum 128 entries on the H8/300
2440 and 64 entries on the H8/300H) and shares space with the interrupt vector.
2442 You must use GAS and GLD from GNU binutils version 2.7 or later for
2443 this attribute to work correctly.
2446 @cindex interrupt handler functions
2447 Use this attribute on the ARM, AVR, C4x, M32R/D and Xstormy16 ports to indicate
2448 that the specified function is an interrupt handler. The compiler will
2449 generate function entry and exit sequences suitable for use in an
2450 interrupt handler when this attribute is present.
2452 Note, interrupt handlers for the H8/300, H8/300H and SH processors can
2453 be specified via the @code{interrupt_handler} attribute.
2455 Note, on the AVR interrupts will be enabled inside the function.
2457 Note, for the ARM you can specify the kind of interrupt to be handled by
2458 adding an optional parameter to the interrupt attribute like this:
2461 void f () __attribute__ ((interrupt ("IRQ")));
2464 Permissible values for this parameter are: IRQ, FIQ, SWI, ABORT and UNDEF@.
2466 @item interrupt_handler
2467 @cindex interrupt handler functions on the H8/300 and SH processors
2468 Use this attribute on the H8/300, H8/300H and SH to indicate that the
2469 specified function is an interrupt handler. The compiler will generate
2470 function entry and exit sequences suitable for use in an interrupt
2471 handler when this attribute is present.
2474 Use this attribute on the SH to indicate an @code{interrupt_handler}
2475 function should switch to an alternate stack. It expects a string
2476 argument that names a global variable holding the address of the
2481 void f () __attribute__ ((interrupt_handler,
2482 sp_switch ("alt_stack")));
2486 Use this attribute on the SH for an @code{interrupt_handle} to return using
2487 @code{trapa} instead of @code{rte}. This attribute expects an integer
2488 argument specifying the trap number to be used.
2491 @cindex eight bit data on the H8/300 and H8/300H
2492 Use this attribute on the H8/300 and H8/300H to indicate that the specified
2493 variable should be placed into the eight bit data section.
2494 The compiler will generate more efficient code for certain operations
2495 on data in the eight bit data area. Note the eight bit data area is limited to
2498 You must use GAS and GLD from GNU binutils version 2.7 or later for
2499 this attribute to work correctly.
2502 @cindex tiny data section on the H8/300H
2503 Use this attribute on the H8/300H to indicate that the specified
2504 variable should be placed into the tiny data section.
2505 The compiler will generate more efficient code for loads and stores
2506 on data in the tiny data section. Note the tiny data area is limited to
2507 slightly under 32kbytes of data.
2510 @cindex signal handler functions on the AVR processors
2511 Use this attribute on the AVR to indicate that the specified
2512 function is an signal handler. The compiler will generate function
2513 entry and exit sequences suitable for use in an signal handler when this
2514 attribute is present. Interrupts will be disabled inside function.
2517 @cindex function without a prologue/epilogue code
2518 Use this attribute on the ARM, AVR, C4x and IP2K ports to indicate that the
2519 specified function do not need prologue/epilogue sequences generated by
2520 the compiler. It is up to the programmer to provide these sequences.
2522 @item model (@var{model-name})
2523 @cindex function addressability on the M32R/D
2524 Use this attribute on the M32R/D to set the addressability of an object,
2525 and the code generated for a function.
2526 The identifier @var{model-name} is one of @code{small}, @code{medium},
2527 or @code{large}, representing each of the code models.
2529 Small model objects live in the lower 16MB of memory (so that their
2530 addresses can be loaded with the @code{ld24} instruction), and are
2531 callable with the @code{bl} instruction.
2533 Medium model objects may live anywhere in the 32-bit address space (the
2534 compiler will generate @code{seth/add3} instructions to load their addresses),
2535 and are callable with the @code{bl} instruction.
2537 Large model objects may live anywhere in the 32-bit address space (the
2538 compiler will generate @code{seth/add3} instructions to load their addresses),
2539 and may not be reachable with the @code{bl} instruction (the compiler will
2540 generate the much slower @code{seth/add3/jl} instruction sequence).
2543 @cindex functions which handle memory bank switching
2544 On 68HC11 and 68HC12 the @code{far} attribute causes the compiler to
2545 use a calling convention that takes care of switching memory banks when
2546 entering and leaving a function. This calling convention is also the
2547 default when using the @option{-mlong-calls} option.
2549 On 68HC12 the compiler will use the @code{call} and @code{rtc} instructions
2550 to call and return from a function.
2552 On 68HC11 the compiler will generate a sequence of instructions
2553 to invoke a board-specific routine to switch the memory bank and call the
2554 real function. The board-specific routine simulates a @code{call}.
2555 At the end of a function, it will jump to a board-specific routine
2556 instead of using @code{rts}. The board-specific return routine simulates
2560 @cindex functions which do not handle memory bank switching on 68HC11/68HC12
2561 On 68HC11 and 68HC12 the @code{near} attribute causes the compiler to
2562 use the normal calling convention based on @code{jsr} and @code{rts}.
2563 This attribute can be used to cancel the effect of the @option{-mlong-calls}
2568 You can specify multiple attributes in a declaration by separating them
2569 by commas within the double parentheses or by immediately following an
2570 attribute declaration with another attribute declaration.
2572 @cindex @code{#pragma}, reason for not using
2573 @cindex pragma, reason for not using
2574 Some people object to the @code{__attribute__} feature, suggesting that
2575 ISO C's @code{#pragma} should be used instead. At the time
2576 @code{__attribute__} was designed, there were two reasons for not doing
2581 It is impossible to generate @code{#pragma} commands from a macro.
2584 There is no telling what the same @code{#pragma} might mean in another
2588 These two reasons applied to almost any application that might have been
2589 proposed for @code{#pragma}. It was basically a mistake to use
2590 @code{#pragma} for @emph{anything}.
2592 The ISO C99 standard includes @code{_Pragma}, which now allows pragmas
2593 to be generated from macros. In addition, a @code{#pragma GCC}
2594 namespace is now in use for GCC-specific pragmas. However, it has been
2595 found convenient to use @code{__attribute__} to achieve a natural
2596 attachment of attributes to their corresponding declarations, whereas
2597 @code{#pragma GCC} is of use for constructs that do not naturally form
2598 part of the grammar. @xref{Other Directives,,Miscellaneous
2599 Preprocessing Directives, cpp, The C Preprocessor}.
2601 @node Attribute Syntax
2602 @section Attribute Syntax
2603 @cindex attribute syntax
2605 This section describes the syntax with which @code{__attribute__} may be
2606 used, and the constructs to which attribute specifiers bind, for the C
2607 language. Some details may vary for C++ and Objective-C@. Because of
2608 infelicities in the grammar for attributes, some forms described here
2609 may not be successfully parsed in all cases.
2611 There are some problems with the semantics of attributes in C++. For
2612 example, there are no manglings for attributes, although they may affect
2613 code generation, so problems may arise when attributed types are used in
2614 conjunction with templates or overloading. Similarly, @code{typeid}
2615 does not distinguish between types with different attributes. Support
2616 for attributes in C++ may be restricted in future to attributes on
2617 declarations only, but not on nested declarators.
2619 @xref{Function Attributes}, for details of the semantics of attributes
2620 applying to functions. @xref{Variable Attributes}, for details of the
2621 semantics of attributes applying to variables. @xref{Type Attributes},
2622 for details of the semantics of attributes applying to structure, union
2623 and enumerated types.
2625 An @dfn{attribute specifier} is of the form
2626 @code{__attribute__ ((@var{attribute-list}))}. An @dfn{attribute list}
2627 is a possibly empty comma-separated sequence of @dfn{attributes}, where
2628 each attribute is one of the following:
2632 Empty. Empty attributes are ignored.
2635 A word (which may be an identifier such as @code{unused}, or a reserved
2636 word such as @code{const}).
2639 A word, followed by, in parentheses, parameters for the attribute.
2640 These parameters take one of the following forms:
2644 An identifier. For example, @code{mode} attributes use this form.
2647 An identifier followed by a comma and a non-empty comma-separated list
2648 of expressions. For example, @code{format} attributes use this form.
2651 A possibly empty comma-separated list of expressions. For example,
2652 @code{format_arg} attributes use this form with the list being a single
2653 integer constant expression, and @code{alias} attributes use this form
2654 with the list being a single string constant.
2658 An @dfn{attribute specifier list} is a sequence of one or more attribute
2659 specifiers, not separated by any other tokens.
2661 An attribute specifier list may appear after the colon following a
2662 label, other than a @code{case} or @code{default} label. The only
2663 attribute it makes sense to use after a label is @code{unused}. This
2664 feature is intended for code generated by programs which contains labels
2665 that may be unused but which is compiled with @option{-Wall}. It would
2666 not normally be appropriate to use in it human-written code, though it
2667 could be useful in cases where the code that jumps to the label is
2668 contained within an @code{#ifdef} conditional.
2670 An attribute specifier list may appear as part of a @code{struct},
2671 @code{union} or @code{enum} specifier. It may go either immediately
2672 after the @code{struct}, @code{union} or @code{enum} keyword, or after
2673 the closing brace. It is ignored if the content of the structure, union
2674 or enumerated type is not defined in the specifier in which the
2675 attribute specifier list is used---that is, in usages such as
2676 @code{struct __attribute__((foo)) bar} with no following opening brace.
2677 Where attribute specifiers follow the closing brace, they are considered
2678 to relate to the structure, union or enumerated type defined, not to any
2679 enclosing declaration the type specifier appears in, and the type
2680 defined is not complete until after the attribute specifiers.
2681 @c Otherwise, there would be the following problems: a shift/reduce
2682 @c conflict between attributes binding the struct/union/enum and
2683 @c binding to the list of specifiers/qualifiers; and "aligned"
2684 @c attributes could use sizeof for the structure, but the size could be
2685 @c changed later by "packed" attributes.
2687 Otherwise, an attribute specifier appears as part of a declaration,
2688 counting declarations of unnamed parameters and type names, and relates
2689 to that declaration (which may be nested in another declaration, for
2690 example in the case of a parameter declaration), or to a particular declarator
2691 within a declaration. Where an
2692 attribute specifier is applied to a parameter declared as a function or
2693 an array, it should apply to the function or array rather than the
2694 pointer to which the parameter is implicitly converted, but this is not
2695 yet correctly implemented.
2697 Any list of specifiers and qualifiers at the start of a declaration may
2698 contain attribute specifiers, whether or not such a list may in that
2699 context contain storage class specifiers. (Some attributes, however,
2700 are essentially in the nature of storage class specifiers, and only make
2701 sense where storage class specifiers may be used; for example,
2702 @code{section}.) There is one necessary limitation to this syntax: the
2703 first old-style parameter declaration in a function definition cannot
2704 begin with an attribute specifier, because such an attribute applies to
2705 the function instead by syntax described below (which, however, is not
2706 yet implemented in this case). In some other cases, attribute
2707 specifiers are permitted by this grammar but not yet supported by the
2708 compiler. All attribute specifiers in this place relate to the
2709 declaration as a whole. In the obsolescent usage where a type of
2710 @code{int} is implied by the absence of type specifiers, such a list of
2711 specifiers and qualifiers may be an attribute specifier list with no
2712 other specifiers or qualifiers.
2714 An attribute specifier list may appear immediately before a declarator
2715 (other than the first) in a comma-separated list of declarators in a
2716 declaration of more than one identifier using a single list of
2717 specifiers and qualifiers. Such attribute specifiers apply
2718 only to the identifier before whose declarator they appear. For
2722 __attribute__((noreturn)) void d0 (void),
2723 __attribute__((format(printf, 1, 2))) d1 (const char *, ...),
2728 the @code{noreturn} attribute applies to all the functions
2729 declared; the @code{format} attribute only applies to @code{d1}.
2731 An attribute specifier list may appear immediately before the comma,
2732 @code{=} or semicolon terminating the declaration of an identifier other
2733 than a function definition. At present, such attribute specifiers apply
2734 to the declared object or function, but in future they may attach to the
2735 outermost adjacent declarator. In simple cases there is no difference,
2736 but, for example, in
2739 void (****f)(void) __attribute__((noreturn));
2743 at present the @code{noreturn} attribute applies to @code{f}, which
2744 causes a warning since @code{f} is not a function, but in future it may
2745 apply to the function @code{****f}. The precise semantics of what
2746 attributes in such cases will apply to are not yet specified. Where an
2747 assembler name for an object or function is specified (@pxref{Asm
2748 Labels}), at present the attribute must follow the @code{asm}
2749 specification; in future, attributes before the @code{asm} specification
2750 may apply to the adjacent declarator, and those after it to the declared
2753 An attribute specifier list may, in future, be permitted to appear after
2754 the declarator in a function definition (before any old-style parameter
2755 declarations or the function body).
2757 Attribute specifiers may be mixed with type qualifiers appearing inside
2758 the @code{[]} of a parameter array declarator, in the C99 construct by
2759 which such qualifiers are applied to the pointer to which the array is
2760 implicitly converted. Such attribute specifiers apply to the pointer,
2761 not to the array, but at present this is not implemented and they are
2764 An attribute specifier list may appear at the start of a nested
2765 declarator. At present, there are some limitations in this usage: the
2766 attributes correctly apply to the declarator, but for most individual
2767 attributes the semantics this implies are not implemented.
2768 When attribute specifiers follow the @code{*} of a pointer
2769 declarator, they may be mixed with any type qualifiers present.
2770 The following describes the formal semantics of this syntax. It will make the
2771 most sense if you are familiar with the formal specification of
2772 declarators in the ISO C standard.
2774 Consider (as in C99 subclause 6.7.5 paragraph 4) a declaration @code{T
2775 D1}, where @code{T} contains declaration specifiers that specify a type
2776 @var{Type} (such as @code{int}) and @code{D1} is a declarator that
2777 contains an identifier @var{ident}. The type specified for @var{ident}
2778 for derived declarators whose type does not include an attribute
2779 specifier is as in the ISO C standard.
2781 If @code{D1} has the form @code{( @var{attribute-specifier-list} D )},
2782 and the declaration @code{T D} specifies the type
2783 ``@var{derived-declarator-type-list} @var{Type}'' for @var{ident}, then
2784 @code{T D1} specifies the type ``@var{derived-declarator-type-list}
2785 @var{attribute-specifier-list} @var{Type}'' for @var{ident}.
2787 If @code{D1} has the form @code{*
2788 @var{type-qualifier-and-attribute-specifier-list} D}, and the
2789 declaration @code{T D} specifies the type
2790 ``@var{derived-declarator-type-list} @var{Type}'' for @var{ident}, then
2791 @code{T D1} specifies the type ``@var{derived-declarator-type-list}
2792 @var{type-qualifier-and-attribute-specifier-list} @var{Type}'' for
2798 void (__attribute__((noreturn)) ****f) (void);
2802 specifies the type ``pointer to pointer to pointer to pointer to
2803 non-returning function returning @code{void}''. As another example,
2806 char *__attribute__((aligned(8))) *f;
2810 specifies the type ``pointer to 8-byte-aligned pointer to @code{char}''.
2811 Note again that this does not work with most attributes; for example,
2812 the usage of @samp{aligned} and @samp{noreturn} attributes given above
2813 is not yet supported.
2815 For compatibility with existing code written for compiler versions that
2816 did not implement attributes on nested declarators, some laxity is
2817 allowed in the placing of attributes. If an attribute that only applies
2818 to types is applied to a declaration, it will be treated as applying to
2819 the type of that declaration. If an attribute that only applies to
2820 declarations is applied to the type of a declaration, it will be treated
2821 as applying to that declaration; and, for compatibility with code
2822 placing the attributes immediately before the identifier declared, such
2823 an attribute applied to a function return type will be treated as
2824 applying to the function type, and such an attribute applied to an array
2825 element type will be treated as applying to the array type. If an
2826 attribute that only applies to function types is applied to a
2827 pointer-to-function type, it will be treated as applying to the pointer
2828 target type; if such an attribute is applied to a function return type
2829 that is not a pointer-to-function type, it will be treated as applying
2830 to the function type.
2832 @node Function Prototypes
2833 @section Prototypes and Old-Style Function Definitions
2834 @cindex function prototype declarations
2835 @cindex old-style function definitions
2836 @cindex promotion of formal parameters
2838 GNU C extends ISO C to allow a function prototype to override a later
2839 old-style non-prototype definition. Consider the following example:
2842 /* @r{Use prototypes unless the compiler is old-fashioned.} */
2849 /* @r{Prototype function declaration.} */
2850 int isroot P((uid_t));
2852 /* @r{Old-style function definition.} */
2854 isroot (x) /* ??? lossage here ??? */
2861 Suppose the type @code{uid_t} happens to be @code{short}. ISO C does
2862 not allow this example, because subword arguments in old-style
2863 non-prototype definitions are promoted. Therefore in this example the
2864 function definition's argument is really an @code{int}, which does not
2865 match the prototype argument type of @code{short}.
2867 This restriction of ISO C makes it hard to write code that is portable
2868 to traditional C compilers, because the programmer does not know
2869 whether the @code{uid_t} type is @code{short}, @code{int}, or
2870 @code{long}. Therefore, in cases like these GNU C allows a prototype
2871 to override a later old-style definition. More precisely, in GNU C, a
2872 function prototype argument type overrides the argument type specified
2873 by a later old-style definition if the former type is the same as the
2874 latter type before promotion. Thus in GNU C the above example is
2875 equivalent to the following:
2888 GNU C++ does not support old-style function definitions, so this
2889 extension is irrelevant.
2892 @section C++ Style Comments
2894 @cindex C++ comments
2895 @cindex comments, C++ style
2897 In GNU C, you may use C++ style comments, which start with @samp{//} and
2898 continue until the end of the line. Many other C implementations allow
2899 such comments, and they are included in the 1999 C standard. However,
2900 C++ style comments are not recognized if you specify an @option{-std}
2901 option specifying a version of ISO C before C99, or @option{-ansi}
2902 (equivalent to @option{-std=c89}).
2905 @section Dollar Signs in Identifier Names
2907 @cindex dollar signs in identifier names
2908 @cindex identifier names, dollar signs in
2910 In GNU C, you may normally use dollar signs in identifier names.
2911 This is because many traditional C implementations allow such identifiers.
2912 However, dollar signs in identifiers are not supported on a few target
2913 machines, typically because the target assembler does not allow them.
2915 @node Character Escapes
2916 @section The Character @key{ESC} in Constants
2918 You can use the sequence @samp{\e} in a string or character constant to
2919 stand for the ASCII character @key{ESC}.
2922 @section Inquiring on Alignment of Types or Variables
2924 @cindex type alignment
2925 @cindex variable alignment
2927 The keyword @code{__alignof__} allows you to inquire about how an object
2928 is aligned, or the minimum alignment usually required by a type. Its
2929 syntax is just like @code{sizeof}.
2931 For example, if the target machine requires a @code{double} value to be
2932 aligned on an 8-byte boundary, then @code{__alignof__ (double)} is 8.
2933 This is true on many RISC machines. On more traditional machine
2934 designs, @code{__alignof__ (double)} is 4 or even 2.
2936 Some machines never actually require alignment; they allow reference to any
2937 data type even at an odd addresses. For these machines, @code{__alignof__}
2938 reports the @emph{recommended} alignment of a type.
2940 If the operand of @code{__alignof__} is an lvalue rather than a type,
2941 its value is the required alignment for its type, taking into account
2942 any minimum alignment specified with GCC's @code{__attribute__}
2943 extension (@pxref{Variable Attributes}). For example, after this
2947 struct foo @{ int x; char y; @} foo1;
2951 the value of @code{__alignof__ (foo1.y)} is 1, even though its actual
2952 alignment is probably 2 or 4, the same as @code{__alignof__ (int)}.
2954 It is an error to ask for the alignment of an incomplete type.
2956 @node Variable Attributes
2957 @section Specifying Attributes of Variables
2958 @cindex attribute of variables
2959 @cindex variable attributes
2961 The keyword @code{__attribute__} allows you to specify special
2962 attributes of variables or structure fields. This keyword is followed
2963 by an attribute specification inside double parentheses. Some
2964 attributes are currently defined generically for variables.
2965 Other attributes are defined for variables on particular target
2966 systems. Other attributes are available for functions
2967 (@pxref{Function Attributes}) and for types (@pxref{Type Attributes}).
2968 Other front ends might define more attributes
2969 (@pxref{C++ Extensions,,Extensions to the C++ Language}).
2971 You may also specify attributes with @samp{__} preceding and following
2972 each keyword. This allows you to use them in header files without
2973 being concerned about a possible macro of the same name. For example,
2974 you may use @code{__aligned__} instead of @code{aligned}.
2976 @xref{Attribute Syntax}, for details of the exact syntax for using
2980 @cindex @code{aligned} attribute
2981 @item aligned (@var{alignment})
2982 This attribute specifies a minimum alignment for the variable or
2983 structure field, measured in bytes. For example, the declaration:
2986 int x __attribute__ ((aligned (16))) = 0;
2990 causes the compiler to allocate the global variable @code{x} on a
2991 16-byte boundary. On a 68040, this could be used in conjunction with
2992 an @code{asm} expression to access the @code{move16} instruction which
2993 requires 16-byte aligned operands.
2995 You can also specify the alignment of structure fields. For example, to
2996 create a double-word aligned @code{int} pair, you could write:
2999 struct foo @{ int x[2] __attribute__ ((aligned (8))); @};
3003 This is an alternative to creating a union with a @code{double} member
3004 that forces the union to be double-word aligned.
3006 As in the preceding examples, you can explicitly specify the alignment
3007 (in bytes) that you wish the compiler to use for a given variable or
3008 structure field. Alternatively, you can leave out the alignment factor
3009 and just ask the compiler to align a variable or field to the maximum
3010 useful alignment for the target machine you are compiling for. For
3011 example, you could write:
3014 short array[3] __attribute__ ((aligned));
3017 Whenever you leave out the alignment factor in an @code{aligned} attribute
3018 specification, the compiler automatically sets the alignment for the declared
3019 variable or field to the largest alignment which is ever used for any data
3020 type on the target machine you are compiling for. Doing this can often make
3021 copy operations more efficient, because the compiler can use whatever
3022 instructions copy the biggest chunks of memory when performing copies to
3023 or from the variables or fields that you have aligned this way.
3025 The @code{aligned} attribute can only increase the alignment; but you
3026 can decrease it by specifying @code{packed} as well. See below.
3028 Note that the effectiveness of @code{aligned} attributes may be limited
3029 by inherent limitations in your linker. On many systems, the linker is
3030 only able to arrange for variables to be aligned up to a certain maximum
3031 alignment. (For some linkers, the maximum supported alignment may
3032 be very very small.) If your linker is only able to align variables
3033 up to a maximum of 8 byte alignment, then specifying @code{aligned(16)}
3034 in an @code{__attribute__} will still only provide you with 8 byte
3035 alignment. See your linker documentation for further information.
3039 @cindex @code{common} attribute
3040 @cindex @code{nocommon} attribute
3043 The @code{common} attribute requests GCC to place a variable in
3044 ``common'' storage. The @code{nocommon} attribute requests the
3045 opposite -- to allocate space for it directly.
3047 These attributes override the default chosen by the
3048 @option{-fno-common} and @option{-fcommon} flags respectively.
3051 @cindex @code{deprecated} attribute
3052 The @code{deprecated} attribute results in a warning if the variable
3053 is used anywhere in the source file. This is useful when identifying
3054 variables that are expected to be removed in a future version of a
3055 program. The warning also includes the location of the declaration
3056 of the deprecated variable, to enable users to easily find further
3057 information about why the variable is deprecated, or what they should
3058 do instead. Note that the warnings only occurs for uses:
3061 extern int old_var __attribute__ ((deprecated));
3063 int new_fn () @{ return old_var; @}
3066 results in a warning on line 3 but not line 2.
3068 The @code{deprecated} attribute can also be used for functions and
3069 types (@pxref{Function Attributes}, @pxref{Type Attributes}.)
3071 @item mode (@var{mode})
3072 @cindex @code{mode} attribute
3073 This attribute specifies the data type for the declaration---whichever
3074 type corresponds to the mode @var{mode}. This in effect lets you
3075 request an integer or floating point type according to its width.
3077 You may also specify a mode of @samp{byte} or @samp{__byte__} to
3078 indicate the mode corresponding to a one-byte integer, @samp{word} or
3079 @samp{__word__} for the mode of a one-word integer, and @samp{pointer}
3080 or @samp{__pointer__} for the mode used to represent pointers.
3083 @cindex @code{packed} attribute
3084 The @code{packed} attribute specifies that a variable or structure field
3085 should have the smallest possible alignment---one byte for a variable,
3086 and one bit for a field, unless you specify a larger value with the
3087 @code{aligned} attribute.
