]>
Commit | Line | Data |
---|---|---|
8d8da227 JM |
1 | @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1996, 1998, 1999, 2000, 2001, |
2 | @c 2002, 2003, 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc. | |
9a8ce21f | 3 | |
c1f7febf RK |
4 | @c This is part of the GCC manual. |
5 | @c For copying conditions, see the file gcc.texi. | |
6 | ||
7 | @node C Extensions | |
8 | @chapter Extensions to the C Language Family | |
9 | @cindex extensions, C language | |
10 | @cindex C language extensions | |
11 | ||
84330467 | 12 | @opindex pedantic |
161d7b59 | 13 | GNU C provides several language features not found in ISO standard C@. |
f0523f02 | 14 | (The @option{-pedantic} option directs GCC to print a warning message if |
c1f7febf RK |
15 | any of these features is used.) To test for the availability of these |
16 | features in conditional compilation, check for a predefined macro | |
161d7b59 | 17 | @code{__GNUC__}, which is always defined under GCC@. |
c1f7febf | 18 | |
161d7b59 | 19 | These extensions are available in C and Objective-C@. Most of them are |
c1f7febf RK |
20 | also available in C++. @xref{C++ Extensions,,Extensions to the |
21 | C++ Language}, for extensions that apply @emph{only} to C++. | |
22 | ||
4b404517 JM |
23 | Some features that are in ISO C99 but not C89 or C++ are also, as |
24 | extensions, accepted by GCC in C89 mode and in C++. | |
5490d604 | 25 | |
c1f7febf RK |
26 | @menu |
27 | * Statement Exprs:: Putting statements and declarations inside expressions. | |
14e33ee8 | 28 | * Local Labels:: Labels local to a block. |
c1f7febf RK |
29 | * Labels as Values:: Getting pointers to labels, and computed gotos. |
30 | * Nested Functions:: As in Algol and Pascal, lexical scoping of functions. | |
6ccde948 | 31 | * Constructing Calls:: Dispatching a call to another function. |
c1f7febf | 32 | * Typeof:: @code{typeof}: referring to the type of an expression. |
c1f7febf | 33 | * Conditionals:: Omitting the middle operand of a @samp{?:} expression. |
6ccde948 | 34 | * Long Long:: Double-word integers---@code{long long int}. |
c1f7febf | 35 | * Complex:: Data types for complex numbers. |
c77cd3d1 | 36 | * Floating Types:: Additional Floating Types. |
85a92f7e | 37 | * Decimal Float:: Decimal Floating Types. |
6f4d7222 | 38 | * Hex Floats:: Hexadecimal floating-point constants. |
0f996086 | 39 | * Fixed-Point:: Fixed-Point Types. |
c1f7febf RK |
40 | * Zero Length:: Zero-length arrays. |
41 | * Variable Length:: Arrays whose length is computed at run time. | |
ba05abd3 | 42 | * Empty Structures:: Structures with no members. |
6ccde948 | 43 | * Variadic Macros:: Macros with a variable number of arguments. |
ccd96f0a | 44 | * Escaped Newlines:: Slightly looser rules for escaped newlines. |
c1f7febf RK |
45 | * Subscripting:: Any array can be subscripted, even if not an lvalue. |
46 | * Pointer Arith:: Arithmetic on @code{void}-pointers and function pointers. | |
47 | * Initializers:: Non-constant initializers. | |
4b404517 | 48 | * Compound Literals:: Compound literals give structures, unions |
6ccde948 RW |
49 | or arrays as values. |
50 | * Designated Inits:: Labeling elements of initializers. | |
c1f7febf | 51 | * Cast to Union:: Casting to union type from any member of the union. |
6ccde948 RW |
52 | * Case Ranges:: `case 1 ... 9' and such. |
53 | * Mixed Declarations:: Mixing declarations and code. | |
c1f7febf | 54 | * Function Attributes:: Declaring that functions have no side effects, |
6ccde948 | 55 | or that they can never return. |
2c5e91d2 | 56 | * Attribute Syntax:: Formal syntax for attributes. |
c1f7febf RK |
57 | * Function Prototypes:: Prototype declarations and old-style definitions. |