3089 Here is a structure in which the field @code{x} is packed, so that it
3090 immediately follows @code{a}:
3096 int x[2] __attribute__ ((packed));
3100 @item section ("@var{section-name}")
3101 @cindex @code{section} variable attribute
3102 Normally, the compiler places the objects it generates in sections like
3103 @code{data} and @code{bss}. Sometimes, however, you need additional sections,
3104 or you need certain particular variables to appear in special sections,
3105 for example to map to special hardware. The @code{section}
3106 attribute specifies that a variable (or function) lives in a particular
3107 section. For example, this small program uses several specific section names:
3110 struct duart a __attribute__ ((section ("DUART_A"))) = @{ 0 @};
3111 struct duart b __attribute__ ((section ("DUART_B"))) = @{ 0 @};
3112 char stack[10000] __attribute__ ((section ("STACK"))) = @{ 0 @};
3113 int init_data __attribute__ ((section ("INITDATA"))) = 0;
3117 /* Initialize stack pointer */
3118 init_sp (stack + sizeof (stack));
3120 /* Initialize initialized data */
3121 memcpy (&init_data, &data, &edata - &data);
3123 /* Turn on the serial ports */
3130 Use the @code{section} attribute with an @emph{initialized} definition
3131 of a @emph{global} variable, as shown in the example. GCC issues
3132 a warning and otherwise ignores the @code{section} attribute in
3133 uninitialized variable declarations.
3135 You may only use the @code{section} attribute with a fully initialized
3136 global definition because of the way linkers work. The linker requires
3137 each object be defined once, with the exception that uninitialized
3138 variables tentatively go in the @code{common} (or @code{bss}) section
3139 and can be multiply ``defined''. You can force a variable to be
3140 initialized with the @option{-fno-common} flag or the @code{nocommon}
3143 Some file formats do not support arbitrary sections so the @code{section}
3144 attribute is not available on all platforms.
3145 If you need to map the entire contents of a module to a particular
3146 section, consider using the facilities of the linker instead.
3149 @cindex @code{shared} variable attribute
3150 On Windows NT, in addition to putting variable definitions in a named
3151 section, the section can also be shared among all running copies of an
3152 executable or DLL@. For example, this small program defines shared data
3153 by putting it in a named section @code{shared} and marking the section
3157 int foo __attribute__((section ("shared"), shared)) = 0;
3162 /* Read and write foo. All running
3163 copies see the same value. */
3169 You may only use the @code{shared} attribute along with @code{section}
3170 attribute with a fully initialized global definition because of the way
3171 linkers work. See @code{section} attribute for more information.
3173 The @code{shared} attribute is only available on Windows NT@.
3175 @item tls_model ("@var{tls_model}")
3176 @cindex @code{tls_model} attribute
3177 The @code{tls_model} attribute sets thread-local storage model
3178 (@pxref{Thread-Local}) of a particular @code{__thread} variable,
3179 overriding @code{-ftls-model=} command line switch on a per-variable
3181 The @var{tls_model} argument should be one of @code{global-dynamic},
3182 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
3184 Not all targets support this attribute.
3186 @item transparent_union
3187 This attribute, attached to a function parameter which is a union, means
3188 that the corresponding argument may have the type of any union member,
3189 but the argument is passed as if its type were that of the first union
3190 member. For more details see @xref{Type Attributes}. You can also use
3191 this attribute on a @code{typedef} for a union data type; then it
3192 applies to all function parameters with that type.
3195 This attribute, attached to a variable, means that the variable is meant
3196 to be possibly unused. GCC will not produce a warning for this
3199 @item vector_size (@var{bytes})
3200 This attribute specifies the vector size for the variable, measured in
3201 bytes. For example, the declaration:
3204 int foo __attribute__ ((vector_size (16)));
3208 causes the compiler to set the mode for @code{foo}, to be 16 bytes,
3209 divided into @code{int} sized units. Assuming a 32-bit int (a vector of
3210 4 units of 4 bytes), the corresponding mode of @code{foo} will be V4SI@.
3212 This attribute is only applicable to integral and float scalars,
3213 although arrays, pointers, and function return values are allowed in
3214 conjunction with this construct.
3216 Aggregates with this attribute are invalid, even if they are of the same
3217 size as a corresponding scalar. For example, the declaration:
3220 struct S @{ int a; @};
3221 struct S __attribute__ ((vector_size (16))) foo;
3225 is invalid even if the size of the structure is the same as the size of
3229 The @code{weak} attribute is described in @xref{Function Attributes}.
3232 @subsection M32R/D Variable Attributes
3234 One attribute is currently defined for the M32R/D.
3237 @item model (@var{model-name})
3238 @cindex variable addressability on the M32R/D
3239 Use this attribute on the M32R/D to set the addressability of an object.
3240 The identifier @var{model-name} is one of @code{small}, @code{medium},
3241 or @code{large}, representing each of the code models.
3243 Small model objects live in the lower 16MB of memory (so that their
3244 addresses can be loaded with the @code{ld24} instruction).
3246 Medium and large model objects may live anywhere in the 32-bit address space
3247 (the compiler will generate @code{seth/add3} instructions to load their
3251 @subsection i386 Variable Attributes
3253 Two attributes are currently defined for i386 configurations:
3254 @code{ms_struct} and @code{gcc_struct}
3259 @cindex @code{ms_struct} attribute
3260 @cindex @code{gcc_struct} attribute
3262 If @code{packed} is used on a structure, or if bit-fields are used
3263 it may be that the Microsoft ABI packs them differently
3264 than GCC would normally pack them. Particularly when moving packed
3265 data between functions compiled with GCC and the native Microsoft compiler
3266 (either via function call or as data in a file), it may be necessary to access
3269 Currently @option{-m[no-]ms-bitfields} is provided for the Windows X86
3270 compilers to match the native Microsoft compiler.
3273 @node Type Attributes
3274 @section Specifying Attributes of Types
3275 @cindex attribute of types
3276 @cindex type attributes
3278 The keyword @code{__attribute__} allows you to specify special
3279 attributes of @code{struct} and @code{union} types when you define such
3280 types. This keyword is followed by an attribute specification inside
3281 double parentheses. Six attributes are currently defined for types:
3282 @code{aligned}, @code{packed}, @code{transparent_union}, @code{unused},
3283 @code{deprecated} and @code{may_alias}. Other attributes are defined for
3284 functions (@pxref{Function Attributes}) and for variables
3285 (@pxref{Variable Attributes}).
3287 You may also specify any one of these attributes with @samp{__}
3288 preceding and following its keyword. This allows you to use these
3289 attributes in header files without being concerned about a possible
3290 macro of the same name. For example, you may use @code{__aligned__}
3291 instead of @code{aligned}.
3293 You may specify the @code{aligned} and @code{transparent_union}
3294 attributes either in a @code{typedef} declaration or just past the
3295 closing curly brace of a complete enum, struct or union type
3296 @emph{definition} and the @code{packed} attribute only past the closing
3297 brace of a definition.
3299 You may also specify attributes between the enum, struct or union
3300 tag and the name of the type rather than after the closing brace.
3302 @xref{Attribute Syntax}, for details of the exact syntax for using
3306 @cindex @code{aligned} attribute
3307 @item aligned (@var{alignment})
3308 This attribute specifies a minimum alignment (in bytes) for variables
3309 of the specified type. For example, the declarations:
3312 struct S @{ short f[3]; @} __attribute__ ((aligned (8)));
3313 typedef int more_aligned_int __attribute__ ((aligned (8)));
3317 force the compiler to insure (as far as it can) that each variable whose
3318 type is @code{struct S} or @code{more_aligned_int} will be allocated and
3319 aligned @emph{at least} on a 8-byte boundary. On a SPARC, having all
3320 variables of type @code{struct S} aligned to 8-byte boundaries allows
3321 the compiler to use the @code{ldd} and @code{std} (doubleword load and
3322 store) instructions when copying one variable of type @code{struct S} to
3323 another, thus improving run-time efficiency.
3325 Note that the alignment of any given @code{struct} or @code{union} type
3326 is required by the ISO C standard to be at least a perfect multiple of
3327 the lowest common multiple of the alignments of all of the members of
3328 the @code{struct} or @code{union} in question. This means that you @emph{can}
3329 effectively adjust the alignment of a @code{struct} or @code{union}
3330 type by attaching an @code{aligned} attribute to any one of the members
3331 of such a type, but the notation illustrated in the example above is a
3332 more obvious, intuitive, and readable way to request the compiler to
3333 adjust the alignment of an entire @code{struct} or @code{union} type.
3335 As in the preceding example, you can explicitly specify the alignment
3336 (in bytes) that you wish the compiler to use for a given @code{struct}
3337 or @code{union} type. Alternatively, you can leave out the alignment factor
3338 and just ask the compiler to align a type to the maximum
3339 useful alignment for the target machine you are compiling for. For
3340 example, you could write:
3343 struct S @{ short f[3]; @} __attribute__ ((aligned));
3346 Whenever you leave out the alignment factor in an @code{aligned}
3347 attribute specification, the compiler automatically sets the alignment
3348 for the type to the largest alignment which is ever used for any data
3349 type on the target machine you are compiling for. Doing this can often
3350 make copy operations more efficient, because the compiler can use
3351 whatever instructions copy the biggest chunks of memory when performing
3352 copies to or from the variables which have types that you have aligned
3355 In the example above, if the size of each @code{short} is 2 bytes, then
3356 the size of the entire @code{struct S} type is 6 bytes. The smallest
3357 power of two which is greater than or equal to that is 8, so the
3358 compiler sets the alignment for the entire @code{struct S} type to 8
3361 Note that although you can ask the compiler to select a time-efficient
3362 alignment for a given type and then declare only individual stand-alone
3363 objects of that type, the compiler's ability to select a time-efficient
3364 alignment is primarily useful only when you plan to create arrays of
3365 variables having the relevant (efficiently aligned) type. If you
3366 declare or use arrays of variables of an efficiently-aligned type, then
3367 it is likely that your program will also be doing pointer arithmetic (or
3368 subscripting, which amounts to the same thing) on pointers to the
3369 relevant type, and the code that the compiler generates for these
3370 pointer arithmetic operations will often be more efficient for
3371 efficiently-aligned types than for other types.
3373 The @code{aligned} attribute can only increase the alignment; but you
3374 can decrease it by specifying @code{packed} as well. See below.
3376 Note that the effectiveness of @code{aligned} attributes may be limited
3377 by inherent limitations in your linker. On many systems, the linker is
3378 only able to arrange for variables to be aligned up to a certain maximum
3379 alignment. (For some linkers, the maximum supported alignment may
3380 be very very small.) If your linker is only able to align variables
3381 up to a maximum of 8 byte alignment, then specifying @code{aligned(16)}
3382 in an @code{__attribute__} will still only provide you with 8 byte
3383 alignment. See your linker documentation for further information.
3386 This attribute, attached to an @code{enum}, @code{struct}, or
3387 @code{union} type definition, specified that the minimum required memory
3388 be used to represent the type.
3390 @opindex fshort-enums
3391 Specifying this attribute for @code{struct} and @code{union} types is
3392 equivalent to specifying the @code{packed} attribute on each of the
3393 structure or union members. Specifying the @option{-fshort-enums}
3394 flag on the line is equivalent to specifying the @code{packed}
3395 attribute on all @code{enum} definitions.
3397 You may only specify this attribute after a closing curly brace on an
3398 @code{enum} definition, not in a @code{typedef} declaration, unless that
3399 declaration also contains the definition of the @code{enum}.
3401 @item transparent_union
3402 This attribute, attached to a @code{union} type definition, indicates
3403 that any function parameter having that union type causes calls to that
3404 function to be treated in a special way.
3406 First, the argument corresponding to a transparent union type can be of
3407 any type in the union; no cast is required. Also, if the union contains
3408 a pointer type, the corresponding argument can be a null pointer
3409 constant or a void pointer expression; and if the union contains a void
3410 pointer type, the corresponding argument can be any pointer expression.
3411 If the union member type is a pointer, qualifiers like @code{const} on
3412 the referenced type must be respected, just as with normal pointer
3415 Second, the argument is passed to the function using the calling
3416 conventions of first member of the transparent union, not the calling
3417 conventions of the union itself. All members of the union must have the
3418 same machine representation; this is necessary for this argument passing
3421 Transparent unions are designed for library functions that have multiple
3422 interfaces for compatibility reasons. For example, suppose the
3423 @code{wait} function must accept either a value of type @code{int *} to
3424 comply with Posix, or a value of type @code{union wait *} to comply with
3425 the 4.1BSD interface. If @code{wait}'s parameter were @code{void *},
3426 @code{wait} would accept both kinds of arguments, but it would also
3427 accept any other pointer type and this would make argument type checking
3428 less useful. Instead, @code{<sys/wait.h>} might define the interface
3436 @} wait_status_ptr_t __attribute__ ((__transparent_union__));
3438 pid_t wait (wait_status_ptr_t);
3441 This interface allows either @code{int *} or @code{union wait *}
3442 arguments to be passed, using the @code{int *} calling convention.
3443 The program can call @code{wait} with arguments of either type:
3446 int w1 () @{ int w; return wait (&w); @}
3447 int w2 () @{ union wait w; return wait (&w); @}
3450 With this interface, @code{wait}'s implementation might look like this:
3453 pid_t wait (wait_status_ptr_t p)
3455 return waitpid (-1, p.__ip, 0);
3460 When attached to a type (including a @code{union} or a @code{struct}),
3461 this attribute means that variables of that type are meant to appear
3462 possibly unused. GCC will not produce a warning for any variables of
3463 that type, even if the variable appears to do nothing. This is often
3464 the case with lock or thread classes, which are usually defined and then
3465 not referenced, but contain constructors and destructors that have
3466 nontrivial bookkeeping functions.
3469 The @code{deprecated} attribute results in a warning if the type
3470 is used anywhere in the source file. This is useful when identifying
3471 types that are expected to be removed in a future version of a program.
3472 If possible, the warning also includes the location of the declaration
3473 of the deprecated type, to enable users to easily find further
3474 information about why the type is deprecated, or what they should do
3475 instead. Note that the warnings only occur for uses and then only
3476 if the type is being applied to an identifier that itself is not being
3477 declared as deprecated.
3480 typedef int T1 __attribute__ ((deprecated));
3484 typedef T1 T3 __attribute__ ((deprecated));
3485 T3 z __attribute__ ((deprecated));
3488 results in a warning on line 2 and 3 but not lines 4, 5, or 6. No
3489 warning is issued for line 4 because T2 is not explicitly
3490 deprecated. Line 5 has no warning because T3 is explicitly
3491 deprecated. Similarly for line 6.
3493 The @code{deprecated} attribute can also be used for functions and
3494 variables (@pxref{Function Attributes}, @pxref{Variable Attributes}.)
3497 Accesses to objects with types with this attribute are not subjected to
3498 type-based alias analysis, but are instead assumed to be able to alias
3499 any other type of objects, just like the @code{char} type. See
3500 @option{-fstrict-aliasing} for more information on aliasing issues.
3505 typedef short __attribute__((__may_alias__)) short_a;
3511 short_a *b = (short_a *) &a;
3515 if (a == 0x12345678)
3522 If you replaced @code{short_a} with @code{short} in the variable
3523 declaration, the above program would abort when compiled with
3524 @option{-fstrict-aliasing}, which is on by default at @option{-O2} or
3525 above in recent GCC versions.
3527 @subsection i386 Type Attributes
3529 Two attributes are currently defined for i386 configurations:
3530 @code{ms_struct} and @code{gcc_struct}
3534 @cindex @code{ms_struct}
3535 @cindex @code{gcc_struct}
3537 If @code{packed} is used on a structure, or if bit-fields are used
3538 it may be that the Microsoft ABI packs them differently
3539 than GCC would normally pack them. Particularly when moving packed
3540 data between functions compiled with GCC and the native Microsoft compiler
3541 (either via function call or as data in a file), it may be necessary to access
3544 Currently @option{-m[no-]ms-bitfields} is provided for the Windows X86
3545 compilers to match the native Microsoft compiler.
3548 To specify multiple attributes, separate them by commas within the
3549 double parentheses: for example, @samp{__attribute__ ((aligned (16),
3553 @section An Inline Function is As Fast As a Macro
3554 @cindex inline functions
3555 @cindex integrating function code
3557 @cindex macros, inline alternative
3559 By declaring a function @code{inline}, you can direct GCC to
3560 integrate that function's code into the code for its callers. This
3561 makes execution faster by eliminating the function-call overhead; in
3562 addition, if any of the actual argument values are constant, their known
3563 values may permit simplifications at compile time so that not all of the
3564 inline function's code needs to be included. The effect on code size is
3565 less predictable; object code may be larger or smaller with function
3566 inlining, depending on the particular case. Inlining of functions is an
3567 optimization and it really ``works'' only in optimizing compilation. If
3568 you don't use @option{-O}, no function is really inline.
3570 Inline functions are included in the ISO C99 standard, but there are
3571 currently substantial differences between what GCC implements and what
3572 the ISO C99 standard requires.
3574 To declare a function inline, use the @code{inline} keyword in its
3575 declaration, like this:
3585 (If you are writing a header file to be included in ISO C programs, write
3586 @code{__inline__} instead of @code{inline}. @xref{Alternate Keywords}.)
3587 You can also make all ``simple enough'' functions inline with the option
3588 @option{-finline-functions}.
3591 Note that certain usages in a function definition can make it unsuitable
3592 for inline substitution. Among these usages are: use of varargs, use of
3593 alloca, use of variable sized data types (@pxref{Variable Length}),
3594 use of computed goto (@pxref{Labels as Values}), use of nonlocal goto,
3595 and nested functions (@pxref{Nested Functions}). Using @option{-Winline}
3596 will warn when a function marked @code{inline} could not be substituted,
3597 and will give the reason for the failure.
3599 Note that in C and Objective-C, unlike C++, the @code{inline} keyword
3600 does not affect the linkage of the function.
3602 @cindex automatic @code{inline} for C++ member fns
3603 @cindex @code{inline} automatic for C++ member fns
3604 @cindex member fns, automatically @code{inline}
3605 @cindex C++ member fns, automatically @code{inline}
3606 @opindex fno-default-inline
3607 GCC automatically inlines member functions defined within the class
3608 body of C++ programs even if they are not explicitly declared
3609 @code{inline}. (You can override this with @option{-fno-default-inline};
3610 @pxref{C++ Dialect Options,,Options Controlling C++ Dialect}.)
3612 @cindex inline functions, omission of
3613 @opindex fkeep-inline-functions
3614 When a function is both inline and @code{static}, if all calls to the
3615 function are integrated into the caller, and the function's address is
3616 never used, then the function's own assembler code is never referenced.
3617 In this case, GCC does not actually output assembler code for the
3618 function, unless you specify the option @option{-fkeep-inline-functions}.
3619 Some calls cannot be integrated for various reasons (in particular,
3620 calls that precede the function's definition cannot be integrated, and
3621 neither can recursive calls within the definition). If there is a
3622 nonintegrated call, then the function is compiled to assembler code as
3623 usual. The function must also be compiled as usual if the program
3624 refers to its address, because that can't be inlined.
3626 @cindex non-static inline function
3627 When an inline function is not @code{static}, then the compiler must assume
3628 that there may be calls from other source files; since a global symbol can
3629 be defined only once in any program, the function must not be defined in
3630 the other source files, so the calls therein cannot be integrated.
3631 Therefore, a non-@code{static} inline function is always compiled on its
3632 own in the usual fashion.
3634 If you specify both @code{inline} and @code{extern} in the function
3635 definition, then the definition is used only for inlining. In no case
3636 is the function compiled on its own, not even if you refer to its
3637 address explicitly. Such an address becomes an external reference, as
3638 if you had only declared the function, and had not defined it.
3640 This combination of @code{inline} and @code{extern} has almost the
3641 effect of a macro. The way to use it is to put a function definition in
3642 a header file with these keywords, and put another copy of the
3643 definition (lacking @code{inline} and @code{extern}) in a library file.
3644 The definition in the header file will cause most calls to the function
3645 to be inlined. If any uses of the function remain, they will refer to
3646 the single copy in the library.
3648 For future compatibility with when GCC implements ISO C99 semantics for
3649 inline functions, it is best to use @code{static inline} only. (The
3650 existing semantics will remain available when @option{-std=gnu89} is
3651 specified, but eventually the default will be @option{-std=gnu99} and
3652 that will implement the C99 semantics, though it does not do so yet.)
3654 GCC does not inline any functions when not optimizing unless you specify
3655 the @samp{always_inline} attribute for the function, like this:
3659 inline void foo (const char) __attribute__((always_inline));
3663 @section Assembler Instructions with C Expression Operands
3664 @cindex extended @code{asm}
3665 @cindex @code{asm} expressions
3666 @cindex assembler instructions
3669 In an assembler instruction using @code{asm}, you can specify the
3670 operands of the instruction using C expressions. This means you need not
3671 guess which registers or memory locations will contain the data you want
3674 You must specify an assembler instruction template much like what
3675 appears in a machine description, plus an operand constraint string for
3678 For example, here is how to use the 68881's @code{fsinx} instruction:
3681 asm ("fsinx %1,%0" : "=f" (result) : "f" (angle));
3685 Here @code{angle} is the C expression for the input operand while
3686 @code{result} is that of the output operand. Each has @samp{"f"} as its
3687 operand constraint, saying that a floating point register is required.
3688 The @samp{=} in @samp{=f} indicates that the operand is an output; all
3689 output operands' constraints must use @samp{=}. The constraints use the
3690 same language used in the machine description (@pxref{Constraints}).
3692 Each operand is described by an operand-constraint string followed by
3693 the C expression in parentheses. A colon separates the assembler
3694 template from the first output operand and another separates the last
3695 output operand from the first input, if any. Commas separate the
3696 operands within each group. The total number of operands is currently
3697 limited to 30; this limitation may be lifted in some future version of
3700 If there are no output operands but there are input operands, you must
3701 place two consecutive colons surrounding the place where the output
3704 As of GCC version 3.1, it is also possible to specify input and output
3705 operands using symbolic names which can be referenced within the
3706 assembler code. These names are specified inside square brackets
3707 preceding the constraint string, and can be referenced inside the
3708 assembler code using @code{%[@var{name}]} instead of a percentage sign
3709 followed by the operand number. Using named operands the above example
3713 asm ("fsinx %[angle],%[output]"
3714 : [output] "=f" (result)
3715 : [angle] "f" (angle));
3719 Note that the symbolic operand names have no relation whatsoever to
3720 other C identifiers. You may use any name you like, even those of
3721 existing C symbols, but must ensure that no two operands within the same
3722 assembler construct use the same symbolic name.
3724 Output operand expressions must be lvalues; the compiler can check this.
3725 The input operands need not be lvalues. The compiler cannot check
3726 whether the operands have data types that are reasonable for the
3727 instruction being executed. It does not parse the assembler instruction
3728 template and does not know what it means or even whether it is valid
3729 assembler input. The extended @code{asm} feature is most often used for
3730 machine instructions the compiler itself does not know exist. If
3731 the output expression cannot be directly addressed (for example, it is a
3732 bit-field), your constraint must allow a register. In that case, GCC
3733 will use the register as the output of the @code{asm}, and then store
3734 that register into the output.
3736 The ordinary output operands must be write-only; GCC will assume that
3737 the values in these operands before the instruction are dead and need
3738 not be generated. Extended asm supports input-output or read-write
3739 operands. Use the constraint character @samp{+} to indicate such an
3740 operand and list it with the output operands.
3742 When the constraints for the read-write operand (or the operand in which
3743 only some of the bits are to be changed) allows a register, you may, as
3744 an alternative, logically split its function into two separate operands,
3745 one input operand and one write-only output operand. The connection
3746 between them is expressed by constraints which say they need to be in
3747 the same location when the instruction executes. You can use the same C
3748 expression for both operands, or different expressions. For example,
3749 here we write the (fictitious) @samp{combine} instruction with
3750 @code{bar} as its read-only source operand and @code{foo} as its
3751 read-write destination:
3754 asm ("combine %2,%0" : "=r" (foo) : "0" (foo), "g" (bar));
3758 The constraint @samp{"0"} for operand 1 says that it must occupy the
3759 same location as operand 0. A number in constraint is allowed only in
3760 an input operand and it must refer to an output operand.
3762 Only a number in the constraint can guarantee that one operand will be in
3763 the same place as another. The mere fact that @code{foo} is the value
3764 of both operands is not enough to guarantee that they will be in the
3765 same place in the generated assembler code. The following would not
3769 asm ("combine %2,%0" : "=r" (foo) : "r" (foo), "g" (bar));
3772 Various optimizations or reloading could cause operands 0 and 1 to be in
3773 different registers; GCC knows no reason not to do so. For example, the
3774 compiler might find a copy of the value of @code{foo} in one register and
3775 use it for operand 1, but generate the output operand 0 in a different
3776 register (copying it afterward to @code{foo}'s own address). Of course,
3777 since the register for operand 1 is not even mentioned in the assembler
3778 code, the result will not work, but GCC can't tell that.
3780 As of GCC version 3.1, one may write @code{[@var{name}]} instead of
3781 the operand number for a matching constraint. For example:
3784 asm ("cmoveq %1,%2,%[result]"
3785 : [result] "=r"(result)
3786 : "r" (test), "r"(new), "[result]"(old));
3789 Some instructions clobber specific hard registers. To describe this,
3790 write a third colon after the input operands, followed by the names of
3791 the clobbered hard registers (given as strings). Here is a realistic
3792 example for the VAX:
3795 asm volatile ("movc3 %0,%1,%2"
3797 : "g" (from), "g" (to), "g" (count)
3798 : "r0", "r1", "r2", "r3", "r4", "r5");
3801 You may not write a clobber description in a way that overlaps with an
3802 input or output operand. For example, you may not have an operand
3803 describing a register class with one member if you mention that register
3804 in the clobber list. Variables declared to live in specific registers
3805 (@pxref{Explicit Reg Vars}), and used as asm input or output operands must
3806 have no part mentioned in the clobber description.
3807 There is no way for you to specify that an input
3808 operand is modified without also specifying it as an output
3809 operand. Note that if all the output operands you specify are for this
3810 purpose (and hence unused), you will then also need to specify
3811 @code{volatile} for the @code{asm} construct, as described below, to
3812 prevent GCC from deleting the @code{asm} statement as unused.
3814 If you refer to a particular hardware register from the assembler code,
3815 you will probably have to list the register after the third colon to
3816 tell the compiler the register's value is modified. In some assemblers,
3817 the register names begin with @samp{%}; to produce one @samp{%} in the
3818 assembler code, you must write @samp{%%} in the input.
3820 If your assembler instruction can alter the condition code register, add
3821 @samp{cc} to the list of clobbered registers. GCC on some machines
3822 represents the condition codes as a specific hardware register;
3823 @samp{cc} serves to name this register. On other machines, the
3824 condition code is handled differently, and specifying @samp{cc} has no
3825 effect. But it is valid no matter what the machine.
3827 If your assembler instruction modifies memory in an unpredictable
3828 fashion, add @samp{memory} to the list of clobbered registers. This
3829 will cause GCC to not keep memory values cached in registers across
3830 the assembler instruction. You will also want to add the
3831 @code{volatile} keyword if the memory affected is not listed in the
3832 inputs or outputs of the @code{asm}, as the @samp{memory} clobber does
3833 not count as a side-effect of the @code{asm}.
3835 You can put multiple assembler instructions together in a single
3836 @code{asm} template, separated by the characters normally used in assembly
3837 code for the system. A combination that works in most places is a newline
3838 to break the line, plus a tab character to move to the instruction field
3839 (written as @samp{\n\t}). Sometimes semicolons can be used, if the
3840 assembler allows semicolons as a line-breaking character. Note that some
3841 assembler dialects use semicolons to start a comment.
3842 The input operands are guaranteed not to use any of the clobbered
3843 registers, and neither will the output operands' addresses, so you can
3844 read and write the clobbered registers as many times as you like. Here
3845 is an example of multiple instructions in a template; it assumes the
3846 subroutine @code{_foo} accepts arguments in registers 9 and 10:
3849 asm ("movl %0,r9\n\tmovl %1,r10\n\tcall _foo"
3851 : "g" (from), "g" (to)
3855 Unless an output operand has the @samp{&} constraint modifier, GCC
3856 may allocate it in the same register as an unrelated input operand, on
3857 the assumption the inputs are consumed before the outputs are produced.
3858 This assumption may be false if the assembler code actually consists of
3859 more than one instruction. In such a case, use @samp{&} for each output
3860 operand that may not overlap an input. @xref{Modifiers}.
3862 If you want to test the condition code produced by an assembler
3863 instruction, you must include a branch and a label in the @code{asm}
3864 construct, as follows:
3867 asm ("clr %0\n\tfrob %1\n\tbeq 0f\n\tmov #1,%0\n0:"
3873 This assumes your assembler supports local labels, as the GNU assembler
3874 and most Unix assemblers do.
3876 Speaking of labels, jumps from one @code{asm} to another are not
3877 supported. The compiler's optimizers do not know about these jumps, and
3878 therefore they cannot take account of them when deciding how to
3881 @cindex macros containing @code{asm}
3882 Usually the most convenient way to use these @code{asm} instructions is to
3883 encapsulate them in macros that look like functions. For example,
3887 (@{ double __value, __arg = (x); \
3888 asm ("fsinx %1,%0": "=f" (__value): "f" (__arg)); \
3893 Here the variable @code{__arg} is used to make sure that the instruction
3894 operates on a proper @code{double} value, and to accept only those
3895 arguments @code{x} which can convert automatically to a @code{double}.
3897 Another way to make sure the instruction operates on the correct data
3898 type is to use a cast in the @code{asm}. This is different from using a
3899 variable @code{__arg} in that it converts more different types. For
3900 example, if the desired type were @code{int}, casting the argument to
3901 @code{int} would accept a pointer with no complaint, while assigning the
3902 argument to an @code{int} variable named @code{__arg} would warn about
3903 using a pointer unless the caller explicitly casts it.
3905 If an @code{asm} has output operands, GCC assumes for optimization
3906 purposes the instruction has no side effects except to change the output
3907 operands. This does not mean instructions with a side effect cannot be
3908 used, but you must be careful, because the compiler may eliminate them
3909 if the output operands aren't used, or move them out of loops, or
3910 replace two with one if they constitute a common subexpression. Also,
3911 if your instruction does have a side effect on a variable that otherwise
3912 appears not to change, the old value of the variable may be reused later
3913 if it happens to be found in a register.
3915 You can prevent an @code{asm} instruction from being deleted, moved
3916 significantly, or combined, by writing the keyword @code{volatile} after
3917 the @code{asm}. For example:
3920 #define get_and_set_priority(new) \
3922 asm volatile ("get_and_set_priority %0, %1" \
3923 : "=g" (__old) : "g" (new)); \
3928 If you write an @code{asm} instruction with no outputs, GCC will know
3929 the instruction has side-effects and will not delete the instruction or
3930 move it outside of loops.
3932 The @code{volatile} keyword indicates that the instruction has
3933 important side-effects. GCC will not delete a volatile @code{asm} if
3934 it is reachable. (The instruction can still be deleted if GCC can
3935 prove that control-flow will never reach the location of the
3936 instruction.) In addition, GCC will not reschedule instructions
3937 across a volatile @code{asm} instruction. For example:
3940 *(volatile int *)addr = foo;
3941 asm volatile ("eieio" : : );
3945 Assume @code{addr} contains the address of a memory mapped device
3946 register. The PowerPC @code{eieio} instruction (Enforce In-order
3947 Execution of I/O) tells the CPU to make sure that the store to that
3948 device register happens before it issues any other I/O@.
3950 Note that even a volatile @code{asm} instruction can be moved in ways
3951 that appear insignificant to the compiler, such as across jump
3952 instructions. You can't expect a sequence of volatile @code{asm}
3953 instructions to remain perfectly consecutive. If you want consecutive
3954 output, use a single @code{asm}. Also, GCC will perform some
3955 optimizations across a volatile @code{asm} instruction; GCC does not
3956 ``forget everything'' when it encounters a volatile @code{asm}
3957 instruction the way some other compilers do.
3959 An @code{asm} instruction without any operands or clobbers (an ``old
3960 style'' @code{asm}) will be treated identically to a volatile
3961 @code{asm} instruction.
3963 It is a natural idea to look for a way to give access to the condition
3964 code left by the assembler instruction. However, when we attempted to
3965 implement this, we found no way to make it work reliably. The problem
3966 is that output operands might need reloading, which would result in
3967 additional following ``store'' instructions. On most machines, these
3968 instructions would alter the condition code before there was time to
3969 test it. This problem doesn't arise for ordinary ``test'' and
3970 ``compare'' instructions because they don't have any output operands.
3972 For reasons similar to those described above, it is not possible to give
3973 an assembler instruction access to the condition code left by previous
3976 If you are writing a header file that should be includable in ISO C
3977 programs, write @code{__asm__} instead of @code{asm}. @xref{Alternate
3980 @subsection i386 floating point asm operands
3982 There are several rules on the usage of stack-like regs in
3983 asm_operands insns. These rules apply only to the operands that are
3988 Given a set of input regs that die in an asm_operands, it is
3989 necessary to know which are implicitly popped by the asm, and
3990 which must be explicitly popped by gcc.
3992 An input reg that is implicitly popped by the asm must be
3993 explicitly clobbered, unless it is constrained to match an
3997 For any input reg that is implicitly popped by an asm, it is
3998 necessary to know how to adjust the stack to compensate for the pop.
3999 If any non-popped input is closer to the top of the reg-stack than
4000 the implicitly popped reg, it would not be possible to know what the
4001 stack looked like---it's not clear how the rest of the stack ``slides
4004 All implicitly popped input regs must be closer to the top of
4005 the reg-stack than any input that is not implicitly popped.
4007 It is possible that if an input dies in an insn, reload might
4008 use the input reg for an output reload. Consider this example:
4011 asm ("foo" : "=t" (a) : "f" (b));
4014 This asm says that input B is not popped by the asm, and that
4015 the asm pushes a result onto the reg-stack, i.e., the stack is one
4016 deeper after the asm than it was before. But, it is possible that
4017 reload will think that it can use the same reg for both the input and
4018 the output, if input B dies in this insn.
4020 If any input operand uses the @code{f} constraint, all output reg
4021 constraints must use the @code{&} earlyclobber.
4023 The asm above would be written as
4026 asm ("foo" : "=&t" (a) : "f" (b));
4030 Some operands need to be in particular places on the stack. All
4031 output operands fall in this category---there is no other way to
4032 know which regs the outputs appear in unless the user indicates
4033 this in the constraints.
4035 Output operands must specifically indicate which reg an output
4036 appears in after an asm. @code{=f} is not allowed: the operand
4037 constraints must select a class with a single reg.
4040 Output operands may not be ``inserted'' between existing stack regs.
4041 Since no 387 opcode uses a read/write operand, all output operands
4042 are dead before the asm_operands, and are pushed by the asm_operands.
4043 It makes no sense to push anywhere but the top of the reg-stack.
4045 Output operands must start at the top of the reg-stack: output
4046 operands may not ``skip'' a reg.
4049 Some asm statements may need extra stack space for internal
4050 calculations. This can be guaranteed by clobbering stack registers
4051 unrelated to the inputs and outputs.
4055 Here are a couple of reasonable asms to want to write. This asm
4056 takes one input, which is internally popped, and produces two outputs.
4059 asm ("fsincos" : "=t" (cos), "=u" (sin) : "0" (inp));
4062 This asm takes two inputs, which are popped by the @code{fyl2xp1} opcode,
4063 and replaces them with one output. The user must code the @code{st(1)}
4064 clobber for reg-stack.c to know that @code{fyl2xp1} pops both inputs.
4067 asm ("fyl2xp1" : "=t" (result) : "0" (x), "u" (y) : "st(1)");
4073 @section Controlling Names Used in Assembler Code
4074 @cindex assembler names for identifiers
4075 @cindex names used in assembler code
4076 @cindex identifiers, names in assembler code
4078 You can specify the name to be used in the assembler code for a C
4079 function or variable by writing the @code{asm} (or @code{__asm__})
4080 keyword after the declarator as follows:
4083 int foo asm ("myfoo") = 2;
4087 This specifies that the name to be used for the variable @code{foo} in
4088 the assembler code should be @samp{myfoo} rather than the usual
4091 On systems where an underscore is normally prepended to the name of a C
4092 function or variable, this feature allows you to define names for the
4093 linker that do not start with an underscore.
4095 It does not make sense to use this feature with a non-static local
4096 variable since such variables do not have assembler names. If you are
4097 trying to put the variable in a particular register, see @ref{Explicit
4098 Reg Vars}. GCC presently accepts such code with a warning, but will
4099 probably be changed to issue an error, rather than a warning, in the
4102 You cannot use @code{asm} in this way in a function @emph{definition}; but
4103 you can get the same effect by writing a declaration for the function
4104 before its definition and putting @code{asm} there, like this:
4107 extern func () asm ("FUNC");
4114 It is up to you to make sure that the assembler names you choose do not
4115 conflict with any other assembler symbols. Also, you must not use a
4116 register name; that would produce completely invalid assembler code. GCC
4117 does not as yet have the ability to store static variables in registers.
4118 Perhaps that will be added.
4120 @node Explicit Reg Vars
4121 @section Variables in Specified Registers
4122 @cindex explicit register variables
4123 @cindex variables in specified registers
4124 @cindex specified registers
4125 @cindex registers, global allocation
4127 GNU C allows you to put a few global variables into specified hardware
4128 registers. You can also specify the register in which an ordinary
4129 register variable should be allocated.
4133 Global register variables reserve registers throughout the program.
4134 This may be useful in programs such as programming language
4135 interpreters which have a couple of global variables that are accessed
4139 Local register variables in specific registers do not reserve the
4140 registers. The compiler's data flow analysis is capable of determining
4141 where the specified registers contain live values, and where they are
4142 available for other uses. Stores into local register variables may be deleted
4143 when they appear to be dead according to dataflow analysis. References
4144 to local register variables may be deleted or moved or simplified.
4146 These local variables are sometimes convenient for use with the extended
4147 @code{asm} feature (@pxref{Extended Asm}), if you want to write one
4148 output of the assembler instruction directly into a particular register.
4149 (This will work provided the register you specify fits the constraints
4150 specified for that operand in the @code{asm}.)
4158 @node Global Reg Vars
4159 @subsection Defining Global Register Variables
4160 @cindex global register variables
4161 @cindex registers, global variables in
4163 You can define a global register variable in GNU C like this:
4166 register int *foo asm ("a5");
4170 Here @code{a5} is the name of the register which should be used. Choose a
4171 register which is normally saved and restored by function calls on your
4172 machine, so that library routines will not clobber it.
4174 Naturally the register name is cpu-dependent, so you would need to
4175 conditionalize your program according to cpu type. The register
4176 @code{a5} would be a good choice on a 68000 for a variable of pointer
4177 type. On machines with register windows, be sure to choose a ``global''
4178 register that is not affected magically by the function call mechanism.
4180 In addition, operating systems on one type of cpu may differ in how they
4181 name the registers; then you would need additional conditionals. For
4182 example, some 68000 operating systems call this register @code{%a5}.
4184 Eventually there may be a way of asking the compiler to choose a register
4185 automatically, but first we need to figure out how it should choose and
4186 how to enable you to guide the choice. No solution is evident.
4188 Defining a global register variable in a certain register reserves that
4189 register entirely for this use, at least within the current compilation.
4190 The register will not be allocated for any other purpose in the functions
4191 in the current compilation. The register will not be saved and restored by
4192 these functions. Stores into this register are never deleted even if they
4193 would appear to be dead, but references may be deleted or moved or
4196 It is not safe to access the global register variables from signal
4197 handlers, or from more than one thread of control, because the system
4198 library routines may temporarily use the register for other things (unless
4199 you recompile them specially for the task at hand).
4201 @cindex @code{qsort}, and global register variables
4202 It is not safe for one function that uses a global register variable to
4203 call another such function @code{foo} by way of a third function
4204 @code{lose} that was compiled without knowledge of this variable (i.e.@: in a
4205 different source file in which the variable wasn't declared). This is
4206 because @code{lose} might save the register and put some other value there.
4207 For example, you can't expect a global register variable to be available in
4208 the comparison-function that you pass to @code{qsort}, since @code{qsort}
4209 might have put something else in that register. (If you are prepared to
4210 recompile @code{qsort} with the same global register variable, you can
4211 solve this problem.)
4213 If you want to recompile @code{qsort} or other source files which do not
4214 actually use your global register variable, so that they will not use that
4215 register for any other purpose, then it suffices to specify the compiler
4216 option @option{-ffixed-@var{reg}}. You need not actually add a global
4217 register declaration to their source code.
4219 A function which can alter the value of a global register variable cannot
4220 safely be called from a function compiled without this variable, because it
4221 could clobber the value the caller expects to find there on return.
4222 Therefore, the function which is the entry point into the part of the
4223 program that uses the global register variable must explicitly save and
4224 restore the value which belongs to its caller.
4226 @cindex register variable after @code{longjmp}
4227 @cindex global register after @code{longjmp}
4228 @cindex value after @code{longjmp}
4231 On most machines, @code{longjmp} will restore to each global register
4232 variable the value it had at the time of the @code{setjmp}. On some
4233 machines, however, @code{longjmp} will not change the value of global
4234 register variables. To be portable, the function that called @code{setjmp}
4235 should make other arrangements to save the values of the global register
4236 variables, and to restore them in a @code{longjmp}. This way, the same
4237 thing will happen regardless of what @code{longjmp} does.
4239 All global register variable declarations must precede all function
4240 definitions. If such a declaration could appear after function
4241 definitions, the declaration would be too late to prevent the register from
4242 being used for other purposes in the preceding functions.
4244 Global register variables may not have initial values, because an
4245 executable file has no means to supply initial contents for a register.
4247 On the SPARC, there are reports that g3 @dots{} g7 are suitable
4248 registers, but certain library functions, such as @code{getwd}, as well
4249 as the subroutines for division and remainder, modify g3 and g4. g1 and
4250 g2 are local temporaries.
4252 On the 68000, a2 @dots{} a5 should be suitable, as should d2 @dots{} d7.
4253 Of course, it will not do to use more than a few of those.
4255 @node Local Reg Vars
4256 @subsection Specifying Registers for Local Variables
4257 @cindex local variables, specifying registers
4258 @cindex specifying registers for local variables
4259 @cindex registers for local variables
4261 You can define a local register variable with a specified register
4265 register int *foo asm ("a5");
4269 Here @code{a5} is the name of the register which should be used. Note
4270 that this is the same syntax used for defining global register
4271 variables, but for a local variable it would appear within a function.
4273 Naturally the register name is cpu-dependent, but this is not a
4274 problem, since specific registers are most often useful with explicit
4275 assembler instructions (@pxref{Extended Asm}). Both of these things
4276 generally require that you conditionalize your program according to
4279 In addition, operating systems on one type of cpu may differ in how they
4280 name the registers; then you would need additional conditionals. For
4281 example, some 68000 operating systems call this register @code{%a5}.
4283 Defining such a register variable does not reserve the register; it
4284 remains available for other uses in places where flow control determines
4285 the variable's value is not live. However, these registers are made
4286 unavailable for use in the reload pass; excessive use of this feature
4287 leaves the compiler too few available registers to compile certain
4290 This option does not guarantee that GCC will generate code that has
4291 this variable in the register you specify at all times. You may not
4292 code an explicit reference to this register in an @code{asm} statement
4293 and assume it will always refer to this variable.
4295 Stores into local register variables may be deleted when they appear to be dead
4296 according to dataflow analysis. References to local register variables may
4297 be deleted or moved or simplified.
4299 @node Alternate Keywords
4300 @section Alternate Keywords
4301 @cindex alternate keywords
4302 @cindex keywords, alternate
4304 @option{-ansi} and the various @option{-std} options disable certain
4305 keywords. This causes trouble when you want to use GNU C extensions, or
4306 a general-purpose header file that should be usable by all programs,
4307 including ISO C programs. The keywords @code{asm}, @code{typeof} and
4308 @code{inline} are not available in programs compiled with
4309 @option{-ansi} or @option{-std} (although @code{inline} can be used in a
4310 program compiled with @option{-std=c99}). The ISO C99 keyword
4311 @code{restrict} is only available when @option{-std=gnu99} (which will
4312 eventually be the default) or @option{-std=c99} (or the equivalent
4313 @option{-std=iso9899:1999}) is used.
4315 The way to solve these problems is to put @samp{__} at the beginning and
4316 end of each problematical keyword. For example, use @code{__asm__}
4317 instead of @code{asm}, and @code{__inline__} instead of @code{inline}.
4319 Other C compilers won't accept these alternative keywords; if you want to
4320 compile with another compiler, you can define the alternate keywords as
4321 macros to replace them with the customary keywords. It looks like this:
4329 @findex __extension__
4331 @option{-pedantic} and other options cause warnings for many GNU C extensions.
4333 prevent such warnings within one expression by writing
4334 @code{__extension__} before the expression. @code{__extension__} has no
4335 effect aside from this.
4337 @node Incomplete Enums
4338 @section Incomplete @code{enum} Types
4340 You can define an @code{enum} tag without specifying its possible values.
4341 This results in an incomplete type, much like what you get if you write
4342 @code{struct foo} without describing the elements. A later declaration
4343 which does specify the possible values completes the type.
4345 You can't allocate variables or storage using the type while it is
4346 incomplete. However, you can work with pointers to that type.
4348 This extension may not be very useful, but it makes the handling of
4349 @code{enum} more consistent with the way @code{struct} and @code{union}
4352 This extension is not supported by GNU C++.
4354 @node Function Names
4355 @section Function Names as Strings
4356 @cindex @code{__FUNCTION__} identifier
4357 @cindex @code{__PRETTY_FUNCTION__} identifier
4358 @cindex @code{__func__} identifier
4360 GCC predefines two magic identifiers to hold the name of the current
4361 function. The identifier @code{__FUNCTION__} holds the name of the function
4362 as it appears in the source. The identifier @code{__PRETTY_FUNCTION__}
4363 holds the name of the function pretty printed in a language specific
4366 These names are always the same in a C function, but in a C++ function
4367 they may be different. For example, this program:
4371 extern int printf (char *, ...);
4378 printf ("__FUNCTION__ = %s\n", __FUNCTION__);
4379 printf ("__PRETTY_FUNCTION__ = %s\n", __PRETTY_FUNCTION__);
4397 __PRETTY_FUNCTION__ = int a::sub (int)
4400 The compiler automagically replaces the identifiers with a string
4401 literal containing the appropriate name. Thus, they are neither
4402 preprocessor macros, like @code{__FILE__} and @code{__LINE__}, nor
4403 variables. This means that they catenate with other string literals, and
4404 that they can be used to initialize char arrays. For example
4407 char here[] = "Function " __FUNCTION__ " in " __FILE__;
4410 On the other hand, @samp{#ifdef __FUNCTION__} does not have any special
4411 meaning inside a function, since the preprocessor does not do anything
4412 special with the identifier @code{__FUNCTION__}.
4414 Note that these semantics are deprecated, and that GCC 3.2 will handle
4415 @code{__FUNCTION__} and @code{__PRETTY_FUNCTION__} the same way as
4416 @code{__func__}. @code{__func__} is defined by the ISO standard C99:
4419 The identifier @code{__func__} is implicitly declared by the translator
4420 as if, immediately following the opening brace of each function
4421 definition, the declaration
4424 static const char __func__[] = "function-name";
4427 appeared, where function-name is the name of the lexically-enclosing
4428 function. This name is the unadorned name of the function.
4431 By this definition, @code{__func__} is a variable, not a string literal.
4432 In particular, @code{__func__} does not catenate with other string
4435 In @code{C++}, @code{__FUNCTION__} and @code{__PRETTY_FUNCTION__} are
4436 variables, declared in the same way as @code{__func__}.
4438 @node Return Address
4439 @section Getting the Return or Frame Address of a Function
4441 These functions may be used to get information about the callers of a
4444 @deftypefn {Built-in Function} {void *} __builtin_return_address (unsigned int @var{level})
4445 This function returns the return address of the current function, or of
4446 one of its callers. The @var{level} argument is number of frames to
4447 scan up the call stack. A value of @code{0} yields the return address
4448 of the current function, a value of @code{1} yields the return address
4449 of the caller of the current function, and so forth. When inlining
4450 the expected behavior is that the function will return the address of
4451 the function that will be returned to. To work around this behavior use
4452 the @code{noinline} function attribute.
4454 The @var{level} argument must be a constant integer.
4456 On some machines it may be impossible to determine the return address of
4457 any function other than the current one; in such cases, or when the top
4458 of the stack has been reached, this function will return @code{0} or a
4459 random value. In addition, @code{__builtin_frame_address} may be used
4460 to determine if the top of the stack has been reached.
4462 This function should only be used with a nonzero argument for debugging
4466 @deftypefn {Built-in Function} {void *} __builtin_frame_address (unsigned int @var{level})
4467 This function is similar to @code{__builtin_return_address}, but it
4468 returns the address of the function frame rather than the return address
4469 of the function. Calling @code{__builtin_frame_address} with a value of
4470 @code{0} yields the frame address of the current function, a value of
4471 @code{1} yields the frame address of the caller of the current function,
4474 The frame is the area on the stack which holds local variables and saved
4475 registers. The frame address is normally the address of the first word
4476 pushed on to the stack by the function. However, the exact definition
4477 depends upon the processor and the calling convention. If the processor
4478 has a dedicated frame pointer register, and the function has a frame,
4479 then @code{__builtin_frame_address} will return the value of the frame
4482 On some machines it may be impossible to determine the frame address of
4483 any function other than the current one; in such cases, or when the top
4484 of the stack has been reached, this function will return @code{0} if
4485 the first frame pointer is properly initialized by the startup code.
4487 This function should only be used with a nonzero argument for debugging
4491 @node Vector Extensions
4492 @section Using vector instructions through built-in functions
4494 On some targets, the instruction set contains SIMD vector instructions that
4495 operate on multiple values contained in one large register at the same time.