58 | * C++ Comments:: C++ comments are recognized. | |
59 | * Dollar Signs:: Dollar sign is allowed in identifiers. | |
60 | * Character Escapes:: @samp{\e} stands for the character @key{ESC}. | |
6ccde948 RW |
61 | * Variable Attributes:: Specifying attributes of variables. |
62 | * Type Attributes:: Specifying attributes of types. | |
c1f7febf RK |
63 | * Alignment:: Inquiring about the alignment of a type or variable. |
64 | * Inline:: Defining inline functions (as fast as macros). | |
65 | * Extended Asm:: Assembler instructions with C expressions as operands. | |
6ccde948 | 66 | (With them you can define ``built-in'' functions.) |
c1f7febf RK |
67 | * Constraints:: Constraints for asm operands |
68 | * Asm Labels:: Specifying the assembler name to use for a C symbol. | |
69 | * Explicit Reg Vars:: Defining variables residing in specified registers. | |
70 | * Alternate Keywords:: @code{__const__}, @code{__asm__}, etc., for header files. | |
71 | * Incomplete Enums:: @code{enum foo;}, with details to follow. | |
6ccde948 RW |
72 | * Function Names:: Printable strings which are the name of the current |
73 | function. | |
c1f7febf | 74 | * Return Address:: Getting the return or frame address of a function. |
1255c85c | 75 | * Vector Extensions:: Using vector instructions through built-in functions. |
7a3ea201 | 76 | * Offsetof:: Special syntax for implementing @code{offsetof}. |
6ccde948 | 77 | * Atomic Builtins:: Built-in functions for atomic memory access. |
10a0d495 JJ |
78 | * Object Size Checking:: Built-in functions for limited buffer overflow |
79 | checking. | |
c5c76735 | 80 | * Other Builtins:: Other built-in functions. |
0975678f | 81 | * Target Builtins:: Built-in functions specific to particular targets. |
a2bec818 | 82 | * Target Format Checks:: Format checks specific to particular targets. |
0168a849 | 83 | * Pragmas:: Pragmas accepted by GCC. |
b11cc610 | 84 | * Unnamed Fields:: Unnamed struct/union fields within structs/unions. |
3d78f2e9 | 85 | * Thread-Local:: Per-thread variables. |
f7fd775f | 86 | * Binary constants:: Binary constants using the @samp{0b} prefix. |
c1f7febf | 87 | @end menu |
c1f7febf RK |
88 | |
89 | @node Statement Exprs | |
90 | @section Statements and Declarations in Expressions | |
91 | @cindex statements inside expressions | |
92 | @cindex declarations inside expressions | |
93 | @cindex expressions containing statements | |
94 | @cindex macros, statements in expressions | |
95 | ||
96 | @c the above section title wrapped and causes an underfull hbox.. i | |
97 | @c changed it from "within" to "in". --mew 4feb93 | |
c1f7febf | 98 | A compound statement enclosed in parentheses may appear as an expression |
161d7b59 | 99 | in GNU C@. This allows you to use loops, switches, and local variables |
c1f7febf RK |
100 | within an expression. |
101 | ||
102 | Recall that a compound statement is a sequence of statements surrounded | |
103 | by braces; in this construct, parentheses go around the braces. For | |
104 | example: | |
105 | ||
3ab51846 | 106 | @smallexample |
c1f7febf RK |
107 | (@{ int y = foo (); int z; |
108 | if (y > 0) z = y; | |
109 | else z = - y; | |
110 | z; @}) | |
3ab51846 | 111 | @end smallexample |
c1f7febf RK |
112 | |
113 | @noindent | |
114 | is a valid (though slightly more complex than necessary) expression | |
115 | for the absolute value of @code{foo ()}. | |
116 | ||
117 | The last thing in the compound statement should be an expression | |
118 | followed by a semicolon; the value of this subexpression serves as the | |
119 | value of the entire construct. (If you use some other kind of statement | |
120 | last within the braces, the construct has type @code{void}, and thus | |
121 | effectively no value.) | |
122 | ||
123 | This feature is especially useful in making macro definitions ``safe'' (so | |
124 | that they evaluate each operand exactly once). For example, the | |
125 | ``maximum'' function is commonly defined as a macro in standard C as | |
126 | follows: | |
127 | ||
3ab51846 | 128 | @smallexample |
c1f7febf | 129 | #define max(a,b) ((a) > (b) ? (a) : (b)) |
3ab51846 | 130 | @end smallexample |
c1f7febf RK |
131 | |
132 | @noindent | |
133 | @cindex side effects, macro argument | |
134 | But this definition computes either @var{a} or @var{b} twice, with bad | |
135 | results if the operand has side effects. In GNU C, if you know the | |
962e6e00 | 136 | type of the operands (here taken as @code{int}), you can define |
c1f7febf RK |
137 | the macro safely as follows: |
138 | ||
3ab51846 | 139 | @smallexample |
c1f7febf RK |
140 | #define maxint(a,b) \ |
141 | (@{int _a = (a), _b = (b); _a > _b ? _a : _b; @}) | |
3ab51846 | 142 | @end smallexample |
c1f7febf RK |
143 | |
144 | Embedded statements are not allowed in constant expressions, such as | |
c771326b | 145 | the value of an enumeration constant, the width of a bit-field, or |
c1f7febf RK |
146 | the initial value of a static variable. |
147 | ||
148 | If you don't know the type of the operand, you can still do this, but you | |
95f79357 | 149 | must use @code{typeof} (@pxref{Typeof}). |
c1f7febf | 150 | |
a5bcc582 NS |
151 | In G++, the result value of a statement expression undergoes array and |
152 | function pointer decay, and is returned by value to the enclosing | |
8a36672b | 153 | expression. For instance, if @code{A} is a class, then |
b98e139b | 154 | |
a5bcc582 NS |
155 | @smallexample |
156 | A a; | |
b98e139b | 157 | |
a5bcc582 NS |
158 | (@{a;@}).Foo () |
159 | @end smallexample | |
b98e139b MM |
160 | |
161 | @noindent | |
a5bcc582 NS |
162 | will construct a temporary @code{A} object to hold the result of the |
163 | statement expression, and that will be used to invoke @code{Foo}. | |
164 | Therefore the @code{this} pointer observed by @code{Foo} will not be the | |
165 | address of @code{a}. | |
166 | ||
167 | Any temporaries created within a statement within a statement expression | |
168 | will be destroyed at the statement's end. This makes statement | |
169 | expressions inside macros slightly different from function calls. In | |
170 | the latter case temporaries introduced during argument evaluation will | |
171 | be destroyed at the end of the statement that includes the function | |
172 | call. In the statement expression case they will be destroyed during | |
173 | the statement expression. For instance, | |
b98e139b | 174 | |
a5bcc582 NS |
175 | @smallexample |
176 | #define macro(a) (@{__typeof__(a) b = (a); b + 3; @}) | |
177 | template<typename T> T function(T a) @{ T b = a; return b + 3; @} | |
178 | ||
179 | void foo () | |
180 | @{ | |
181 | macro (X ()); | |
182 | function (X ()); | |
183 | @} | |
184 | @end smallexample | |
b98e139b MM |
185 | |
186 | @noindent | |
a5bcc582 NS |
187 | will have different places where temporaries are destroyed. For the |
188 | @code{macro} case, the temporary @code{X} will be destroyed just after | |
189 | the initialization of @code{b}. In the @code{function} case that | |
190 | temporary will be destroyed when the function returns. | |
b98e139b MM |
191 | |
192 | These considerations mean that it is probably a bad idea to use | |
193 | statement-expressions of this form in header files that are designed to | |
54e1d3a6 MM |
194 | work with C++. (Note that some versions of the GNU C Library contained |