4496 For example, on the i386 the MMX, 3Dnow! and SSE extensions can be used
4499 The first step in using these extensions is to provide the necessary data
4500 types. This should be done using an appropriate @code{typedef}:
4503 typedef int v4si __attribute__ ((mode(V4SI)));
4506 The base type @code{int} is effectively ignored by the compiler, the
4507 actual properties of the new type @code{v4si} are defined by the
4508 @code{__attribute__}. It defines the machine mode to be used; for vector
4509 types these have the form @code{V@var{n}@var{B}}; @var{n} should be the
4510 number of elements in the vector, and @var{B} should be the base mode of the
4511 individual elements. The following can be used as base modes:
4515 An integer that is as wide as the smallest addressable unit, usually 8 bits.
4517 An integer, twice as wide as a QI mode integer, usually 16 bits.
4519 An integer, four times as wide as a QI mode integer, usually 32 bits.
4521 An integer, eight times as wide as a QI mode integer, usually 64 bits.
4523 A floating point value, as wide as a SI mode integer, usually 32 bits.
4525 A floating point value, as wide as a DI mode integer, usually 64 bits.
4528 Specifying a combination that is not valid for the current architecture
4529 will cause gcc to synthesize the instructions using a narrower mode.
4530 For example, if you specify a variable of type @code{V4SI} and your
4531 architecture does not allow for this specific SIMD type, gcc will
4532 produce code that uses 4 @code{SIs}.
4534 The types defined in this manner can be used with a subset of normal C
4535 operations. Currently, gcc will allow using the following operators on
4536 these types: @code{+, -, *, /, unary minus}@.
4538 The operations behave like C++ @code{valarrays}. Addition is defined as
4539 the addition of the corresponding elements of the operands. For
4540 example, in the code below, each of the 4 elements in @var{a} will be
4541 added to the corresponding 4 elements in @var{b} and the resulting
4542 vector will be stored in @var{c}.
4545 typedef int v4si __attribute__ ((mode(V4SI)));
4552 Subtraction, multiplication, and division operate in a similar manner.
4553 Likewise, the result of using the unary minus operator on a vector type
4554 is a vector whose elements are the negative value of the corresponding
4555 elements in the operand.
4557 You can declare variables and use them in function calls and returns, as
4558 well as in assignments and some casts. You can specify a vector type as
4559 a return type for a function. Vector types can also be used as function
4560 arguments. It is possible to cast from one vector type to another,
4561 provided they are of the same size (in fact, you can also cast vectors
4562 to and from other datatypes of the same size).
4564 You cannot operate between vectors of different lengths or different
4565 signedness without a cast.
4567 A port that supports hardware vector operations, usually provides a set
4568 of built-in functions that can be used to operate on vectors. For
4569 example, a function to add two vectors and multiply the result by a
4570 third could look like this:
4573 v4si f (v4si a, v4si b, v4si c)
4575 v4si tmp = __builtin_addv4si (a, b);
4576 return __builtin_mulv4si (tmp, c);
4581 @node Other Builtins
4582 @section Other built-in functions provided by GCC
4583 @cindex built-in functions
4584 @findex __builtin_isgreater
4585 @findex __builtin_isgreaterequal
4586 @findex __builtin_isless
4587 @findex __builtin_islessequal
4588 @findex __builtin_islessgreater
4589 @findex __builtin_isunordered
4631 @findex fprintf_unlocked
4633 @findex fputs_unlocked
4653 @findex printf_unlocked
4694 GCC provides a large number of built-in functions other than the ones
4695 mentioned above. Some of these are for internal use in the processing
4696 of exceptions or variable-length argument lists and will not be
4697 documented here because they may change from time to time; we do not
4698 recommend general use of these functions.
4700 The remaining functions are provided for optimization purposes.
4702 @opindex fno-builtin
4703 GCC includes built-in versions of many of the functions in the standard
4704 C library. The versions prefixed with @code{__builtin_} will always be
4705 treated as having the same meaning as the C library function even if you
4706 specify the @option{-fno-builtin} option. (@pxref{C Dialect Options})
4707 Many of these functions are only optimized in certain cases; if they are
4708 not optimized in a particular case, a call to the library function will
4713 Outside strict ISO C mode (@option{-ansi}, @option{-std=c89} or
4714 @option{-std=c99}), the functions @code{alloca}, @code{bcmp},
4715 @code{bzero}, @code{_exit}, @code{ffs}, @code{fprintf_unlocked},
4716 @code{fputs_unlocked}, @code{index}, @code{mempcpy}, @code{printf_unlocked},
4717 @code{rindex}, @code{stpcpy} and @code{strdup}
4718 may be handled as built-in functions.
4719 All these functions have corresponding versions
4720 prefixed with @code{__builtin_}, which may be used even in strict C89
4723 The ISO C99 functions @code{conj}, @code{conjf}, @code{conjl}, @code{creal},
4724 @code{crealf}, @code{creall}, @code{cimag}, @code{cimagf}, @code{cimagl},
4725 @code{_Exit}, @code{imaxabs}, @code{llabs},
4726 @code{nearbyint}, @code{nearbyintf}, @code{nearbyintl},
4727 @code{round}, @code{roundf}, @code{roundl}, @code{snprintf},
4728 @code{trunc}, @code{truncf}, @code{truncl},
4729 @code{vscanf}, @code{vsnprintf} and @code{vsscanf}
4730 are handled as built-in functions
4731 except in strict ISO C90 mode (@option{-ansi} or @option{-std=c89}).
4733 There are also built-in versions of the ISO C99 functions @code{atan2f},
4734 @code{atan2l}, @code{ceilf}, @code{ceill}, @code{cosf}, @code{cosl},
4735 @code{expf}, @code{expl}, @code{fabsf}, @code{fabsl}, @code{floorf},
4736 @code{floorl}, @code{fmodf}, @code{fmodl},
4737 @code{logf}, @code{logl}, @code{powf}, @code{powl},
4738 @code{sinf}, @code{sinl}, @code{sqrtf} and @code{sqrtl}
4739 that are recognized in any mode since ISO C90 reserves these names for
4740 the purpose to which ISO C99 puts them. All these functions have
4741 corresponding versions prefixed with @code{__builtin_}.
4743 The ISO C90 functions @code{abort}, @code{abs}, @code{atan2},
4744 @code{calloc}, @code{ceil}, @code{cos}, @code{exit},
4745 @code{exp}, @code{fabs}, @code{floor}, @code{fmod},
4746 @code{fprintf}, @code{fputs}, @code{labs}, @code{log}, @code{malloc},
4747 @code{memcmp}, @code{memcpy}, @code{memset}, @code{pow}, @code{printf},
4748 @code{putchar}, @code{puts}, @code{scanf}, @code{sin}, @code{snprintf},
4749 @code{sprintf}, @code{sqrt}, @code{sscanf},
4750 @code{strcat}, @code{strchr}, @code{strcmp},
4751 @code{strcpy}, @code{strcspn}, @code{strlen}, @code{strncat}, @code{strncmp},
4752 @code{strncpy}, @code{strpbrk}, @code{strrchr}, @code{strspn}, @code{strstr},
4753 @code{vprintf} and @code{vsprintf}
4754 are all recognized as built-in functions unless
4755 @option{-fno-builtin} is specified (or @option{-fno-builtin-@var{function}}
4756 is specified for an individual function). All of these functions have
4757 corresponding versions prefixed with @code{__builtin_}.
4759 GCC provides built-in versions of the ISO C99 floating point comparison
4760 macros that avoid raising exceptions for unordered operands. They have
4761 the same names as the standard macros ( @code{isgreater},
4762 @code{isgreaterequal}, @code{isless}, @code{islessequal},
4763 @code{islessgreater}, and @code{isunordered}) , with @code{__builtin_}
4764 prefixed. We intend for a library implementor to be able to simply
4765 @code{#define} each standard macro to its built-in equivalent.
4767 @deftypefn {Built-in Function} int __builtin_types_compatible_p (@var{type1}, @var{type2})
4769 You can use the built-in function @code{__builtin_types_compatible_p} to
4770 determine whether two types are the same.
4772 This built-in function returns 1 if the unqualified versions of the
4773 types @var{type1} and @var{type2} (which are types, not expressions) are
4774 compatible, 0 otherwise. The result of this built-in function can be
4775 used in integer constant expressions.
4777 This built-in function ignores top level qualifiers (e.g., @code{const},
4778 @code{volatile}). For example, @code{int} is equivalent to @code{const
4781 The type @code{int[]} and @code{int[5]} are compatible. On the other
4782 hand, @code{int} and @code{char *} are not compatible, even if the size
4783 of their types, on the particular architecture are the same. Also, the
4784 amount of pointer indirection is taken into account when determining
4785 similarity. Consequently, @code{short *} is not similar to
4786 @code{short **}. Furthermore, two types that are typedefed are
4787 considered compatible if their underlying types are compatible.
4789 An @code{enum} type is considered to be compatible with another
4790 @code{enum} type. For example, @code{enum @{foo, bar@}} is similar to
4791 @code{enum @{hot, dog@}}.
4793 You would typically use this function in code whose execution varies
4794 depending on the arguments' types. For example:
4800 if (__builtin_types_compatible_p (typeof (x), long double)) \
4801 tmp = foo_long_double (tmp); \
4802 else if (__builtin_types_compatible_p (typeof (x), double)) \
4803 tmp = foo_double (tmp); \
4804 else if (__builtin_types_compatible_p (typeof (x), float)) \
4805 tmp = foo_float (tmp); \
4812 @emph{Note:} This construct is only available for C.
4816 @deftypefn {Built-in Function} @var{type} __builtin_choose_expr (@var{const_exp}, @var{exp1}, @var{exp2})
4818 You can use the built-in function @code{__builtin_choose_expr} to
4819 evaluate code depending on the value of a constant expression. This
4820 built-in function returns @var{exp1} if @var{const_exp}, which is a
4821 constant expression that must be able to be determined at compile time,
4822 is nonzero. Otherwise it returns 0.
4824 This built-in function is analogous to the @samp{? :} operator in C,
4825 except that the expression returned has its type unaltered by promotion
4826 rules. Also, the built-in function does not evaluate the expression
4827 that was not chosen. For example, if @var{const_exp} evaluates to true,
4828 @var{exp2} is not evaluated even if it has side-effects.
4830 This built-in function can return an lvalue if the chosen argument is an
4833 If @var{exp1} is returned, the return type is the same as @var{exp1}'s
4834 type. Similarly, if @var{exp2} is returned, its return type is the same
4841 __builtin_choose_expr ( \
4842 __builtin_types_compatible_p (typeof (x), double), \
4844 __builtin_choose_expr ( \
4845 __builtin_types_compatible_p (typeof (x), float), \
4847 /* @r{The void expression results in a compile-time error} \
4848 @r{when assigning the result to something.} */ \
4852 @emph{Note:} This construct is only available for C. Furthermore, the
4853 unused expression (@var{exp1} or @var{exp2} depending on the value of
4854 @var{const_exp}) may still generate syntax errors. This may change in
4859 @deftypefn {Built-in Function} int __builtin_constant_p (@var{exp})
4860 You can use the built-in function @code{__builtin_constant_p} to
4861 determine if a value is known to be constant at compile-time and hence
4862 that GCC can perform constant-folding on expressions involving that
4863 value. The argument of the function is the value to test. The function
4864 returns the integer 1 if the argument is known to be a compile-time
4865 constant and 0 if it is not known to be a compile-time constant. A
4866 return of 0 does not indicate that the value is @emph{not} a constant,
4867 but merely that GCC cannot prove it is a constant with the specified
4868 value of the @option{-O} option.
4870 You would typically use this function in an embedded application where
4871 memory was a critical resource. If you have some complex calculation,
4872 you may want it to be folded if it involves constants, but need to call
4873 a function if it does not. For example:
4876 #define Scale_Value(X) \
4877 (__builtin_constant_p (X) \
4878 ? ((X) * SCALE + OFFSET) : Scale (X))
4881 You may use this built-in function in either a macro or an inline
4882 function. However, if you use it in an inlined function and pass an
4883 argument of the function as the argument to the built-in, GCC will
4884 never return 1 when you call the inline function with a string constant
4885 or compound literal (@pxref{Compound Literals}) and will not return 1
4886 when you pass a constant numeric value to the inline function unless you
4887 specify the @option{-O} option.
4889 You may also use @code{__builtin_constant_p} in initializers for static
4890 data. For instance, you can write
4893 static const int table[] = @{
4894 __builtin_constant_p (EXPRESSION) ? (EXPRESSION) : -1,
4900 This is an acceptable initializer even if @var{EXPRESSION} is not a
4901 constant expression. GCC must be more conservative about evaluating the
4902 built-in in this case, because it has no opportunity to perform
4905 Previous versions of GCC did not accept this built-in in data
4906 initializers. The earliest version where it is completely safe is
4910 @deftypefn {Built-in Function} long __builtin_expect (long @var{exp}, long @var{c})
4911 @opindex fprofile-arcs
4912 You may use @code{__builtin_expect} to provide the compiler with
4913 branch prediction information. In general, you should prefer to
4914 use actual profile feedback for this (@option{-fprofile-arcs}), as
4915 programmers are notoriously bad at predicting how their programs
4916 actually perform. However, there are applications in which this
4917 data is hard to collect.
4919 The return value is the value of @var{exp}, which should be an
4920 integral expression. The value of @var{c} must be a compile-time
4921 constant. The semantics of the built-in are that it is expected
4922 that @var{exp} == @var{c}. For example:
4925 if (__builtin_expect (x, 0))
4930 would indicate that we do not expect to call @code{foo}, since
4931 we expect @code{x} to be zero. Since you are limited to integral
4932 expressions for @var{exp}, you should use constructions such as
4935 if (__builtin_expect (ptr != NULL, 1))
4940 when testing pointer or floating-point values.
4943 @deftypefn {Built-in Function} void __builtin_prefetch (const void *@var{addr}, ...)
4944 This function is used to minimize cache-miss latency by moving data into
4945 a cache before it is accessed.
4946 You can insert calls to @code{__builtin_prefetch} into code for which
4947 you know addresses of data in memory that is likely to be accessed soon.
4948 If the target supports them, data prefetch instructions will be generated.
4949 If the prefetch is done early enough before the access then the data will
4950 be in the cache by the time it is accessed.
4952 The value of @var{addr} is the address of the memory to prefetch.
4953 There are two optional arguments, @var{rw} and @var{locality}.
4954 The value of @var{rw} is a compile-time constant one or zero; one
4955 means that the prefetch is preparing for a write to the memory address
4956 and zero, the default, means that the prefetch is preparing for a read.
4957 The value @var{locality} must be a compile-time constant integer between
4958 zero and three. A value of zero means that the data has no temporal
4959 locality, so it need not be left in the cache after the access. A value
4960 of three means that the data has a high degree of temporal locality and
4961 should be left in all levels of cache possible. Values of one and two
4962 mean, respectively, a low or moderate degree of temporal locality. The
4966 for (i = 0; i < n; i++)
4969 __builtin_prefetch (&a[i+j], 1, 1);
4970 __builtin_prefetch (&b[i+j], 0, 1);
4975 Data prefetch does not generate faults if @var{addr} is invalid, but
4976 the address expression itself must be valid. For example, a prefetch
4977 of @code{p->next} will not fault if @code{p->next} is not a valid
4978 address, but evaluation will fault if @code{p} is not a valid address.
4980 If the target does not support data prefetch, the address expression
4981 is evaluated if it includes side effects but no other code is generated
4982 and GCC does not issue a warning.
4985 @deftypefn {Built-in Function} double __builtin_huge_val (void)
4986 Returns a positive infinity, if supported by the floating-point format,
4987 else @code{DBL_MAX}. This function is suitable for implementing the
4988 ISO C macro @code{HUGE_VAL}.
4991 @deftypefn {Built-in Function} float __builtin_huge_valf (void)
4992 Similar to @code{__builtin_huge_val}, except the return type is @code{float}.
4995 @deftypefn {Built-in Function} {long double} __builtin_huge_vall (void)
4996 Similar to @code{__builtin_huge_val}, except the return
4997 type is @code{long double}.
5000 @deftypefn {Built-in Function} double __builtin_inf (void)
5001 Similar to @code{__builtin_huge_val}, except a warning is generated
5002 if the target floating-point format does not support infinities.
5003 This function is suitable for implementing the ISO C99 macro @code{INFINITY}.
5006 @deftypefn {Built-in Function} float __builtin_inff (void)
5007 Similar to @code{__builtin_inf}, except the return type is @code{float}.
5010 @deftypefn {Built-in Function} {long double} __builtin_infl (void)
5011 Similar to @code{__builtin_inf}, except the return
5012 type is @code{long double}.
5015 @deftypefn {Built-in Function} double __builtin_nan (const char *str)
5016 This is an implementation of the ISO C99 function @code{nan}.
5018 Since ISO C99 defines this function in terms of @code{strtod}, which we
5019 do not implement, a description of the parsing is in order. The string
5020 is parsed as by @code{strtol}; that is, the base is recognized by
5021 leading @samp{0} or @samp{0x} prefixes. The number parsed is placed
5022 in the significand such that the least significant bit of the number
5023 is at the least significant bit of the significand. The number is
5024 truncated to fit the significand field provided. The significand is
5025 forced to be a quiet NaN.
5027 This function, if given a string literal, is evaluated early enough
5028 that it is considered a compile-time constant.
5031 @deftypefn {Built-in Function} float __builtin_nanf (const char *str)
5032 Similar to @code{__builtin_nan}, except the return type is @code{float}.
5035 @deftypefn {Built-in Function} {long double} __builtin_nanl (const char *str)
5036 Similar to @code{__builtin_nan}, except the return type is @code{long double}.
5039 @deftypefn {Built-in Function} double __builtin_nans (const char *str)
5040 Similar to @code{__builtin_nan}, except the significand is forced
5041 to be a signaling NaN. The @code{nans} function is proposed by
5042 @uref{http://std.dkuug.dk/JTC1/SC22/WG14/www/docs/n965.htm,,WG14 N965}.
5045 @deftypefn {Built-in Function} float __builtin_nansf (const char *str)
5046 Similar to @code{__builtin_nans}, except the return type is @code{float}.
5049 @deftypefn {Built-in Function} {long double} __builtin_nansl (const char *str)
5050 Similar to @code{__builtin_nans}, except the return type is @code{long double}.
5053 @deftypefn {Built-in Function} int __builtin_ffs (unsigned int x)
5054 Returns one plus the index of the least significant 1-bit of @var{x}, or
5055 if @var{x} is zero, returns zero.
5058 @deftypefn {Built-in Function} int __builtin_clz (unsigned int x)
5059 Returns the number of leading 0-bits in @var{x}, starting at the most
5060 significant bit position. If @var{x} is 0, the result is undefined.
5063 @deftypefn {Built-in Function} int __builtin_ctz (unsigned int x)
5064 Returns the number of trailing 0-bits in @var{x}, starting at the least
5065 significant bit position. If @var{x} is 0, the result is undefined.
5068 @deftypefn {Built-in Function} int __builtin_popcount (unsigned int x)
5069 Returns the number of 1-bits in @var{x}.
5072 @deftypefn {Built-in Function} int __builtin_parity (unsigned int x)
5073 Returns the parity of @var{x}, i.@:e. the number of 1-bits in @var{x}
5077 @deftypefn {Built-in Function} int __builtin_ffsl (unsigned long)
5078 Similar to @code{__builtin_ffs}, except the argument type is
5079 @code{unsigned long}.
5082 @deftypefn {Built-in Function} int __builtin_clzl (unsigned long)
5083 Similar to @code{__builtin_clz}, except the argument type is
5084 @code{unsigned long}.
5087 @deftypefn {Built-in Function} int __builtin_ctzl (unsigned long)
5088 Similar to @code{__builtin_ctz}, except the argument type is
5089 @code{unsigned long}.
5092 @deftypefn {Built-in Function} int __builtin_popcountl (unsigned long)
5093 Similar to @code{__builtin_popcount}, except the argument type is
5094 @code{unsigned long}.
5097 @deftypefn {Built-in Function} int __builtin_parityl (unsigned long)
5098 Similar to @code{__builtin_parity}, except the argument type is
5099 @code{unsigned long}.
5102 @deftypefn {Built-in Function} int __builtin_ffsll (unsigned long long)
5103 Similar to @code{__builtin_ffs}, except the argument type is
5104 @code{unsigned long long}.
5107 @deftypefn {Built-in Function} int __builtin_clzll (unsigned long long)
5108 Similar to @code{__builtin_clz}, except the argument type is
5109 @code{unsigned long long}.
5112 @deftypefn {Built-in Function} int __builtin_ctzll (unsigned long long)
5113 Similar to @code{__builtin_ctz}, except the argument type is
5114 @code{unsigned long long}.
5117 @deftypefn {Built-in Function} int __builtin_popcountll (unsigned long long)
5118 Similar to @code{__builtin_popcount}, except the argument type is
5119 @code{unsigned long long}.
5122 @deftypefn {Built-in Function} int __builtin_parityll (unsigned long long)
5123 Similar to @code{__builtin_parity}, except the argument type is
5124 @code{unsigned long long}.
5128 @node Target Builtins
5129 @section Built-in Functions Specific to Particular Target Machines
5131 On some target machines, GCC supports many built-in functions specific
5132 to those machines. Generally these generate calls to specific machine
5133 instructions, but allow the compiler to schedule those calls.
5136 * Alpha Built-in Functions::
5137 * X86 Built-in Functions::
5138 * PowerPC AltiVec Built-in Functions::
5141 @node Alpha Built-in Functions
5142 @subsection Alpha Built-in Functions
5144 These built-in functions are available for the Alpha family of
5145 processors, depending on the command-line switches used.
5147 The following built-in functions are always available. They
5148 all generate the machine instruction that is part of the name.
5151 long __builtin_alpha_implver (void)
5152 long __builtin_alpha_rpcc (void)
5153 long __builtin_alpha_amask (long)
5154 long __builtin_alpha_cmpbge (long, long)
5155 long __builtin_alpha_extbl (long, long)
5156 long __builtin_alpha_extwl (long, long)
5157 long __builtin_alpha_extll (long, long)
5158 long __builtin_alpha_extql (long, long)
5159 long __builtin_alpha_extwh (long, long)
5160 long __builtin_alpha_extlh (long, long)
5161 long __builtin_alpha_extqh (long, long)
5162 long __builtin_alpha_insbl (long, long)
5163 long __builtin_alpha_inswl (long, long)
5164 long __builtin_alpha_insll (long, long)
5165 long __builtin_alpha_insql (long, long)
5166 long __builtin_alpha_inswh (long, long)
5167 long __builtin_alpha_inslh (long, long)
5168 long __builtin_alpha_insqh (long, long)
5169 long __builtin_alpha_mskbl (long, long)
5170 long __builtin_alpha_mskwl (long, long)
5171 long __builtin_alpha_mskll (long, long)
5172 long __builtin_alpha_mskql (long, long)
5173 long __builtin_alpha_mskwh (long, long)
5174 long __builtin_alpha_msklh (long, long)
5175 long __builtin_alpha_mskqh (long, long)
5176 long __builtin_alpha_umulh (long, long)
5177 long __builtin_alpha_zap (long, long)
5178 long __builtin_alpha_zapnot (long, long)
5181 The following built-in functions are always with @option{-mmax}
5182 or @option{-mcpu=@var{cpu}} where @var{cpu} is @code{pca56} or
5183 later. They all generate the machine instruction that is part
5187 long __builtin_alpha_pklb (long)
5188 long __builtin_alpha_pkwb (long)
5189 long __builtin_alpha_unpkbl (long)
5190 long __builtin_alpha_unpkbw (long)
5191 long __builtin_alpha_minub8 (long, long)
5192 long __builtin_alpha_minsb8 (long, long)
5193 long __builtin_alpha_minuw4 (long, long)
5194 long __builtin_alpha_minsw4 (long, long)
5195 long __builtin_alpha_maxub8 (long, long)
5196 long __builtin_alpha_maxsb8 (long, long)
5197 long __builtin_alpha_maxuw4 (long, long)
5198 long __builtin_alpha_maxsw4 (long, long)
5199 long __builtin_alpha_perr (long, long)
5202 The following built-in functions are always with @option{-mcix}
5203 or @option{-mcpu=@var{cpu}} where @var{cpu} is @code{ev67} or
5204 later. They all generate the machine instruction that is part
5208 long __builtin_alpha_cttz (long)
5209 long __builtin_alpha_ctlz (long)
5210 long __builtin_alpha_ctpop (long)
5213 The following builtins are available on systems that use the OSF/1
5214 PALcode. Normally they invoke the @code{rduniq} and @code{wruniq}
5215 PAL calls, but when invoked with @option{-mtls-kernel}, they invoke
5216 @code{rdval} and @code{wrval}.
5219 void *__builtin_thread_pointer (void)
5220 void __builtin_set_thread_pointer (void *)
5223 @node X86 Built-in Functions
5224 @subsection X86 Built-in Functions
5226 These built-in functions are available for the i386 and x86-64 family
5227 of computers, depending on the command-line switches used.
5229 The following machine modes are available for use with MMX built-in functions
5230 (@pxref{Vector Extensions}): @code{V2SI} for a vector of two 32-bit integers,
5231 @code{V4HI} for a vector of four 16-bit integers, and @code{V8QI} for a
5232 vector of eight 8-bit integers. Some of the built-in functions operate on
5233 MMX registers as a whole 64-bit entity, these use @code{DI} as their mode.
5235 If 3Dnow extensions are enabled, @code{V2SF} is used as a mode for a vector
5236 of two 32-bit floating point values.
5238 If SSE extensions are enabled, @code{V4SF} is used for a vector of four 32-bit
5239 floating point values. Some instructions use a vector of four 32-bit
5240 integers, these use @code{V4SI}. Finally, some instructions operate on an
5241 entire vector register, interpreting it as a 128-bit integer, these use mode
5244 The following built-in functions are made available by @option{-mmmx}.
5245 All of them generate the machine instruction that is part of the name.
5248 v8qi __builtin_ia32_paddb (v8qi, v8qi)
5249 v4hi __builtin_ia32_paddw (v4hi, v4hi)
5250 v2si __builtin_ia32_paddd (v2si, v2si)
5251 v8qi __builtin_ia32_psubb (v8qi, v8qi)
5252 v4hi __builtin_ia32_psubw (v4hi, v4hi)
5253 v2si __builtin_ia32_psubd (v2si, v2si)
5254 v8qi __builtin_ia32_paddsb (v8qi, v8qi)
5255 v4hi __builtin_ia32_paddsw (v4hi, v4hi)
5256 v8qi __builtin_ia32_psubsb (v8qi, v8qi)
5257 v4hi __builtin_ia32_psubsw (v4hi, v4hi)
5258 v8qi __builtin_ia32_paddusb (v8qi, v8qi)
5259 v4hi __builtin_ia32_paddusw (v4hi, v4hi)
5260 v8qi __builtin_ia32_psubusb (v8qi, v8qi)
5261 v4hi __builtin_ia32_psubusw (v4hi, v4hi)
5262 v4hi __builtin_ia32_pmullw (v4hi, v4hi)
5263 v4hi __builtin_ia32_pmulhw (v4hi, v4hi)
5264 di __builtin_ia32_pand (di, di)
5265 di __builtin_ia32_pandn (di,di)
5266 di __builtin_ia32_por (di, di)
5267 di __builtin_ia32_pxor (di, di)
5268 v8qi __builtin_ia32_pcmpeqb (v8qi, v8qi)
5269 v4hi __builtin_ia32_pcmpeqw (v4hi, v4hi)
5270 v2si __builtin_ia32_pcmpeqd (v2si, v2si)
5271 v8qi __builtin_ia32_pcmpgtb (v8qi, v8qi)
5272 v4hi __builtin_ia32_pcmpgtw (v4hi, v4hi)
5273 v2si __builtin_ia32_pcmpgtd (v2si, v2si)
5274 v8qi __builtin_ia32_punpckhbw (v8qi, v8qi)
5275 v4hi __builtin_ia32_punpckhwd (v4hi, v4hi)
5276 v2si __builtin_ia32_punpckhdq (v2si, v2si)
5277 v8qi __builtin_ia32_punpcklbw (v8qi, v8qi)
5278 v4hi __builtin_ia32_punpcklwd (v4hi, v4hi)
5279 v2si __builtin_ia32_punpckldq (v2si, v2si)
5280 v8qi __builtin_ia32_packsswb (v4hi, v4hi)
5281 v4hi __builtin_ia32_packssdw (v2si, v2si)
5282 v8qi __builtin_ia32_packuswb (v4hi, v4hi)
5285 The following built-in functions are made available either with
5286 @option{-msse}, or with a combination of @option{-m3dnow} and
5287 @option{-march=athlon}. All of them generate the machine
5288 instruction that is part of the name.
5291 v4hi __builtin_ia32_pmulhuw (v4hi, v4hi)
5292 v8qi __builtin_ia32_pavgb (v8qi, v8qi)
5293 v4hi __builtin_ia32_pavgw (v4hi, v4hi)
5294 v4hi __builtin_ia32_psadbw (v8qi, v8qi)
5295 v8qi __builtin_ia32_pmaxub (v8qi, v8qi)
5296 v4hi __builtin_ia32_pmaxsw (v4hi, v4hi)
5297 v8qi __builtin_ia32_pminub (v8qi, v8qi)
5298 v4hi __builtin_ia32_pminsw (v4hi, v4hi)
5299 int __builtin_ia32_pextrw (v4hi, int)
5300 v4hi __builtin_ia32_pinsrw (v4hi, int, int)
5301 int __builtin_ia32_pmovmskb (v8qi)
5302 void __builtin_ia32_maskmovq (v8qi, v8qi, char *)
5303 void __builtin_ia32_movntq (di *, di)
5304 void __builtin_ia32_sfence (void)
5307 The following built-in functions are available when @option{-msse} is used.
5308 All of them generate the machine instruction that is part of the name.
5311 int __builtin_ia32_comieq (v4sf, v4sf)
5312 int __builtin_ia32_comineq (v4sf, v4sf)
5313 int __builtin_ia32_comilt (v4sf, v4sf)
5314 int __builtin_ia32_comile (v4sf, v4sf)
5315 int __builtin_ia32_comigt (v4sf, v4sf)
5316 int __builtin_ia32_comige (v4sf, v4sf)
5317 int __builtin_ia32_ucomieq (v4sf, v4sf)
5318 int __builtin_ia32_ucomineq (v4sf, v4sf)
5319 int __builtin_ia32_ucomilt (v4sf, v4sf)
5320 int __builtin_ia32_ucomile (v4sf, v4sf)
5321 int __builtin_ia32_ucomigt (v4sf, v4sf)
5322 int __builtin_ia32_ucomige (v4sf, v4sf)
5323 v4sf __builtin_ia32_addps (v4sf, v4sf)
5324 v4sf __builtin_ia32_subps (v4sf, v4sf)
5325 v4sf __builtin_ia32_mulps (v4sf, v4sf)
5326 v4sf __builtin_ia32_divps (v4sf, v4sf)
5327 v4sf __builtin_ia32_addss (v4sf, v4sf)
5328 v4sf __builtin_ia32_subss (v4sf, v4sf)
5329 v4sf __builtin_ia32_mulss (v4sf, v4sf)
5330 v4sf __builtin_ia32_divss (v4sf, v4sf)
5331 v4si __builtin_ia32_cmpeqps (v4sf, v4sf)
5332 v4si __builtin_ia32_cmpltps (v4sf, v4sf)
5333 v4si __builtin_ia32_cmpleps (v4sf, v4sf)
5334 v4si __builtin_ia32_cmpgtps (v4sf, v4sf)
5335 v4si __builtin_ia32_cmpgeps (v4sf, v4sf)
5336 v4si __builtin_ia32_cmpunordps (v4sf, v4sf)
5337 v4si __builtin_ia32_cmpneqps (v4sf, v4sf)
5338 v4si __builtin_ia32_cmpnltps (v4sf, v4sf)
5339 v4si __builtin_ia32_cmpnleps (v4sf, v4sf)
5340 v4si __builtin_ia32_cmpngtps (v4sf, v4sf)
5341 v4si __builtin_ia32_cmpngeps (v4sf, v4sf)
5342 v4si __builtin_ia32_cmpordps (v4sf, v4sf)
5343 v4si __builtin_ia32_cmpeqss (v4sf, v4sf)
5344 v4si __builtin_ia32_cmpltss (v4sf, v4sf)
5345 v4si __builtin_ia32_cmpless (v4sf, v4sf)
5346 v4si __builtin_ia32_cmpunordss (v4sf, v4sf)
5347 v4si __builtin_ia32_cmpneqss (v4sf, v4sf)
5348 v4si __builtin_ia32_cmpnlts (v4sf, v4sf)
5349 v4si __builtin_ia32_cmpnless (v4sf, v4sf)
5350 v4si __builtin_ia32_cmpordss (v4sf, v4sf)
5351 v4sf __builtin_ia32_maxps (v4sf, v4sf)
5352 v4sf __builtin_ia32_maxss (v4sf, v4sf)
5353 v4sf __builtin_ia32_minps (v4sf, v4sf)
5354 v4sf __builtin_ia32_minss (v4sf, v4sf)
5355 v4sf __builtin_ia32_andps (v4sf, v4sf)
5356 v4sf __builtin_ia32_andnps (v4sf, v4sf)
5357 v4sf __builtin_ia32_orps (v4sf, v4sf)
5358 v4sf __builtin_ia32_xorps (v4sf, v4sf)
5359 v4sf __builtin_ia32_movss (v4sf, v4sf)
5360 v4sf __builtin_ia32_movhlps (v4sf, v4sf)
5361 v4sf __builtin_ia32_movlhps (v4sf, v4sf)
5362 v4sf __builtin_ia32_unpckhps (v4sf, v4sf)
5363 v4sf __builtin_ia32_unpcklps (v4sf, v4sf)
5364 v4sf __builtin_ia32_cvtpi2ps (v4sf, v2si)
5365 v4sf __builtin_ia32_cvtsi2ss (v4sf, int)
5366 v2si __builtin_ia32_cvtps2pi (v4sf)
5367 int __builtin_ia32_cvtss2si (v4sf)
5368 v2si __builtin_ia32_cvttps2pi (v4sf)
5369 int __builtin_ia32_cvttss2si (v4sf)
5370 v4sf __builtin_ia32_rcpps (v4sf)
5371 v4sf __builtin_ia32_rsqrtps (v4sf)
5372 v4sf __builtin_ia32_sqrtps (v4sf)
5373 v4sf __builtin_ia32_rcpss (v4sf)
5374 v4sf __builtin_ia32_rsqrtss (v4sf)
5375 v4sf __builtin_ia32_sqrtss (v4sf)
5376 v4sf __builtin_ia32_shufps (v4sf, v4sf, int)
5377 void __builtin_ia32_movntps (float *, v4sf)
5378 int __builtin_ia32_movmskps (v4sf)
5381 The following built-in functions are available when @option{-msse} is used.
5384 @item v4sf __builtin_ia32_loadaps (float *)
5385 Generates the @code{movaps} machine instruction as a load from memory.
5386 @item void __builtin_ia32_storeaps (float *, v4sf)
5387 Generates the @code{movaps} machine instruction as a store to memory.
5388 @item v4sf __builtin_ia32_loadups (float *)
5389 Generates the @code{movups} machine instruction as a load from memory.
5390 @item void __builtin_ia32_storeups (float *, v4sf)
5391 Generates the @code{movups} machine instruction as a store to memory.
5392 @item v4sf __builtin_ia32_loadsss (float *)
5393 Generates the @code{movss} machine instruction as a load from memory.
5394 @item void __builtin_ia32_storess (float *, v4sf)
5395 Generates the @code{movss} machine instruction as a store to memory.
5396 @item v4sf __builtin_ia32_loadhps (v4sf, v2si *)
5397 Generates the @code{movhps} machine instruction as a load from memory.
5398 @item v4sf __builtin_ia32_loadlps (v4sf, v2si *)
5399 Generates the @code{movlps} machine instruction as a load from memory
5400 @item void __builtin_ia32_storehps (v4sf, v2si *)
5401 Generates the @code{movhps} machine instruction as a store to memory.
5402 @item void __builtin_ia32_storelps (v4sf, v2si *)
5403 Generates the @code{movlps} machine instruction as a store to memory.
5406 The following built-in functions are available when @option{-m3dnow} is used.
5407 All of them generate the machine instruction that is part of the name.
5410 void __builtin_ia32_femms (void)
5411 v8qi __builtin_ia32_pavgusb (v8qi, v8qi)
5412 v2si __builtin_ia32_pf2id (v2sf)
5413 v2sf __builtin_ia32_pfacc (v2sf, v2sf)
5414 v2sf __builtin_ia32_pfadd (v2sf, v2sf)
5415 v2si __builtin_ia32_pfcmpeq (v2sf, v2sf)
5416 v2si __builtin_ia32_pfcmpge (v2sf, v2sf)
5417 v2si __builtin_ia32_pfcmpgt (v2sf, v2sf)
5418 v2sf __builtin_ia32_pfmax (v2sf, v2sf)
5419 v2sf __builtin_ia32_pfmin (v2sf, v2sf)
5420 v2sf __builtin_ia32_pfmul (v2sf, v2sf)
5421 v2sf __builtin_ia32_pfrcp (v2sf)
5422 v2sf __builtin_ia32_pfrcpit1 (v2sf, v2sf)
5423 v2sf __builtin_ia32_pfrcpit2 (v2sf, v2sf)
5424 v2sf __builtin_ia32_pfrsqrt (v2sf)
5425 v2sf __builtin_ia32_pfrsqrtit1 (v2sf, v2sf)
5426 v2sf __builtin_ia32_pfsub (v2sf, v2sf)
5427 v2sf __builtin_ia32_pfsubr (v2sf, v2sf)
5428 v2sf __builtin_ia32_pi2fd (v2si)
5429 v4hi __builtin_ia32_pmulhrw (v4hi, v4hi)
5432 The following built-in functions are available when both @option{-m3dnow}
5433 and @option{-march=athlon} are used. All of them generate the machine
5434 instruction that is part of the name.
5437 v2si __builtin_ia32_pf2iw (v2sf)
5438 v2sf __builtin_ia32_pfnacc (v2sf, v2sf)
5439 v2sf __builtin_ia32_pfpnacc (v2sf, v2sf)
5440 v2sf __builtin_ia32_pi2fw (v2si)
5441 v2sf __builtin_ia32_pswapdsf (v2sf)
5442 v2si __builtin_ia32_pswapdsi (v2si)
5445 @node PowerPC AltiVec Built-in Functions
5446 @subsection PowerPC AltiVec Built-in Functions
5448 These built-in functions are available for the PowerPC family
5449 of computers, depending on the command-line switches used.
5451 The following machine modes are available for use with AltiVec built-in
5452 functions (@pxref{Vector Extensions}): @code{V4SI} for a vector of four
5453 32-bit integers, @code{V4SF} for a vector of four 32-bit floating point
5454 numbers, @code{V8HI} for a vector of eight 16-bit integers, and
5455 @code{V16QI} for a vector of sixteen 8-bit integers.
5457 The following functions are made available by including
5458 @code{<altivec.h>} and using @option{-maltivec} and
5459 @option{-mabi=altivec}. The functions implement the functionality
5460 described in Motorola's AltiVec Programming Interface Manual.
5462 There are a few differences from Motorola's documentation and GCC's
5463 implementation. Vector constants are done with curly braces (not
5464 parentheses). Vector initializers require no casts if the vector
5465 constant is of the same type as the variable it is initializing. The
5466 @code{vector bool} type is deprecated and will be discontinued in
5467 further revisions. Use @code{vector signed} instead. If @code{signed}
5468 or @code{unsigned} is omitted, the vector type will default to
5469 @code{signed}. Lastly, all overloaded functions are implemented with macros
5470 for the C implementation. So code the following example will not work:
5473 vec_add ((vector signed int)@{1, 2, 3, 4@}, foo);
5476 Since vec_add is a macro, the vector constant in the above example will
5477 be treated as four different arguments. Wrap the entire argument in
5478 parentheses for this to work. The C++ implementation does not use
5481 @emph{Note:} Only the @code{<altivec.h>} interface is supported.
5482 Internally, GCC uses built-in functions to achieve the functionality in
5483 the aforementioned header file, but they are not supported and are
5484 subject to change without notice.
5487 vector signed char vec_abs (vector signed char, vector signed char);
5488 vector signed short vec_abs (vector signed short, vector signed short);
5489 vector signed int vec_abs (vector signed int, vector signed int);
5490 vector signed float vec_abs (vector signed float, vector signed float);
5492 vector signed char vec_abss (vector signed char, vector signed char);
5493 vector signed short vec_abss (vector signed short, vector signed short);
5495 vector signed char vec_add (vector signed char, vector signed char);
5496 vector unsigned char vec_add (vector signed char, vector unsigned char);
5498 vector unsigned char vec_add (vector unsigned char, vector signed char);
5500 vector unsigned char vec_add (vector unsigned char,
5501 vector unsigned char);
5502 vector signed short vec_add (vector signed short, vector signed short);
5503 vector unsigned short vec_add (vector signed short,
5504 vector unsigned short);
5505 vector unsigned short vec_add (vector unsigned short,
5506 vector signed short);
5507 vector unsigned short vec_add (vector unsigned short,
5508 vector unsigned short);
5509 vector signed int vec_add (vector signed int, vector signed int);
5510 vector unsigned int vec_add (vector signed int, vector unsigned int);
5511 vector unsigned int vec_add (vector unsigned int, vector signed int);
5512 vector unsigned int vec_add (vector unsigned int, vector unsigned int);
5513 vector float vec_add (vector float, vector float);
5515 vector unsigned int vec_addc (vector unsigned int, vector unsigned int);
5517 vector unsigned char vec_adds (vector signed char,
5518 vector unsigned char);
5519 vector unsigned char vec_adds (vector unsigned char,
5520 vector signed char);
5521 vector unsigned char vec_adds (vector unsigned char,
5522 vector unsigned char);
5523 vector signed char vec_adds (vector signed char, vector signed char);
5524 vector unsigned short vec_adds (vector signed short,
5525 vector unsigned short);
5526 vector unsigned short vec_adds (vector unsigned short,
5527 vector signed short);
5528 vector unsigned short vec_adds (vector unsigned short,
5529 vector unsigned short);
5530 vector signed short vec_adds (vector signed short, vector signed short);
5532 vector unsigned int vec_adds (vector signed int, vector unsigned int);
5533 vector unsigned int vec_adds (vector unsigned int, vector signed int);
5534 vector unsigned int vec_adds (vector unsigned int, vector unsigned int);
5536 vector signed int vec_adds (vector signed int, vector signed int);
5538 vector float vec_and (vector float, vector float);
5539 vector float vec_and (vector float, vector signed int);
5540 vector float vec_and (vector signed int, vector float);
5541 vector signed int vec_and (vector signed int, vector signed int);
5542 vector unsigned int vec_and (vector signed int, vector unsigned int);
5543 vector unsigned int vec_and (vector unsigned int, vector signed int);
5544 vector unsigned int vec_and (vector unsigned int, vector unsigned int);
5545 vector signed short vec_and (vector signed short, vector signed short);
5546 vector unsigned short vec_and (vector signed short,
5547 vector unsigned short);
5548 vector unsigned short vec_and (vector unsigned short,
5549 vector signed short);
5550 vector unsigned short vec_and (vector unsigned short,
5551 vector unsigned short);
5552 vector signed char vec_and (vector signed char, vector signed char);
5553 vector unsigned char vec_and (vector signed char, vector unsigned char);
5555 vector unsigned char vec_and (vector unsigned char, vector signed char);
5557 vector unsigned char vec_and (vector unsigned char,
5558 vector unsigned char);
5560 vector float vec_andc (vector float, vector float);
5561 vector float vec_andc (vector float, vector signed int);
5562 vector float vec_andc (vector signed int, vector float);
5563 vector signed int vec_andc (vector signed int, vector signed int);
5564 vector unsigned int vec_andc (vector signed int, vector unsigned int);
5565 vector unsigned int vec_andc (vector unsigned int, vector signed int);
5566 vector unsigned int vec_andc (vector unsigned int, vector unsigned int);
5568 vector signed short vec_andc (vector signed short, vector signed short);
5570 vector unsigned short vec_andc (vector signed short,
5571 vector unsigned short);
5572 vector unsigned short vec_andc (vector unsigned short,
5573 vector signed short);
5574 vector unsigned short vec_andc (vector unsigned short,
5575 vector unsigned short);
5576 vector signed char vec_andc (vector signed char, vector signed char);
5577 vector unsigned char vec_andc (vector signed char,
5578 vector unsigned char);
5579 vector unsigned char vec_andc (vector unsigned char,
5580 vector signed char);
5581 vector unsigned char vec_andc (vector unsigned char,
5582 vector unsigned char);
5584 vector unsigned char vec_avg (vector unsigned char,
5585 vector unsigned char);
5586 vector signed char vec_avg (vector signed char, vector signed char);
5587 vector unsigned short vec_avg (vector unsigned short,
5588 vector unsigned short);
5589 vector signed short vec_avg (vector signed short, vector signed short);
5590 vector unsigned int vec_avg (vector unsigned int, vector unsigned int);
5591 vector signed int vec_avg (vector signed int, vector signed int);
5593 vector float vec_ceil (vector float);
5595 vector signed int vec_cmpb (vector float, vector float);
5597 vector signed char vec_cmpeq (vector signed char, vector signed char);
5598 vector signed char vec_cmpeq (vector unsigned char,
5599 vector unsigned char);
5600 vector signed short vec_cmpeq (vector signed short,
5601 vector signed short);
5602 vector signed short vec_cmpeq (vector unsigned short,
5603 vector unsigned short);
5604 vector signed int vec_cmpeq (vector signed int, vector signed int);
5605 vector signed int vec_cmpeq (vector unsigned int, vector unsigned int);
5606 vector signed int vec_cmpeq (vector float, vector float);
5608 vector signed int vec_cmpge (vector float, vector float);
5610 vector signed char vec_cmpgt (vector unsigned char,
5611 vector unsigned char);
5612 vector signed char vec_cmpgt (vector signed char, vector signed char);
5613 vector signed short vec_cmpgt (vector unsigned short,
5614 vector unsigned short);
5615 vector signed short vec_cmpgt (vector signed short,
5616 vector signed short);
5617 vector signed int vec_cmpgt (vector unsigned int, vector unsigned int);
5618 vector signed int vec_cmpgt (vector signed int, vector signed int);
5619 vector signed int vec_cmpgt (vector float, vector float);
5621 vector signed int vec_cmple (vector float, vector float);
5623 vector signed char vec_cmplt (vector unsigned char,
5624 vector unsigned char);
5625 vector signed char vec_cmplt (vector signed char, vector signed char);
5626 vector signed short vec_cmplt (vector unsigned short,
5627 vector unsigned short);
5628 vector signed short vec_cmplt (vector signed short,
5629 vector signed short);
5630 vector signed int vec_cmplt (vector unsigned int, vector unsigned int);
5631 vector signed int vec_cmplt (vector signed int, vector signed int);
5632 vector signed int vec_cmplt (vector float, vector float);
5634 vector float vec_ctf (vector unsigned int, const char);
5635 vector float vec_ctf (vector signed int, const char);
5637 vector signed int vec_cts (vector float, const char);
5639 vector unsigned int vec_ctu (vector float, const char);
5641 void vec_dss (const char);
5643 void vec_dssall (void);
5645 void vec_dst (void *, int, const char);
5647 void vec_dstst (void *, int, const char);
5649 void vec_dststt (void *, int, const char);
5651 void vec_dstt (void *, int, const char);
5653 vector float vec_expte (vector float, vector float);
5655 vector float vec_floor (vector float, vector float);
5657 vector float vec_ld (int, vector float *);
5658 vector float vec_ld (int, float *):
5659 vector signed int vec_ld (int, int *);
5660 vector signed int vec_ld (int, vector signed int *);
5661 vector unsigned int vec_ld (int, vector unsigned int *);
5662 vector unsigned int vec_ld (int, unsigned int *);
5663 vector signed short vec_ld (int, short *, vector signed short *);
5664 vector unsigned short vec_ld (int, unsigned short *,
5665 vector unsigned short *);
5666 vector signed char vec_ld (int, signed char *);
5667 vector signed char vec_ld (int, vector signed char *);
5668 vector unsigned char vec_ld (int, unsigned char *);
5669 vector unsigned char vec_ld (int, vector unsigned char *);
5671 vector signed char vec_lde (int, signed char *);
5672 vector unsigned char vec_lde (int, unsigned char *);
5673 vector signed short vec_lde (int, short *);
5674 vector unsigned short vec_lde (int, unsigned short *);
5675 vector float vec_lde (int, float *);
5676 vector signed int vec_lde (int, int *);
5677 vector unsigned int vec_lde (int, unsigned int *);
5679 void float vec_ldl (int, float *);
5680 void float vec_ldl (int, vector float *);
5681 void signed int vec_ldl (int, vector signed int *);
5682 void signed int vec_ldl (int, int *);
5683 void unsigned int vec_ldl (int, unsigned int *);
5684 void unsigned int vec_ldl (int, vector unsigned int *);
5685 void signed short vec_ldl (int, vector signed short *);
5686 void signed short vec_ldl (int, short *);
5687 void unsigned short vec_ldl (int, vector unsigned short *);
5688 void unsigned short vec_ldl (int, unsigned short *);
5689 void signed char vec_ldl (int, vector signed char *);
5690 void signed char vec_ldl (int, signed char *);
5691 void unsigned char vec_ldl (int, vector unsigned char *);
5692 void unsigned char vec_ldl (int, unsigned char *);
5694 vector float vec_loge (vector float);
5696 vector unsigned char vec_lvsl (int, void *, int *);
5698 vector unsigned char vec_lvsr (int, void *, int *);
5700 vector float vec_madd (vector float, vector float, vector float);
5702 vector signed short vec_madds (vector signed short, vector signed short,
5703 vector signed short);
5705 vector unsigned char vec_max (vector signed char, vector unsigned char);
5707 vector unsigned char vec_max (vector unsigned char, vector signed char);
5709 vector unsigned char vec_max (vector unsigned char,
5710 vector unsigned char);
5711 vector signed char vec_max (vector signed char, vector signed char);
5712 vector unsigned short vec_max (vector signed short,
5713 vector unsigned short);
5714 vector unsigned short vec_max (vector unsigned short,
5715 vector signed short);
5716 vector unsigned short vec_max (vector unsigned short,
5717 vector unsigned short);
5718 vector signed short vec_max (vector signed short, vector signed short);
5719 vector unsigned int vec_max (vector signed int, vector unsigned int);
5720 vector unsigned int vec_max (vector unsigned int, vector signed int);
5721 vector unsigned int vec_max (vector unsigned int, vector unsigned int);
5722 vector signed int vec_max (vector signed int, vector signed int);
5723 vector float vec_max (vector float, vector float);
5725 vector signed char vec_mergeh (vector signed char, vector signed char);
5726 vector unsigned char vec_mergeh (vector unsigned char,
5727 vector unsigned char);
5728 vector signed short vec_mergeh (vector signed short,
5729 vector signed short);
5730 vector unsigned short vec_mergeh (vector unsigned short,
5731 vector unsigned short);
5732 vector float vec_mergeh (vector float, vector float);
5733 vector signed int vec_mergeh (vector signed int, vector signed int);
5734 vector unsigned int vec_mergeh (vector unsigned int,
5735 vector unsigned int);
5737 vector signed char vec_mergel (vector signed char, vector signed char);
5738 vector unsigned char vec_mergel (vector unsigned char,
5739 vector unsigned char);
5740 vector signed short vec_mergel (vector signed short,
5741 vector signed short);
5742 vector unsigned short vec_mergel (vector unsigned short,
5743 vector unsigned short);
5744 vector float vec_mergel (vector float, vector float);
5745 vector signed int vec_mergel (vector signed int, vector signed int);
5746 vector unsigned int vec_mergel (vector unsigned int,
5747 vector unsigned int);
5749 vector unsigned short vec_mfvscr (void);
5751 vector unsigned char vec_min (vector signed char, vector unsigned char);
5753 vector unsigned char vec_min (vector unsigned char, vector signed char);
5755 vector unsigned char vec_min (vector unsigned char,
5756 vector unsigned char);
5757 vector signed char vec_min (vector signed char, vector signed char);
5758 vector unsigned short vec_min (vector signed short,
5759 vector unsigned short);
5760 vector unsigned short vec_min (vector unsigned short,
5761 vector signed short);
5762 vector unsigned short vec_min (vector unsigned short,
5763 vector unsigned short);
5764 vector signed short vec_min (vector signed short, vector signed short);
5765 vector unsigned int vec_min (vector signed int, vector unsigned int);
5766 vector unsigned int vec_min (vector unsigned int, vector signed int);
5767 vector unsigned int vec_min (vector unsigned int, vector unsigned int);
5768 vector signed int vec_min (vector signed int, vector signed int);
5769 vector float vec_min (vector float, vector float);
5771 vector signed short vec_mladd (vector signed short, vector signed short,
5772 vector signed short);
5773 vector signed short vec_mladd (vector signed short,
5774 vector unsigned short,
5775 vector unsigned short);
5776 vector signed short vec_mladd (vector unsigned short,
5777 vector signed short,
5778 vector signed short);
5779 vector unsigned short vec_mladd (vector unsigned short,
5780 vector unsigned short,
5781 vector unsigned short);
5783 vector signed short vec_mradds (vector signed short,
5784 vector signed short,
5785 vector signed short);
5787 vector unsigned int vec_msum (vector unsigned char,
5788 vector unsigned char,
5789 vector unsigned int);
5790 vector signed int vec_msum (vector signed char, vector unsigned char,
5792 vector unsigned int vec_msum (vector unsigned short,
5793 vector unsigned short,
5794 vector unsigned int);
5795 vector signed int vec_msum (vector signed short, vector signed short,
5798 vector unsigned int vec_msums (vector unsigned short,
5799 vector unsigned short,
5800 vector unsigned int);
5801 vector signed int vec_msums (vector signed short, vector signed short,
5804 void vec_mtvscr (vector signed int);
5805 void vec_mtvscr (vector unsigned int);
5806 void vec_mtvscr (vector signed short);
5807 void vec_mtvscr (vector unsigned short);
5808 void vec_mtvscr (vector signed char);
5809 void vec_mtvscr (vector unsigned char);
5811 vector unsigned short vec_mule (vector unsigned char,
5812 vector unsigned char);
5813 vector signed short vec_mule (vector signed char, vector signed char);
5814 vector unsigned int vec_mule (vector unsigned short,
5815 vector unsigned short);
5816 vector signed int vec_mule (vector signed short, vector signed short);
5818 vector unsigned short vec_mulo (vector unsigned char,
5819 vector unsigned char);
5820 vector signed short vec_mulo (vector signed char, vector signed char);
5821 vector unsigned int vec_mulo (vector unsigned short,
5822 vector unsigned short);
5823 vector signed int vec_mulo (vector signed short, vector signed short);
5825 vector float vec_nmsub (vector float, vector float, vector float);
5827 vector float vec_nor (vector float, vector float);
5828 vector signed int vec_nor (vector signed int, vector signed int);
5829 vector unsigned int vec_nor (vector unsigned int, vector unsigned int);
5830 vector signed short vec_nor (vector signed short, vector signed short);
5831 vector unsigned short vec_nor (vector unsigned short,
5832 vector unsigned short);
5833 vector signed char vec_nor (vector signed char, vector signed char);
5834 vector unsigned char vec_nor (vector unsigned char,
5835 vector unsigned char);
5837 vector float vec_or (vector float, vector float);
5838 vector float vec_or (vector float, vector signed int);
5839 vector float vec_or (vector signed int, vector float);
5840 vector signed int vec_or (vector signed int, vector signed int);
5841 vector unsigned int vec_or (vector signed int, vector unsigned int);
5842 vector unsigned int vec_or (vector unsigned int, vector signed int);
5843 vector unsigned int vec_or (vector unsigned int, vector unsigned int);
5844 vector signed short vec_or (vector signed short, vector signed short);
5845 vector unsigned short vec_or (vector signed short,
5846 vector unsigned short);
5847 vector unsigned short vec_or (vector unsigned short,
5848 vector signed short);
5849 vector unsigned short vec_or (vector unsigned short,
5850 vector unsigned short);
5851 vector signed char vec_or (vector signed char, vector signed char);
5852 vector unsigned char vec_or (vector signed char, vector unsigned char);
5853 vector unsigned char vec_or (vector unsigned char, vector signed char);
5854 vector unsigned char vec_or (vector unsigned char,
5855 vector unsigned char);
5857 vector signed char vec_pack (vector signed short, vector signed short);
5858 vector unsigned char vec_pack (vector unsigned short,
5859 vector unsigned short);
5860 vector signed short vec_pack (vector signed int, vector signed int);
5861 vector unsigned short vec_pack (vector unsigned int,
5862 vector unsigned int);
5864 vector signed short vec_packpx (vector unsigned int,
5865 vector unsigned int);
5867 vector unsigned char vec_packs (vector unsigned short,
5868 vector unsigned short);
5869 vector signed char vec_packs (vector signed short, vector signed short);
5871 vector unsigned short vec_packs (vector unsigned int,
5872 vector unsigned int);
5873 vector signed short vec_packs (vector signed int, vector signed int);
5875 vector unsigned char vec_packsu (vector unsigned short,
5876 vector unsigned short);
5877 vector unsigned char vec_packsu (vector signed short,
5878 vector signed short);
5879 vector unsigned short vec_packsu (vector unsigned int,
5880 vector unsigned int);
5881 vector unsigned short vec_packsu (vector signed int, vector signed int);
5883 vector float vec_perm (vector float, vector float,
5884 vector unsigned char);
5885 vector signed int vec_perm (vector signed int, vector signed int,
5886 vector unsigned char);
5887 vector unsigned int vec_perm (vector unsigned int, vector unsigned int,
5888 vector unsigned char);
5889 vector signed short vec_perm (vector signed short, vector signed short,
5890 vector unsigned char);
5891 vector unsigned short vec_perm (vector unsigned short,
5892 vector unsigned short,
5893 vector unsigned char);
5894 vector signed char vec_perm (vector signed char, vector signed char,
5895 vector unsigned char);
5896 vector unsigned char vec_perm (vector unsigned char,
5897 vector unsigned char,
5898 vector unsigned char);
5900 vector float vec_re (vector float);
5902 vector signed char vec_rl (vector signed char, vector unsigned char);
5903 vector unsigned char vec_rl (vector unsigned char,
5904 vector unsigned char);
5905 vector signed short vec_rl (vector signed short, vector unsigned short);
5907 vector unsigned short vec_rl (vector unsigned short,
5908 vector unsigned short);
5909 vector signed int vec_rl (vector signed int, vector unsigned int);
5910 vector unsigned int vec_rl (vector unsigned int, vector unsigned int);
5912 vector float vec_round (vector float);
5914 vector float vec_rsqrte (vector float);
5916 vector float vec_sel (vector float, vector float, vector signed int);
5917 vector float vec_sel (vector float, vector float, vector unsigned int);
5918 vector signed int vec_sel (vector signed int, vector signed int,
5920 vector signed int vec_sel (vector signed int, vector signed int,
5921 vector unsigned int);
5922 vector unsigned int vec_sel (vector unsigned int, vector unsigned int,
5924 vector unsigned int vec_sel (vector unsigned int, vector unsigned int,
5925 vector unsigned int);
5926 vector signed short vec_sel (vector signed short, vector signed short,
5927 vector signed short);
5928 vector signed short vec_sel (vector signed short, vector signed short,
5929 vector unsigned short);
5930 vector unsigned short vec_sel (vector unsigned short,
5931 vector unsigned short,
5932 vector signed short);
5933 vector unsigned short vec_sel (vector unsigned short,
5934 vector unsigned short,
5935 vector unsigned short);
5936 vector signed char vec_sel (vector signed char, vector signed char,
5937 vector signed char);
5938 vector signed char vec_sel (vector signed char, vector signed char,
5939 vector unsigned char);
5940 vector unsigned char vec_sel (vector unsigned char,
5941 vector unsigned char,
5942 vector signed char);
5943 vector unsigned char vec_sel (vector unsigned char,
5944 vector unsigned char,
5945 vector unsigned char);
5947 vector signed char vec_sl (vector signed char, vector unsigned char);
5948 vector unsigned char vec_sl (vector unsigned char,
5949 vector unsigned char);
5950 vector signed short vec_sl (vector signed short, vector unsigned short);
5952 vector unsigned short vec_sl (vector unsigned short,
5953 vector unsigned short);
5954 vector signed int vec_sl (vector signed int, vector unsigned int);
5955 vector unsigned int vec_sl (vector unsigned int, vector unsigned int);
5957 vector float vec_sld (vector float, vector float, const char);
5958 vector signed int vec_sld (vector signed int, vector signed int,
5960 vector unsigned int vec_sld (vector unsigned int, vector unsigned int,
5962 vector signed short vec_sld (vector signed short, vector signed short,
5964 vector unsigned short vec_sld (vector unsigned short,
5965 vector unsigned short, const char);
5966 vector signed char vec_sld (vector signed char, vector signed char,
5968 vector unsigned char vec_sld (vector unsigned char,
5969 vector unsigned char,
5972 vector signed int vec_sll (vector signed int, vector unsigned int);
5973 vector signed int vec_sll (vector signed int, vector unsigned short);
5974 vector signed int vec_sll (vector signed int, vector unsigned char);
5975 vector unsigned int vec_sll (vector unsigned int, vector unsigned int);
5976 vector unsigned int vec_sll (vector unsigned int,
5977 vector unsigned short);
5978 vector unsigned int vec_sll (vector unsigned int, vector unsigned char);
5980 vector signed short vec_sll (vector signed short, vector unsigned int);
5981 vector signed short vec_sll (vector signed short,
5982 vector unsigned short);
5983 vector signed short vec_sll (vector signed short, vector unsigned char);
5985 vector unsigned short vec_sll (vector unsigned short,
5986 vector unsigned int);
5987 vector unsigned short vec_sll (vector unsigned short,
5988 vector unsigned short);
5989 vector unsigned short vec_sll (vector unsigned short,
5990 vector unsigned char);
5991 vector signed char vec_sll (vector signed char, vector unsigned int);
5992 vector signed char vec_sll (vector signed char, vector unsigned short);
5993 vector signed char vec_sll (vector signed char, vector unsigned char);
5994 vector unsigned char vec_sll (vector unsigned char,
5995 vector unsigned int);
5996 vector unsigned char vec_sll (vector unsigned char,
5997 vector unsigned short);
5998 vector unsigned char vec_sll (vector unsigned char,
5999 vector unsigned char);
6001 vector float vec_slo (vector float, vector signed char);
6002 vector float vec_slo (vector float, vector unsigned char);
6003 vector signed int vec_slo (vector signed int, vector signed char);
6004 vector signed int vec_slo (vector signed int, vector unsigned char);
6005 vector unsigned int vec_slo (vector unsigned int, vector signed char);
6006 vector unsigned int vec_slo (vector unsigned int, vector unsigned char);
6008 vector signed short vec_slo (vector signed short, vector signed char);
6009 vector signed short vec_slo (vector signed short, vector unsigned char);
6011 vector unsigned short vec_slo (vector unsigned short,
6012 vector signed char);
6013 vector unsigned short vec_slo (vector unsigned short,
6014 vector unsigned char);
6015 vector signed char vec_slo (vector signed char, vector signed char);
6016 vector signed char vec_slo (vector signed char, vector unsigned char);
6017 vector unsigned char vec_slo (vector unsigned char, vector signed char);
6019 vector unsigned char vec_slo (vector unsigned char,
6020 vector unsigned char);
6022 vector signed char vec_splat (vector signed char, const char);
6023 vector unsigned char vec_splat (vector unsigned char, const char);
6024 vector signed short vec_splat (vector signed short, const char);
6025 vector unsigned short vec_splat (vector unsigned short, const char);
6026 vector float vec_splat (vector float, const char);
6027 vector signed int vec_splat (vector signed int, const char);
6028 vector unsigned int vec_splat (vector unsigned int, const char);
6030 vector signed char vec_splat_s8 (const char);
6032 vector signed short vec_splat_s16 (const char);
6034 vector signed int vec_splat_s32 (const char);
6036 vector unsigned char vec_splat_u8 (const char);
6038 vector unsigned short vec_splat_u16 (const char);
6040 vector unsigned int vec_splat_u32 (const char);
6042 vector signed char vec_sr (vector signed char, vector unsigned char);
6043 vector unsigned char vec_sr (vector unsigned char,
6044 vector unsigned char);
6045 vector signed short vec_sr (vector signed short, vector unsigned short);
6047 vector unsigned short vec_sr (vector unsigned short,
6048 vector unsigned short);
6049 vector signed int vec_sr (vector signed int, vector unsigned int);
6050 vector unsigned int vec_sr (vector unsigned int, vector unsigned int);
6052 vector signed char vec_sra (vector signed char, vector unsigned char);
6053 vector unsigned char vec_sra (vector unsigned char,
6054 vector unsigned char);
6055 vector signed short vec_sra (vector signed short,
6056 vector unsigned short);
6057 vector unsigned short vec_sra (vector unsigned short,
6058 vector unsigned short);
6059 vector signed int vec_sra (vector signed int, vector unsigned int);
6060 vector unsigned int vec_sra (vector unsigned int, vector unsigned int);
6062 vector signed int vec_srl (vector signed int, vector unsigned int);
6063 vector signed int vec_srl (vector signed int, vector unsigned short);
6064 vector signed int vec_srl (vector signed int, vector unsigned char);
6065 vector unsigned int vec_srl (vector unsigned int, vector unsigned int);
6066 vector unsigned int vec_srl (vector unsigned int,
6067 vector unsigned short);
6068 vector unsigned int vec_srl (vector unsigned int, vector unsigned char);
6070 vector signed short vec_srl (vector signed short, vector unsigned int);
6071 vector signed short vec_srl (vector signed short,
6072 vector unsigned short);
6073 vector signed short vec_srl (vector signed short, vector unsigned char);
6075 vector unsigned short vec_srl (vector unsigned short,
6076 vector unsigned int);
6077 vector unsigned short vec_srl (vector unsigned short,
6078 vector unsigned short);
6079 vector unsigned short vec_srl (vector unsigned short,
6080 vector unsigned char);
6081 vector signed char vec_srl (vector signed char, vector unsigned int);
6082 vector signed char vec_srl (vector signed char, vector unsigned short);
6083 vector signed char vec_srl (vector signed char, vector unsigned char);
6084 vector unsigned char vec_srl (vector unsigned char,
6085 vector unsigned int);
6086 vector unsigned char vec_srl (vector unsigned char,
6087 vector unsigned short);
6088 vector unsigned char vec_srl (vector unsigned char,
6089 vector unsigned char);
6091 vector float vec_sro (vector float, vector signed char);
6092 vector float vec_sro (vector float, vector unsigned char);
6093 vector signed int vec_sro (vector signed int, vector signed char);
6094 vector signed int vec_sro (vector signed int, vector unsigned char);
6095 vector unsigned int vec_sro (vector unsigned int, vector signed char);
6096 vector unsigned int vec_sro (vector unsigned int, vector unsigned char);
6098 vector signed short vec_sro (vector signed short, vector signed char);
6099 vector signed short vec_sro (vector signed short, vector unsigned char);
6101 vector unsigned short vec_sro (vector unsigned short,
6102 vector signed char);
6103 vector unsigned short vec_sro (vector unsigned short,
6104 vector unsigned char);
6105 vector signed char vec_sro (vector signed char, vector signed char);
6106 vector signed char vec_sro (vector signed char, vector unsigned char);
6107 vector unsigned char vec_sro (vector unsigned char, vector signed char);
6109 vector unsigned char vec_sro (vector unsigned char,
6110 vector unsigned char);
6112 void vec_st (vector float, int, float *);
6113 void vec_st (vector float, int, vector float *);
6114 void vec_st (vector signed int, int, int *);
6115 void vec_st (vector signed int, int, unsigned int *);
6116 void vec_st (vector unsigned int, int, unsigned int *);
6117 void vec_st (vector unsigned int, int, vector unsigned int *);
6118 void vec_st (vector signed short, int, short *);
6119 void vec_st (vector signed short, int, vector unsigned short *);
6120 void vec_st (vector signed short, int, vector signed short *);
6121 void vec_st (vector unsigned short, int, unsigned short *);
6122 void vec_st (vector unsigned short, int, vector unsigned short *);
6123 void vec_st (vector signed char, int, signed char *);
6124 void vec_st (vector signed char, int, unsigned char *);
6125 void vec_st (vector signed char, int, vector signed char *);
6126 void vec_st (vector unsigned char, int, unsigned char *);
6127 void vec_st (vector unsigned char, int, vector unsigned char *);
6129 void vec_ste (vector signed char, int, unsigned char *);
6130 void vec_ste (vector signed char, int, signed char *);
6131 void vec_ste (vector unsigned char, int, unsigned char *);
6132 void vec_ste (vector signed short, int, short *);
6133 void vec_ste (vector signed short, int, unsigned short *);
6134 void vec_ste (vector unsigned short, int, void *);
6135 void vec_ste (vector signed int, int, unsigned int *);
6136 void vec_ste (vector signed int, int, int *);
6137 void vec_ste (vector unsigned int, int, unsigned int *);
6138 void vec_ste (vector float, int, float *);
6140 void vec_stl (vector float, int, vector float *);
6141 void vec_stl (vector float, int, float *);
6142 void vec_stl (vector signed int, int, vector signed int *);
6143 void vec_stl (vector signed int, int, int *);
6144 void vec_stl (vector signed int, int, unsigned int *);
6145 void vec_stl (vector unsigned int, int, vector unsigned int *);
6146 void vec_stl (vector unsigned int, int, unsigned int *);
6147 void vec_stl (vector signed short, int, short *);
6148 void vec_stl (vector signed short, int, unsigned short *);
6149 void vec_stl (vector signed short, int, vector signed short *);
6150 void vec_stl (vector unsigned short, int, unsigned short *);
6151 void vec_stl (vector unsigned short, int, vector signed short *);
6152 void vec_stl (vector signed char, int, signed char *);
6153 void vec_stl (vector signed char, int, unsigned char *);
6154 void vec_stl (vector signed char, int, vector signed char *);
6155 void vec_stl (vector unsigned char, int, unsigned char *);
6156 void vec_stl (vector unsigned char, int, vector unsigned char *);
6158 vector signed char vec_sub (vector signed char, vector signed char);
6159 vector unsigned char vec_sub (vector signed char, vector unsigned char);
6161 vector unsigned char vec_sub (vector unsigned char, vector signed char);
6163 vector unsigned char vec_sub (vector unsigned char,
6164 vector unsigned char);
6165 vector signed short vec_sub (vector signed short, vector signed short);
6166 vector unsigned short vec_sub (vector signed short,
6167 vector unsigned short);
6168 vector unsigned short vec_sub (vector unsigned short,
6169 vector signed short);
6170 vector unsigned short vec_sub (vector unsigned short,
6171 vector unsigned short);
6172 vector signed int vec_sub (vector signed int, vector signed int);
6173 vector unsigned int vec_sub (vector signed int, vector unsigned int);
6174 vector unsigned int vec_sub (vector unsigned int, vector signed int);
6175 vector unsigned int vec_sub (vector unsigned int, vector unsigned int);
6176 vector float vec_sub (vector float, vector float);
6178 vector unsigned int vec_subc (vector unsigned int, vector unsigned int);
6180 vector unsigned char vec_subs (vector signed char,
6181 vector unsigned char);
6182 vector unsigned char vec_subs (vector unsigned char,
6183 vector signed char);
6184 vector unsigned char vec_subs (vector unsigned char,
6185 vector unsigned char);
6186 vector signed char vec_subs (vector signed char, vector signed char);
6187 vector unsigned short vec_subs (vector signed short,
6188 vector unsigned short);
6189 vector unsigned short vec_subs (vector unsigned short,
6190 vector signed short);
6191 vector unsigned short vec_subs (vector unsigned short,
6192 vector unsigned short);
6193 vector signed short vec_subs (vector signed short, vector signed short);
6195 vector unsigned int vec_subs (vector signed int, vector unsigned int);
6196 vector unsigned int vec_subs (vector unsigned int, vector signed int);
6197 vector unsigned int vec_subs (vector unsigned int, vector unsigned int);
6199 vector signed int vec_subs (vector signed int, vector signed int);
6201 vector unsigned int vec_sum4s (vector unsigned char,
6202 vector unsigned int);
6203 vector signed int vec_sum4s (vector signed char, vector signed int);
6204 vector signed int vec_sum4s (vector signed short, vector signed int);
6206 vector signed int vec_sum2s (vector signed int, vector signed int);
6208 vector signed int vec_sums (vector signed int, vector signed int);
6210 vector float vec_trunc (vector float);
6212 vector signed short vec_unpackh (vector signed char);
6213 vector unsigned int vec_unpackh (vector signed short);
6214 vector signed int vec_unpackh (vector signed short);
6216 vector signed short vec_unpackl (vector signed char);
6217 vector unsigned int vec_unpackl (vector signed short);
6218 vector signed int vec_unpackl (vector signed short);
6220 vector float vec_xor (vector float, vector float);
6221 vector float vec_xor (vector float, vector signed int);
6222 vector float vec_xor (vector signed int, vector float);
6223 vector signed int vec_xor (vector signed int, vector signed int);
6224 vector unsigned int vec_xor (vector signed int, vector unsigned int);
6225 vector unsigned int vec_xor (vector unsigned int, vector signed int);
6226 vector unsigned int vec_xor (vector unsigned int, vector unsigned int);
6227 vector signed short vec_xor (vector signed short, vector signed short);
6228 vector unsigned short vec_xor (vector signed short,
6229 vector unsigned short);
6230 vector unsigned short vec_xor (vector unsigned short,
6231 vector signed short);
6232 vector unsigned short vec_xor (vector unsigned short,
6233 vector unsigned short);
6234 vector signed char vec_xor (vector signed char, vector signed char);
6235 vector unsigned char vec_xor (vector signed char, vector unsigned char);
6237 vector unsigned char vec_xor (vector unsigned char, vector signed char);
6239 vector unsigned char vec_xor (vector unsigned char,
6240 vector unsigned char);
6242 vector signed int vec_all_eq (vector signed char, vector unsigned char);
6244 vector signed int vec_all_eq (vector signed char, vector signed char);
6245 vector signed int vec_all_eq (vector unsigned char, vector signed char);
6247 vector signed int vec_all_eq (vector unsigned char,
6248 vector unsigned char);
6249 vector signed int vec_all_eq (vector signed short,
6250 vector unsigned short);
6251 vector signed int vec_all_eq (vector signed short, vector signed short);
6253 vector signed int vec_all_eq (vector unsigned short,
6254 vector signed short);
6255 vector signed int vec_all_eq (vector unsigned short,
6256 vector unsigned short);
6257 vector signed int vec_all_eq (vector signed int, vector unsigned int);
6258 vector signed int vec_all_eq (vector signed int, vector signed int);
6259 vector signed int vec_all_eq (vector unsigned int, vector signed int);
6260 vector signed int vec_all_eq (vector unsigned int, vector unsigned int);
6262 vector signed int vec_all_eq (vector float, vector float);
6264 vector signed int vec_all_ge (vector signed char, vector unsigned char);
6266 vector signed int vec_all_ge (vector unsigned char, vector signed char);
6268 vector signed int vec_all_ge (vector unsigned char,
6269 vector unsigned char);
6270 vector signed int vec_all_ge (vector signed char, vector signed char);
6271 vector signed int vec_all_ge (vector signed short,
6272 vector unsigned short);
6273 vector signed int vec_all_ge (vector unsigned short,
6274 vector signed short);
6275 vector signed int vec_all_ge (vector unsigned short,
6276 vector unsigned short);
6277 vector signed int vec_all_ge (vector signed short, vector signed short);
6279 vector signed int vec_all_ge (vector signed int, vector unsigned int);
6280 vector signed int vec_all_ge (vector unsigned int, vector signed int);
6281 vector signed int vec_all_ge (vector unsigned int, vector unsigned int);
6283 vector signed int vec_all_ge (vector signed int, vector signed int);
6284 vector signed int vec_all_ge (vector float, vector float);
6286 vector signed int vec_all_gt (vector signed char, vector unsigned char);
6288 vector signed int vec_all_gt (vector unsigned char, vector signed char);
6290 vector signed int vec_all_gt (vector unsigned char,
6291 vector unsigned char);
6292 vector signed int vec_all_gt (vector signed char, vector signed char);
6293 vector signed int vec_all_gt (vector signed short,
6294 vector unsigned short);
6295 vector signed int vec_all_gt (vector unsigned short,
6296 vector signed short);
6297 vector signed int vec_all_gt (vector unsigned short,
6298 vector unsigned short);
6299 vector signed int vec_all_gt (vector signed short, vector signed short);
6301 vector signed int vec_all_gt (vector signed int, vector unsigned int);
6302 vector signed int vec_all_gt (vector unsigned int, vector signed int);
6303 vector signed int vec_all_gt (vector unsigned int, vector unsigned int);
6305 vector signed int vec_all_gt (vector signed int, vector signed int);
6306 vector signed int vec_all_gt (vector float, vector float);
6308 vector signed int vec_all_in (vector float, vector float);
6310 vector signed int vec_all_le (vector signed char, vector unsigned char);
6312 vector signed int vec_all_le (vector unsigned char, vector signed char);
6314 vector signed int vec_all_le (vector unsigned char,
6315 vector unsigned char);
6316 vector signed int vec_all_le (vector signed char, vector signed char);
6317 vector signed int vec_all_le (vector signed short,
6318 vector unsigned short);
6319 vector signed int vec_all_le (vector unsigned short,
6320 vector signed short);
6321 vector signed int vec_all_le (vector unsigned short,
6322 vector unsigned short);
6323 vector signed int vec_all_le (vector signed short, vector signed short);
6325 vector signed int vec_all_le (vector signed int, vector unsigned int);
6326 vector signed int vec_all_le (vector unsigned int, vector signed int);
6327 vector signed int vec_all_le (vector unsigned int, vector unsigned int);
6329 vector signed int vec_all_le (vector signed int, vector signed int);
6330 vector signed int vec_all_le (vector float, vector float);
6332 vector signed int vec_all_lt (vector signed char, vector unsigned char);
6334 vector signed int vec_all_lt (vector unsigned char, vector signed char);
6336 vector signed int vec_all_lt (vector unsigned char,
6337 vector unsigned char);
6338 vector signed int vec_all_lt (vector signed char, vector signed char);
6339 vector signed int vec_all_lt (vector signed short,
6340 vector unsigned short);
6341 vector signed int vec_all_lt (vector unsigned short,
6342 vector signed short);
6343 vector signed int vec_all_lt (vector unsigned short,
6344 vector unsigned short);
6345 vector signed int vec_all_lt (vector signed short, vector signed short);
6347 vector signed int vec_all_lt (vector signed int, vector unsigned int);
6348 vector signed int vec_all_lt (vector unsigned int, vector signed int);
6349 vector signed int vec_all_lt (vector unsigned int, vector unsigned int);
6351 vector signed int vec_all_lt (vector signed int, vector signed int);
6352 vector signed int vec_all_lt (vector float, vector float);
6354 vector signed int vec_all_nan (vector float);
6356 vector signed int vec_all_ne (vector signed char, vector unsigned char);
6358 vector signed int vec_all_ne (vector signed char, vector signed char);
6359 vector signed int vec_all_ne (vector unsigned char, vector signed char);
6361 vector signed int vec_all_ne (vector unsigned char,
6362 vector unsigned char);
6363 vector signed int vec_all_ne (vector signed short,
6364 vector unsigned short);
6365 vector signed int vec_all_ne (vector signed short, vector signed short);
6367 vector signed int vec_all_ne (vector unsigned short,
6368 vector signed short);
6369 vector signed int vec_all_ne (vector unsigned short,
6370 vector unsigned short);
6371 vector signed int vec_all_ne (vector signed int, vector unsigned int);
6372 vector signed int vec_all_ne (vector signed int, vector signed int);
6373 vector signed int vec_all_ne (vector unsigned int, vector signed int);
6374 vector signed int vec_all_ne (vector unsigned int, vector unsigned int);
6376 vector signed int vec_all_ne (vector float, vector float);
6378 vector signed int vec_all_nge (vector float, vector float);
6380 vector signed int vec_all_ngt (vector float, vector float);
6382 vector signed int vec_all_nle (vector float, vector float);
6384 vector signed int vec_all_nlt (vector float, vector float);
6386 vector signed int vec_all_numeric (vector float);
6388 vector signed int vec_any_eq (vector signed char, vector unsigned char);
6390 vector signed int vec_any_eq (vector signed char, vector signed char);
6391 vector signed int vec_any_eq (vector unsigned char, vector signed char);
6393 vector signed int vec_any_eq (vector unsigned char,
6394 vector unsigned char);
6395 vector signed int vec_any_eq (vector signed short,
6396 vector unsigned short);
6397 vector signed int vec_any_eq (vector signed short, vector signed short);
6399 vector signed int vec_any_eq (vector unsigned short,
6400 vector signed short);
6401 vector signed int vec_any_eq (vector unsigned short,
6402 vector unsigned short);
6403 vector signed int vec_any_eq (vector signed int, vector unsigned int);
6404 vector signed int vec_any_eq (vector signed int, vector signed int);
6405 vector signed int vec_any_eq (vector unsigned int, vector signed int);
6406 vector signed int vec_any_eq (vector unsigned int, vector unsigned int);
6408 vector signed int vec_any_eq (vector float, vector float);
6410 vector signed int vec_any_ge (vector signed char, vector unsigned char);
6412 vector signed int vec_any_ge (vector unsigned char, vector signed char);
6414 vector signed int vec_any_ge (vector unsigned char,
6415 vector unsigned char);
6416 vector signed int vec_any_ge (vector signed char, vector signed char);
6417 vector signed int vec_any_ge (vector signed short,
6418 vector unsigned short);
6419 vector signed int vec_any_ge (vector unsigned short,
6420 vector signed short);
6421 vector signed int vec_any_ge (vector unsigned short,
6422 vector unsigned short);
6423 vector signed int vec_any_ge (vector signed short, vector signed short);
6425 vector signed int vec_any_ge (vector signed int, vector unsigned int);
6426 vector signed int vec_any_ge (vector unsigned int, vector signed int);
6427 vector signed int vec_any_ge (vector unsigned int, vector unsigned int);
6429 vector signed int vec_any_ge (vector signed int, vector signed int);
6430 vector signed int vec_any_ge (vector float, vector float);
6432 vector signed int vec_any_gt (vector signed char, vector unsigned char);
6434 vector signed int vec_any_gt (vector unsigned char, vector signed char);
6436 vector signed int vec_any_gt (vector unsigned char,
6437 vector unsigned char);
6438 vector signed int vec_any_gt (vector signed char, vector signed char);
6439 vector signed int vec_any_gt (vector signed short,
6440 vector unsigned short);
6441 vector signed int vec_any_gt (vector unsigned short,
6442 vector signed short);
6443 vector signed int vec_any_gt (vector unsigned short,
6444 vector unsigned short);
6445 vector signed int vec_any_gt (vector signed short, vector signed short);
6447 vector signed int vec_any_gt (vector signed int, vector unsigned int);
6448 vector signed int vec_any_gt (vector unsigned int, vector signed int);
6449 vector signed int vec_any_gt (vector unsigned int, vector unsigned int);
6451 vector signed int vec_any_gt (vector signed int, vector signed int);
6452 vector signed int vec_any_gt (vector float, vector float);
6454 vector signed int vec_any_le (vector signed char, vector unsigned char);
6456 vector signed int vec_any_le (vector unsigned char, vector signed char);
6458 vector signed int vec_any_le (vector unsigned char,
6459 vector unsigned char);
6460 vector signed int vec_any_le (vector signed char, vector signed char);
6461 vector signed int vec_any_le (vector signed short,
6462 vector unsigned short);
6463 vector signed int vec_any_le (vector unsigned short,
6464 vector signed short);
6465 vector signed int vec_any_le (vector unsigned short,
6466 vector unsigned short);
6467 vector signed int vec_any_le (vector signed short, vector signed short);
6469 vector signed int vec_any_le (vector signed int, vector unsigned int);
6470 vector signed int vec_any_le (vector unsigned int, vector signed int);
6471 vector signed int vec_any_le (vector unsigned int, vector unsigned int);
6473 vector signed int vec_any_le (vector signed int, vector signed int);
6474 vector signed int vec_any_le (vector float, vector float);
6476 vector signed int vec_any_lt (vector signed char, vector unsigned char);
6478 vector signed int vec_any_lt (vector unsigned char, vector signed char);
6480 vector signed int vec_any_lt (vector unsigned char,
6481 vector unsigned char);
6482 vector signed int vec_any_lt (vector signed char, vector signed char);
6483 vector signed int vec_any_lt (vector signed short,
6484 vector unsigned short);
6485 vector signed int vec_any_lt (vector unsigned short,
6486 vector signed short);
6487 vector signed int vec_any_lt (vector unsigned short,
6488 vector unsigned short);
6489 vector signed int vec_any_lt (vector signed short, vector signed short);
6491 vector signed int vec_any_lt (vector signed int, vector unsigned int);
6492 vector signed int vec_any_lt (vector unsigned int, vector signed int);
6493 vector signed int vec_any_lt (vector unsigned int, vector unsigned int);
6495 vector signed int vec_any_lt (vector signed int, vector signed int);
6496 vector signed int vec_any_lt (vector float, vector float);
6498 vector signed int vec_any_nan (vector float);
6500 vector signed int vec_any_ne (vector signed char, vector unsigned char);
6502 vector signed int vec_any_ne (vector signed char, vector signed char);
6503 vector signed int vec_any_ne (vector unsigned char, vector signed char);
6505 vector signed int vec_any_ne (vector unsigned char,
6506 vector unsigned char);
6507 vector signed int vec_any_ne (vector signed short,
6508 vector unsigned short);
6509 vector signed int vec_any_ne (vector signed short, vector signed short);
6511 vector signed int vec_any_ne (vector unsigned short,
6512 vector signed short);
6513 vector signed int vec_any_ne (vector unsigned short,
6514 vector unsigned short);
6515 vector signed int vec_any_ne (vector signed int, vector unsigned int);
6516 vector signed int vec_any_ne (vector signed int, vector signed int);
6517 vector signed int vec_any_ne (vector unsigned int, vector signed int);
6518 vector signed int vec_any_ne (vector unsigned int, vector unsigned int);
6520 vector signed int vec_any_ne (vector float, vector float);
6522 vector signed int vec_any_nge (vector float, vector float);
6524 vector signed int vec_any_ngt (vector float, vector float);
6526 vector signed int vec_any_nle (vector float, vector float);
6528 vector signed int vec_any_nlt (vector float, vector float);
6530 vector signed int vec_any_numeric (vector float);
6532 vector signed int vec_any_out (vector float, vector float);
6536 @section Pragmas Accepted by GCC
6540 GCC supports several types of pragmas, primarily in order to compile
6541 code originally written for other compilers. Note that in general
6542 we do not recommend the use of pragmas; @xref{Function Attributes},
6543 for further explanation.
6547 * RS/6000 and PowerPC Pragmas::
6554 @subsection ARM Pragmas
6556 The ARM target defines pragmas for controlling the default addition of
6557 @code{long_call} and @code{short_call} attributes to functions.
6558 @xref{Function Attributes}, for information about the effects of these
6563 @cindex pragma, long_calls
6564 Set all subsequent functions to have the @code{long_call} attribute.
6567 @cindex pragma, no_long_calls
6568 Set all subsequent functions to have the @code{short_call} attribute.
6570 @item long_calls_off
6571 @cindex pragma, long_calls_off
6572 Do not affect the @code{long_call} or @code{short_call} attributes of
6573 subsequent functions.
6576 @node RS/6000 and PowerPC Pragmas
6577 @subsection RS/6000 and PowerPC Pragmas
6579 The RS/6000 and PowerPC targets define one pragma for controlling
6580 whether or not the @code{longcall} attribute is added to function
6581 declarations by default. This pragma overrides the @option{-mlongcall}
6582 option, but not the @code{longcall} and @code{shortcall} attributes.
6583 @xref{RS/6000 and PowerPC Options}, for more information about when long
6584 calls are and are not necessary.
6588 @cindex pragma, longcall
6589 Apply the @code{longcall} attribute to all subsequent function
6593 Do not apply the @code{longcall} attribute to subsequent function
6597 @c Describe c4x pragmas here.
6598 @c Describe h8300 pragmas here.
6599 @c Describe i370 pragmas here.
6600 @c Describe i960 pragmas here.
6601 @c Describe sh pragmas here.
6602 @c Describe v850 pragmas here.
6604 @node Darwin Pragmas
6605 @subsection Darwin Pragmas
6607 The following pragmas are available for all architectures running the
6608 Darwin operating system. These are useful for compatibility with other
6612 @item mark @var{tokens}@dots{}
6613 @cindex pragma, mark
6614 This pragma is accepted, but has no effect.
6616 @item options align=@var{alignment}
6617 @cindex pragma, options align
6618 This pragma sets the alignment of fields in structures. The values of
6619 @var{alignment} may be @code{mac68k}, to emulate m68k alignment, or
6620 @code{power}, to emulate PowerPC alignment. Uses of this pragma nest
6621 properly; to restore the previous setting, use @code{reset} for the
6624 @item segment @var{tokens}@dots{}
6625 @cindex pragma, segment
6626 This pragma is accepted, but has no effect.
6628 @item unused (@var{var} [, @var{var}]@dots{})
6629 @cindex pragma, unused
6630 This pragma declares variables to be possibly unused. GCC will not
6631 produce warnings for the listed variables. The effect is similar to
6632 that of the @code{unused} attribute, except that this pragma may appear
6633 anywhere within the variables' scopes.
6636 @node Solaris Pragmas
6637 @subsection Solaris Pragmas
6639 For compatibility with the SunPRO compiler, the following pragma
6643 @item redefine_extname @var{oldname} @var{newname}
6644 @cindex pragma, redefine_extname
6646 This pragma gives the C function @var{oldname} the assembler label
6647 @var{newname}. The pragma must appear before the function declaration.
6648 This pragma is equivalent to the asm labels extension (@pxref{Asm
6649 Labels}). The preprocessor defines @code{__PRAGMA_REDEFINE_EXTNAME}
6650 if the pragma is available.
6654 @subsection Tru64 Pragmas
6656 For compatibility with the Compaq C compiler, the following pragma
6660 @item extern_prefix @var{string}
6661 @cindex pragma, extern_prefix
6663 This pragma renames all subsequent function and variable declarations
6664 such that @var{string} is prepended to the name. This effect may be
6665 terminated by using another @code{extern_prefix} pragma with the
6668 This pragma is similar in intent to to the asm labels extension
6669 (@pxref{Asm Labels}) in that the system programmer wants to change
6670 the assembly-level ABI without changing the source-level API. The
6671 preprocessor defines @code{__PRAGMA_EXTERN_PREFIX} if the pragma is
6675 @node Unnamed Fields
6676 @section Unnamed struct/union fields within structs/unions.
6680 For compatibility with other compilers, GCC allows you to define
6681 a structure or union that contains, as fields, structures and unions
6682 without names. For example:
6695 In this example, the user would be able to access members of the unnamed
6696 union with code like @samp{foo.b}. Note that only unnamed structs and
6697 unions are allowed, you may not have, for example, an unnamed
6700 You must never create such structures that cause ambiguous field definitions.
6701 For example, this structure:
6712 It is ambiguous which @code{a} is being referred to with @samp{foo.a}.
6713 Such constructs are not supported and must be avoided. In the future,
6714 such constructs may be detected and treated as compilation errors.
6717 @section Thread-Local Storage
6718 @cindex Thread-Local Storage
6719 @cindex @acronym{TLS}
6722 Thread-local storage (@acronym{TLS}) is a mechanism by which variables
6723 are allocated such that there is one instance of the variable per extant
6724 thread. The run-time model GCC uses to implement this originates
6725 in the IA-64 processor-specific ABI, but has since been migrated
6726 to other processors as well. It requires significant support from
6727 the linker (@command{ld}), dynamic linker (@command{ld.so}), and
6728 system libraries (@file{libc.so} and @file{libpthread.so}), so it
6729 is not available everywhere.
6731 At the user level, the extension is visible with a new storage
6732 class keyword: @code{__thread}. For example:
6736 extern __thread struct state s;
6737 static __thread char *p;
6740 The @code{__thread} specifier may be used alone, with the @code{extern}
6741 or @code{static} specifiers, but with no other storage class specifier.
6742 When used with @code{extern} or @code{static}, @code{__thread} must appear
6743 immediately after the other storage class specifier.
6745 The @code{__thread} specifier may be applied to any global, file-scoped
6746 static, function-scoped static, or static data member of a class. It may
6747 not be applied to block-scoped automatic or non-static data member.
6749 When the address-of operator is applied to a thread-local variable, it is
6750 evaluated at run-time and returns the address of the current thread's
6751 instance of that variable. An address so obtained may be used by any
6752 thread. When a thread terminates, any pointers to thread-local variables
6753 in that thread become invalid.
6755 No static initialization may refer to the address of a thread-local variable.
6757 In C++, if an initializer is present for a thread-local variable, it must
6758 be a @var{constant-expression}, as defined in 5.19.2 of the ANSI/ISO C++
6761 See @uref{http://people.redhat.com/drepper/tls.pdf,
6762 ELF Handling For Thread-Local Storage} for a detailed explanation of
6763 the four thread-local storage addressing models, and how the run-time
6764 is expected to function.
6767 * C99 Thread-Local Edits::
6768 * C++98 Thread-Local Edits::
6771 @node C99 Thread-Local Edits
6772 @subsection ISO/IEC 9899:1999 Edits for Thread-Local Storage
6774 The following are a set of changes to ISO/IEC 9899:1999 (aka C99)
6775 that document the exact semantics of the language extension.
6779 @cite{5.1.2 Execution environments}
6781 Add new text after paragraph 1
6784 Within either execution environment, a @dfn{thread} is a flow of
6785 control within a program. It is implementation defined whether
6786 or not there may be more than one thread associated with a program.
6787 It is implementation defined how threads beyond the first are
6788 created, the name and type of the function called at thread
6789 startup, and how threads may be terminated. However, objects
6790 with thread storage duration shall be initialized before thread
6795 @cite{6.2.4 Storage durations of objects}
6797 Add new text before paragraph 3
6800 An object whose identifier is declared with the storage-class
6801 specifier @w{@code{__thread}} has @dfn{thread storage duration}.
6802 Its lifetime is the entire execution of the thread, and its
6803 stored value is initialized only once, prior to thread startup.
6807 @cite{6.4.1 Keywords}
6809 Add @code{__thread}.
6812 @cite{6.7.1 Storage-class specifiers}
6814 Add @code{__thread} to the list of storage class specifiers in
6817 Change paragraph 2 to
6820 With the exception of @code{__thread}, at most one storage-class
6821 specifier may be given [@dots{}]. The @code{__thread} specifier may
6822 be used alone, or immediately following @code{extern} or
6826 Add new text after paragraph 6
6829 The declaration of an identifier for a variable that has
6830 block scope that specifies @code{__thread} shall also
6831 specify either @code{extern} or @code{static}.
6833 The @code{__thread} specifier shall be used only with
6838 @node C++98 Thread-Local Edits
6839 @subsection ISO/IEC 14882:1998 Edits for Thread-Local Storage
6841 The following are a set of changes to ISO/IEC 14882:1998 (aka C++98)
6842 that document the exact semantics of the language extension.
6846 @b{[intro.execution]}
6848 New text after paragraph 4
6851 A @dfn{thread} is a flow of control within the abstract machine.
6852 It is implementation defined whether or not there may be more than
6856 New text after paragraph 7
6859 It is unspecified whether additional action must be taken to
6860 ensure when and whether side effects are visible to other threads.
6866 Add @code{__thread}.
6869 @b{[basic.start.main]}
6871 Add after paragraph 5
6874 The thread that begins execution at the @code{main} function is called
6875 the @dfn{main thread}. It is implementation defined how functions
6876 beginning threads other than the main thread are designated or typed.
6877 A function so designated, as well as the @code{main} function, is called
6878 a @dfn{thread startup function}. It is implementation defined what
6879 happens if a thread startup function returns. It is implementation
6880 defined what happens to other threads when any thread calls @code{exit}.
6884 @b{[basic.start.init]}
6886 Add after paragraph 4
6889 The storage for an object of thread storage duration shall be
6890 statically initialized before the first statement of the thread startup
6891 function. An object of thread storage duration shall not require
6892 dynamic initialization.
6896 @b{[basic.start.term]}
6898 Add after paragraph 3
6901 The type of an object with thread storage duration shall not have a
6902 non-trivial destructor, nor shall it be an array type whose elements
6903 (directly or indirectly) have non-trivial destructors.
6909 Add ``thread storage duration'' to the list in paragraph 1.
6914 Thread, static, and automatic storage durations are associated with
6915 objects introduced by declarations [@dots{}].
6918 Add @code{__thread} to the list of specifiers in paragraph 3.
6921 @b{[basic.stc.thread]}
6923 New section before @b{[basic.stc.static]}
6926 The keyword @code{__thread} applied to an non-local object gives the
6927 object thread storage duration.
6929 A local variable or class data member declared both @code{static}
6930 and @code{__thread} gives the variable or member thread storage
6935 @b{[basic.stc.static]}
6940 All objects which have neither thread storage duration, dynamic
6941 storage duration nor are local [@dots{}].
6947 Add @code{__thread} to the list in paragraph 1.
6952 With the exception of @code{__thread}, at most one
6953 @var{storage-class-specifier} shall appear in a given
6954 @var{decl-specifier-seq}. The @code{__thread} specifier may
6955 be used alone, or immediately following the @code{extern} or
6956 @code{static} specifiers. [@dots{}]
6959 Add after paragraph 5
6962 The @code{__thread} specifier can be applied only to the names of objects
6963 and to anonymous unions.
6969 Add after paragraph 6
6972 Non-@code{static} members shall not be @code{__thread}.
6976 @node C++ Extensions
6977 @chapter Extensions to the C++ Language
6978 @cindex extensions, C++ language
6979 @cindex C++ language extensions
6981 The GNU compiler provides these extensions to the C++ language (and you
6982 can also use most of the C language extensions in your C++ programs). If you
6983 want to write code that checks whether these features are available, you can
6984 test for the GNU compiler the same way as for C programs: check for a
6985 predefined macro @code{__GNUC__}. You can also use @code{__GNUG__} to
6986 test specifically for GNU C++ (@pxref{Standard Predefined,,Standard
6987 Predefined Macros,cpp.info,The C Preprocessor}).
6990 * Min and Max:: C++ Minimum and maximum operators.
6991 * Volatiles:: What constitutes an access to a volatile object.
6992 * Restricted Pointers:: C99 restricted pointers and references.
6993 * Vague Linkage:: Where G++ puts inlines, vtables and such.
6994 * C++ Interface:: You can use a single C++ header file for both
6995 declarations and definitions.
6996 * Template Instantiation:: Methods for ensuring that exactly one copy of
6997 each needed template instantiation is emitted.
6998 * Bound member functions:: You can extract a function pointer to the
6999 method denoted by a @samp{->*} or @samp{.*} expression.
7000 * C++ Attributes:: Variable, function, and type attributes for C++ only.
7001 * Java Exceptions:: Tweaking exception handling to work with Java.
7002 * Deprecated Features:: Things might disappear from g++.
7003 * Backwards Compatibility:: Compatibilities with earlier definitions of C++.
7007 @section Minimum and Maximum Operators in C++
7009 It is very convenient to have operators which return the ``minimum'' or the
7010 ``maximum'' of two arguments. In GNU C++ (but not in GNU C),
7013 @item @var{a} <? @var{b}
7015 @cindex minimum operator
7016 is the @dfn{minimum}, returning the smaller of the numeric values
7017 @var{a} and @var{b};
7019 @item @var{a} >? @var{b}
7021 @cindex maximum operator
7022 is the @dfn{maximum}, returning the larger of the numeric values @var{a}
7026 These operations are not primitive in ordinary C++, since you can
7027 use a macro to return the minimum of two things in C++, as in the
7031 #define MIN(X,Y) ((X) < (Y) ? : (X) : (Y))
7035 You might then use @w{@samp{int min = MIN (i, j);}} to set @var{min} to
7036 the minimum value of variables @var{i} and @var{j}.
7038 However, side effects in @code{X} or @code{Y} may cause unintended
7039 behavior. For example, @code{MIN (i++, j++)} will fail, incrementing
7040 the smaller counter twice. The GNU C @code{typeof} extension allows you
7041 to write safe macros that avoid this kind of problem (@pxref{Typeof}).
7042 However, writing @code{MIN} and @code{MAX} as macros also forces you to
7043 use function-call notation for a fundamental arithmetic operation.
7044 Using GNU C++ extensions, you can write @w{@samp{int min = i <? j;}}
7047 Since @code{<?} and @code{>?} are built into the compiler, they properly
7048 handle expressions with side-effects; @w{@samp{int min = i++ <? j++;}}
7052 @section When is a Volatile Object Accessed?
7053 @cindex accessing volatiles
7054 @cindex volatile read
7055 @cindex volatile write
7056 @cindex volatile access
7058 Both the C and C++ standard have the concept of volatile objects. These
7059 are normally accessed by pointers and used for accessing hardware. The
7060 standards encourage compilers to refrain from optimizations
7061 concerning accesses to volatile objects that it might perform on
7062 non-volatile objects. The C standard leaves it implementation defined
7063 as to what constitutes a volatile access. The C++ standard omits to
7064 specify this, except to say that C++ should behave in a similar manner
7065 to C with respect to volatiles, where possible. The minimum either
7066 standard specifies is that at a sequence point all previous accesses to
7067 volatile objects have stabilized and no subsequent accesses have
7068 occurred. Thus an implementation is free to reorder and combine
7069 volatile accesses which occur between sequence points, but cannot do so
7070 for accesses across a sequence point. The use of volatiles does not
7071 allow you to violate the restriction on updating objects multiple times
7072 within a sequence point.
7074 In most expressions, it is intuitively obvious what is a read and what is
7075 a write. For instance
7078 volatile int *dst = @var{somevalue};
7079 volatile int *src = @var{someothervalue};
7084 will cause a read of the volatile object pointed to by @var{src} and stores the
7085 value into the volatile object pointed to by @var{dst}. There is no
7086 guarantee that these reads and writes are atomic, especially for objects
7087 larger than @code{int}.
7089 Less obvious expressions are where something which looks like an access
7090 is used in a void context. An example would be,
7093 volatile int *src = @var{somevalue};
7097 With C, such expressions are rvalues, and as rvalues cause a read of
7098 the object, GCC interprets this as a read of the volatile being pointed
7099 to. The C++ standard specifies that such expressions do not undergo
7100 lvalue to rvalue conversion, and that the type of the dereferenced
7101 object may be incomplete. The C++ standard does not specify explicitly
7102 that it is this lvalue to rvalue conversion which is responsible for
7103 causing an access. However, there is reason to believe that it is,
7104 because otherwise certain simple expressions become undefined. However,
7105 because it would surprise most programmers, G++ treats dereferencing a
7106 pointer to volatile object of complete type in a void context as a read
7107 of the object. When the object has incomplete type, G++ issues a
7112 struct T @{int m;@};
7113 volatile S *ptr1 = @var{somevalue};
7114 volatile T *ptr2 = @var{somevalue};
7119 In this example, a warning is issued for @code{*ptr1}, and @code{*ptr2}
7120 causes a read of the object pointed to. If you wish to force an error on
7121 the first case, you must force a conversion to rvalue with, for instance
7122 a static cast, @code{static_cast<S>(*ptr1)}.
7124 When using a reference to volatile, G++ does not treat equivalent
7125 expressions as accesses to volatiles, but instead issues a warning that
7126 no volatile is accessed. The rationale for this is that otherwise it
7127 becomes difficult to determine where volatile access occur, and not
7128 possible to ignore the return value from functions returning volatile
7129 references. Again, if you wish to force a read, cast the reference to
7132 @node Restricted Pointers
7133 @section Restricting Pointer Aliasing
7134 @cindex restricted pointers
7135 @cindex restricted references
7136 @cindex restricted this pointer
7138 As with gcc, g++ understands the C99 feature of restricted pointers,
7139 specified with the @code{__restrict__}, or @code{__restrict} type
7140 qualifier. Because you cannot compile C++ by specifying the @option{-std=c99}
7141 language flag, @code{restrict} is not a keyword in C++.
7143 In addition to allowing restricted pointers, you can specify restricted
7144 references, which indicate that the reference is not aliased in the local
7148 void fn (int *__restrict__ rptr, int &__restrict__ rref)
7155 In the body of @code{fn}, @var{rptr} points to an unaliased integer and
7156 @var{rref} refers to a (different) unaliased integer.
7158 You may also specify whether a member function's @var{this} pointer is
7159 unaliased by using @code{__restrict__} as a member function qualifier.
7162 void T::fn () __restrict__
7169 Within the body of @code{T::fn}, @var{this} will have the effective
7170 definition @code{T *__restrict__ const this}. Notice that the
7171 interpretation of a @code{__restrict__} member function qualifier is
7172 different to that of @code{const} or @code{volatile} qualifier, in that it
7173 is applied to the pointer rather than the object. This is consistent with
7174 other compilers which implement restricted pointers.
7176 As with all outermost parameter qualifiers, @code{__restrict__} is
7177 ignored in function definition matching. This means you only need to
7178 specify @code{__restrict__} in a function definition, rather than
7179 in a function prototype as well.
7182 @section Vague Linkage
7183 @cindex vague linkage
7185 There are several constructs in C++ which require space in the object
7186 file but are not clearly tied to a single translation unit. We say that
7187 these constructs have ``vague linkage''. Typically such constructs are
7188 emitted wherever they are needed, though sometimes we can be more
7192 @item Inline Functions
7193 Inline functions are typically defined in a header file which can be
7194 included in many different compilations. Hopefully they can usually be
7195 inlined, but sometimes an out-of-line copy is necessary, if the address
7196 of the function is taken or if inlining fails. In general, we emit an
7197 out-of-line copy in all translation units where one is needed. As an
7198 exception, we only emit inline virtual functions with the vtable, since
7199 it will always require a copy.
7201 Local static variables and string constants used in an inline function
7202 are also considered to have vague linkage, since they must be shared
7203 between all inlined and out-of-line instances of the function.
7207 C++ virtual functions are implemented in most compilers using a lookup
7208 table, known as a vtable. The vtable contains pointers to the virtual
7209 functions provided by a class, and each object of the class contains a
7210 pointer to its vtable (or vtables, in some multiple-inheritance
7211 situations). If the class declares any non-inline, non-pure virtual
7212 functions, the first one is chosen as the ``key method'' for the class,
7213 and the vtable is only emitted in the translation unit where the key
7216 @emph{Note:} If the chosen key method is later defined as inline, the
7217 vtable will still be emitted in every translation unit which defines it.
7218 Make sure that any inline virtuals are declared inline in the class
7219 body, even if they are not defined there.
7221 @item type_info objects
7224 C++ requires information about types to be written out in order to
7225 implement @samp{dynamic_cast}, @samp{typeid} and exception handling.
7226 For polymorphic classes (classes with virtual functions), the type_info
7227 object is written out along with the vtable so that @samp{dynamic_cast}
7228 can determine the dynamic type of a class object at runtime. For all
7229 other types, we write out the type_info object when it is used: when
7230 applying @samp{typeid} to an expression, throwing an object, or
7231 referring to a type in a catch clause or exception specification.
7233 @item Template Instantiations
7234 Most everything in this section also applies to template instantiations,
7235 but there are other options as well.
7236 @xref{Template Instantiation,,Where's the Template?}.
7240 When used with GNU ld version 2.8 or later on an ELF system such as
7241 Linux/GNU or Solaris 2, or on Microsoft Windows, duplicate copies of
7242 these constructs will be discarded at link time. This is known as
7245 On targets that don't support COMDAT, but do support weak symbols, GCC
7246 will use them. This way one copy will override all the others, but
7247 the unused copies will still take up space in the executable.
7249 For targets which do not support either COMDAT or weak symbols,
7250 most entities with vague linkage will be emitted as local symbols to
7251 avoid duplicate definition errors from the linker. This will not happen
7252 for local statics in inlines, however, as having multiple copies will
7253 almost certainly break things.
7255 @xref{C++ Interface,,Declarations and Definitions in One Header}, for
7256 another way to control placement of these constructs.
7259 @section Declarations and Definitions in One Header
7261 @cindex interface and implementation headers, C++
7262 @cindex C++ interface and implementation headers
7263 C++ object definitions can be quite complex. In principle, your source
7264 code will need two kinds of things for each object that you use across
7265 more than one source file. First, you need an @dfn{interface}
7266 specification, describing its structure with type declarations and
7267 function prototypes. Second, you need the @dfn{implementation} itself.
7268 It can be tedious to maintain a separate interface description in a
7269 header file, in parallel to the actual implementation. It is also
7270 dangerous, since separate interface and implementation definitions may
7271 not remain parallel.
7273 @cindex pragmas, interface and implementation
7274 With GNU C++, you can use a single header file for both purposes.
7277 @emph{Warning:} The mechanism to specify this is in transition. For the
7278 nonce, you must use one of two @code{#pragma} commands; in a future
7279 release of GNU C++, an alternative mechanism will make these
7280 @code{#pragma} commands unnecessary.
7283 The header file contains the full definitions, but is marked with
7284 @samp{#pragma interface} in the source code. This allows the compiler
7285 to use the header file only as an interface specification when ordinary
7286 source files incorporate it with @code{#include}. In the single source
7287 file where the full implementation belongs, you can use either a naming
7288 convention or @samp{#pragma implementation} to indicate this alternate
7289 use of the header file.
7292 @item #pragma interface
7293 @itemx #pragma interface "@var{subdir}/@var{objects}.h"
7294 @kindex #pragma interface
7295 Use this directive in @emph{header files} that define object classes, to save
7296 space in most of the object files that use those classes. Normally,
7297 local copies of certain information (backup copies of inline member
7298 functions, debugging information, and the internal tables that implement
7299 virtual functions) must be kept in each object file that includes class
7300 definitions. You can use this pragma to avoid such duplication. When a
7301 header file containing @samp{#pragma interface} is included in a
7302 compilation, this auxiliary information will not be generated (unless
7303 the main input source file itself uses @samp{#pragma implementation}).
7304 Instead, the object files will contain references to be resolved at link
7307 The second form of this directive is useful for the case where you have
7308 multiple headers with the same name in different directories. If you
7309 use this form, you must specify the same string to @samp{#pragma
7312 @item #pragma implementation
7313 @itemx #pragma implementation "@var{objects}.h"
7314 @kindex #pragma implementation
7315 Use this pragma in a @emph{main input file}, when you want full output from
7316 included header files to be generated (and made globally visible). The
7317 included header file, in turn, should use @samp{#pragma interface}.
7318 Backup copies of inline member functions, debugging information, and the
7319 internal tables used to implement virtual functions are all generated in
7320 implementation files.
7322 @cindex implied @code{#pragma implementation}
7323 @cindex @code{#pragma implementation}, implied
7324 @cindex naming convention, implementation headers
7325 If you use @samp{#pragma implementation} with no argument, it applies to
7326 an include file with the same basename@footnote{A file's @dfn{basename}
7327 was the name stripped of all leading path information and of trailing
7328 suffixes, such as @samp{.h} or @samp{.C} or @samp{.cc}.} as your source
7329 file. For example, in @file{allclass.cc}, giving just
7330 @samp{#pragma implementation}
7331 by itself is equivalent to @samp{#pragma implementation "allclass.h"}.
7333 In versions of GNU C++ prior to 2.6.0 @file{allclass.h} was treated as
7334 an implementation file whenever you would include it from
7335 @file{allclass.cc} even if you never specified @samp{#pragma
7336 implementation}. This was deemed to be more trouble than it was worth,
7337 however, and disabled.
7339 If you use an explicit @samp{#pragma implementation}, it must appear in
7340 your source file @emph{before} you include the affected header files.
7342 Use the string argument if you want a single implementation file to
7343 include code from multiple header files. (You must also use
7344 @samp{#include} to include the header file; @samp{#pragma
7345 implementation} only specifies how to use the file---it doesn't actually
7348 There is no way to split up the contents of a single header file into
7349 multiple implementation files.
7352 @cindex inlining and C++ pragmas
7353 @cindex C++ pragmas, effect on inlining
7354 @cindex pragmas in C++, effect on inlining
7355 @samp{#pragma implementation} and @samp{#pragma interface} also have an
7356 effect on function inlining.
7358 If you define a class in a header file marked with @samp{#pragma
7359 interface}, the effect on a function defined in that class is similar to
7360 an explicit @code{extern} declaration---the compiler emits no code at
7361 all to define an independent version of the function. Its definition
7362 is used only for inlining with its callers.
7364 @opindex fno-implement-inlines
7365 Conversely, when you include the same header file in a main source file
7366 that declares it as @samp{#pragma implementation}, the compiler emits
7367 code for the function itself; this defines a version of the function
7368 that can be found via pointers (or by callers compiled without
7369 inlining). If all calls to the function can be inlined, you can avoid
7370 emitting the function by compiling with @option{-fno-implement-inlines}.
7371 If any calls were not inlined, you will get linker errors.
7373 @node Template Instantiation
7374 @section Where's the Template?
7375 @cindex template instantiation
7377 C++ templates are the first language feature to require more
7378 intelligence from the environment than one usually finds on a UNIX
7379 system. Somehow the compiler and linker have to make sure that each
7380 template instance occurs exactly once in the executable if it is needed,
7381 and not at all otherwise. There are two basic approaches to this
7382 problem, which I will refer to as the Borland model and the Cfront model.
7386 Borland C++ solved the template instantiation problem by adding the code
7387 equivalent of common blocks to their linker; the compiler emits template
7388 instances in each translation unit that uses them, and the linker
7389 collapses them together. The advantage of this model is that the linker
7390 only has to consider the object files themselves; there is no external
7391 complexity to worry about. This disadvantage is that compilation time
7392 is increased because the template code is being compiled repeatedly.
7393 Code written for this model tends to include definitions of all
7394 templates in the header file, since they must be seen to be
7398 The AT&T C++ translator, Cfront, solved the template instantiation
7399 problem by creating the notion of a template repository, an
7400 automatically maintained place where template instances are stored. A
7401 more modern version of the repository works as follows: As individual
7402 object files are built, the compiler places any template definitions and
7403 instantiations encountered in the repository. At link time, the link
7404 wrapper adds in the objects in the repository and compiles any needed
7405 instances that were not previously emitted. The advantages of this
7406 model are more optimal compilation speed and the ability to use the
7407 system linker; to implement the Borland model a compiler vendor also
7408 needs to replace the linker. The disadvantages are vastly increased
7409 complexity, and thus potential for error; for some code this can be
7410 just as transparent, but in practice it can been very difficult to build
7411 multiple programs in one directory and one program in multiple
7412 directories. Code written for this model tends to separate definitions
7413 of non-inline member templates into a separate file, which should be
7414 compiled separately.
7417 When used with GNU ld version 2.8 or later on an ELF system such as
7418 Linux/GNU or Solaris 2, or on Microsoft Windows, g++ supports the
7419 Borland model. On other systems, g++ implements neither automatic
7422 A future version of g++ will support a hybrid model whereby the compiler
7423 will emit any instantiations for which the template definition is
7424 included in the compile, and store template definitions and
7425 instantiation context information into the object file for the rest.
7426 The link wrapper will extract that information as necessary and invoke
7427 the compiler to produce the remaining instantiations. The linker will
7428 then combine duplicate instantiations.
7430 In the mean time, you have the following options for dealing with
7431 template instantiations:
7436 Compile your template-using code with @option{-frepo}. The compiler will
7437 generate files with the extension @samp{.rpo} listing all of the
7438 template instantiations used in the corresponding object files which
7439 could be instantiated there; the link wrapper, @samp{collect2}, will
7440 then update the @samp{.rpo} files to tell the compiler where to place
7441 those instantiations and rebuild any affected object files. The
7442 link-time overhead is negligible after the first pass, as the compiler
7443 will continue to place the instantiations in the same files.
7445 This is your best option for application code written for the Borland
7446 model, as it will just work. Code written for the Cfront model will
7447 need to be modified so that the template definitions are available at
7448 one or more points of instantiation; usually this is as simple as adding
7449 @code{#include <tmethods.cc>} to the end of each template header.
7451 For library code, if you want the library to provide all of the template
7452 instantiations it needs, just try to link all of its object files
7453 together; the link will fail, but cause the instantiations to be
7454 generated as a side effect. Be warned, however, that this may cause
7455 conflicts if multiple libraries try to provide the same instantiations.
7456 For greater control, use explicit instantiation as described in the next
7460 @opindex fno-implicit-templates
7461 Compile your code with @option{-fno-implicit-templates} to disable the
7462 implicit generation of template instances, and explicitly instantiate
7463 all the ones you use. This approach requires more knowledge of exactly
7464 which instances you need than do the others, but it's less
7465 mysterious and allows greater control. You can scatter the explicit
7466 instantiations throughout your program, perhaps putting them in the
7467 translation units where the instances are used or the translation units
7468 that define the templates themselves; you can put all of the explicit
7469 instantiations you need into one big file; or you can create small files
7476 template class Foo<int>;
7477 template ostream& operator <<
7478 (ostream&, const Foo<int>&);
7481 for each of the instances you need, and create a template instantiation
7484 If you are using Cfront-model code, you can probably get away with not
7485 using @option{-fno-implicit-templates} when compiling files that don't
7486 @samp{#include} the member template definitions.
7488 If you use one big file to do the instantiations, you may want to
7489 compile it without @option{-fno-implicit-templates} so you get all of the
7490 instances required by your explicit instantiations (but not by any
7491 other files) without having to specify them as well.
7493 g++ has extended the template instantiation syntax given in the ISO
7494 standard to allow forward declaration of explicit instantiations
7495 (with @code{extern}), instantiation of the compiler support data for a
7496 template class (i.e.@: the vtable) without instantiating any of its
7497 members (with @code{inline}), and instantiation of only the static data
7498 members of a template class, without the support data or member
7499 functions (with (@code{static}):
7502 extern template int max (int, int);
7503 inline template class Foo<int>;
7504 static template class Foo<int>;
7508 Do nothing. Pretend g++ does implement automatic instantiation
7509 management. Code written for the Borland model will work fine, but
7510 each translation unit will contain instances of each of the templates it
7511 uses. In a large program, this can lead to an unacceptable amount of code
7514 @xref{C++ Interface,,Declarations and Definitions in One Header}, for
7515 more discussion of these pragmas.
7518 @node Bound member functions
7519 @section Extracting the function pointer from a bound pointer to member function
7521 @cindex pointer to member function
7522 @cindex bound pointer to member function
7524 In C++, pointer to member functions (PMFs) are implemented using a wide
7525 pointer of sorts to handle all the possible call mechanisms; the PMF
7526 needs to store information about how to adjust the @samp{this} pointer,
7527 and if the function pointed to is virtual, where to find the vtable, and
7528 where in the vtable to look for the member function. If you are using
7529 PMFs in an inner loop, you should really reconsider that decision. If
7530 that is not an option, you can extract the pointer to the function that
7531 would be called for a given object/PMF pair and call it directly inside
7532 the inner loop, to save a bit of time.
7534 Note that you will still be paying the penalty for the call through a
7535 function pointer; on most modern architectures, such a call defeats the
7536 branch prediction features of the CPU@. This is also true of normal
7537 virtual function calls.
7539 The syntax for this extension is
7543 extern int (A::*fp)();
7544 typedef int (*fptr)(A *);
7546 fptr p = (fptr)(a.*fp);
7549 For PMF constants (i.e.@: expressions of the form @samp{&Klasse::Member}),
7550 no object is needed to obtain the address of the function. They can be
7551 converted to function pointers directly:
7554 fptr p1 = (fptr)(&A::foo);
7557 @opindex Wno-pmf-conversions
7558 You must specify @option{-Wno-pmf-conversions} to use this extension.
7560 @node C++ Attributes
7561 @section C++-Specific Variable, Function, and Type Attributes
7563 Some attributes only make sense for C++ programs.
7566 @item init_priority (@var{priority})
7567 @cindex init_priority attribute
7570 In Standard C++, objects defined at namespace scope are guaranteed to be
7571 initialized in an order in strict accordance with that of their definitions
7572 @emph{in a given translation unit}. No guarantee is made for initializations
7573 across translation units. However, GNU C++ allows users to control the
7574 order of initialization of objects defined at namespace scope with the
7575 @code{init_priority} attribute by specifying a relative @var{priority},
7576 a constant integral expression currently bounded between 101 and 65535
7577 inclusive. Lower numbers indicate a higher priority.
7579 In the following example, @code{A} would normally be created before
7580 @code{B}, but the @code{init_priority} attribute has reversed that order:
7583 Some_Class A __attribute__ ((init_priority (2000)));
7584 Some_Class B __attribute__ ((init_priority (543)));
7588 Note that the particular values of @var{priority} do not matter; only their
7591 @item java_interface
7592 @cindex java_interface attribute
7594 This type attribute informs C++ that the class is a Java interface. It may
7595 only be applied to classes declared within an @code{extern "Java"} block.
7596 Calls to methods declared in this interface will be dispatched using GCJ's
7597 interface table mechanism, instead of regular virtual table dispatch.
7601 @node Java Exceptions
7602 @section Java Exceptions
7604 The Java language uses a slightly different exception handling model
7605 from C++. Normally, GNU C++ will automatically detect when you are
7606 writing C++ code that uses Java exceptions, and handle them
7607 appropriately. However, if C++ code only needs to execute destructors
7608 when Java exceptions are thrown through it, GCC will guess incorrectly.
7609 Sample problematic code is:
7612 struct S @{ ~S(); @};
7613 extern void bar(); // is written in Java, and may throw exceptions
7622 The usual effect of an incorrect guess is a link failure, complaining of
7623 a missing routine called @samp{__gxx_personality_v0}.
7625 You can inform the compiler that Java exceptions are to be used in a
7626 translation unit, irrespective of what it might think, by writing
7627 @samp{@w{#pragma GCC java_exceptions}} at the head of the file. This
7628 @samp{#pragma} must appear before any functions that throw or catch
7629 exceptions, or run destructors when exceptions are thrown through them.
7631 You cannot mix Java and C++ exceptions in the same translation unit. It
7632 is believed to be safe to throw a C++ exception from one file through
7633 another file compiled for the Java exception model, or vice versa, but
7634 there may be bugs in this area.
7636 @node Deprecated Features
7637 @section Deprecated Features
7639 In the past, the GNU C++ compiler was extended to experiment with new
7640 features, at a time when the C++ language was still evolving. Now that
7641 the C++ standard is complete, some of those features are superseded by
7642 superior alternatives. Using the old features might cause a warning in
7643 some cases that the feature will be dropped in the future. In other
7644 cases, the feature might be gone already.
7646 While the list below is not exhaustive, it documents some of the options
7647 that are now deprecated:
7650 @item -fexternal-templates
7651 @itemx -falt-external-templates
7652 These are two of the many ways for g++ to implement template
7653 instantiation. @xref{Template Instantiation}. The C++ standard clearly
7654 defines how template definitions have to be organized across
7655 implementation units. g++ has an implicit instantiation mechanism that
7656 should work just fine for standard-conforming code.
7658 @item -fstrict-prototype
7659 @itemx -fno-strict-prototype
7660 Previously it was possible to use an empty prototype parameter list to
7661 indicate an unspecified number of parameters (like C), rather than no
7662 parameters, as C++ demands. This feature has been removed, except where
7663 it is required for backwards compatibility @xref{Backwards Compatibility}.
7666 The named return value extension has been deprecated, and is now
7669 The use of initializer lists with new expressions has been deprecated,
7670 and is now removed from g++.
7672 Floating and complex non-type template parameters have been deprecated,
7673 and are now removed from g++.
7675 The implicit typename extension has been deprecated and will be removed
7676 from g++ at some point. In some cases g++ determines that a dependent
7677 type such as @code{TPL<T>::X} is a type without needing a
7678 @code{typename} keyword, contrary to the standard.
7680 @node Backwards Compatibility
7681 @section Backwards Compatibility
7682 @cindex Backwards Compatibility
7683 @cindex ARM [Annotated C++ Reference Manual]
7685 Now that there is a definitive ISO standard C++, G++ has a specification
7686 to adhere to. The C++ language evolved over time, and features that
7687 used to be acceptable in previous drafts of the standard, such as the ARM
7688 [Annotated C++ Reference Manual], are no longer accepted. In order to allow
7689 compilation of C++ written to such drafts, G++ contains some backwards
7690 compatibilities. @emph{All such backwards compatibility features are
7691 liable to disappear in future versions of G++.} They should be considered
7692 deprecated @xref{Deprecated Features}.
7696 If a variable is declared at for scope, it used to remain in scope until
7697 the end of the scope which contained the for statement (rather than just
7698 within the for scope). G++ retains this, but issues a warning, if such a
7699 variable is accessed outside the for scope.
7701 @item Implicit C language
7702 Old C system header files did not contain an @code{extern "C" @{@dots{}@}}
7703 scope to set the language. On such systems, all header files are
7704 implicitly scoped inside a C language scope. Also, an empty prototype
7705 @code{()} will be treated as an unspecified number of arguments, rather
7706 than no arguments, as C++ demands.