195 | header files using statement-expression that lead to precisely this | |
196 | bug.) | |
b98e139b | 197 | |
16ef3acc JM |
198 | Jumping into a statement expression with @code{goto} or using a |
199 | @code{switch} statement outside the statement expression with a | |
200 | @code{case} or @code{default} label inside the statement expression is | |
201 | not permitted. Jumping into a statement expression with a computed | |
202 | @code{goto} (@pxref{Labels as Values}) yields undefined behavior. | |
203 | Jumping out of a statement expression is permitted, but if the | |
204 | statement expression is part of a larger expression then it is | |
205 | unspecified which other subexpressions of that expression have been | |
206 | evaluated except where the language definition requires certain | |
207 | subexpressions to be evaluated before or after the statement | |
208 | expression. In any case, as with a function call the evaluation of a | |
209 | statement expression is not interleaved with the evaluation of other | |
210 | parts of the containing expression. For example, | |
211 | ||
212 | @smallexample | |
213 | foo (), ((@{ bar1 (); goto a; 0; @}) + bar2 ()), baz(); | |
214 | @end smallexample | |
215 | ||
216 | @noindent | |
217 | will call @code{foo} and @code{bar1} and will not call @code{baz} but | |
218 | may or may not call @code{bar2}. If @code{bar2} is called, it will be | |
219 | called after @code{foo} and before @code{bar1} | |
220 | ||
c1f7febf RK |
221 | @node Local Labels |
222 | @section Locally Declared Labels | |
223 | @cindex local labels | |
224 | @cindex macros, local labels | |
225 | ||
14e33ee8 | 226 | GCC allows you to declare @dfn{local labels} in any nested block |
8a36672b | 227 | scope. A local label is just like an ordinary label, but you can |
14e33ee8 | 228 | only reference it (with a @code{goto} statement, or by taking its |
daf2f129 | 229 | address) within the block in which it was declared. |
c1f7febf RK |
230 | |
231 | A local label declaration looks like this: | |
232 | ||
3ab51846 | 233 | @smallexample |
c1f7febf | 234 | __label__ @var{label}; |
3ab51846 | 235 | @end smallexample |
c1f7febf RK |
236 | |
237 | @noindent | |
238 | or | |
239 | ||
3ab51846 | 240 | @smallexample |
0d893a63 | 241 | __label__ @var{label1}, @var{label2}, /* @r{@dots{}} */; |
3ab51846 | 242 | @end smallexample |
c1f7febf | 243 | |
14e33ee8 ZW |
244 | Local label declarations must come at the beginning of the block, |
245 | before any ordinary declarations or statements. | |
c1f7febf RK |
246 | |
247 | The label declaration defines the label @emph{name}, but does not define | |
248 | the label itself. You must do this in the usual way, with | |
249 | @code{@var{label}:}, within the statements of the statement expression. | |
250 | ||
14e33ee8 ZW |
251 | The local label feature is useful for complex macros. If a macro |
252 | contains nested loops, a @code{goto} can be useful for breaking out of | |
253 | them. However, an ordinary label whose scope is the whole function | |
254 | cannot be used: if the macro can be expanded several times in one | |
255 | function, the label will be multiply defined in that function. A | |
256 | local label avoids this problem. For example: | |
257 | ||
3ab51846 | 258 | @smallexample |
14e33ee8 ZW |
259 | #define SEARCH(value, array, target) \ |
260 | do @{ \ | |
261 | __label__ found; \ | |
262 | typeof (target) _SEARCH_target = (target); \ | |
263 | typeof (*(array)) *_SEARCH_array = (array); \ | |
264 | int i, j; \ | |
265 | int value; \ | |
266 | for (i = 0; i < max; i++) \ | |
267 | for (j = 0; j < max; j++) \ | |
268 | if (_SEARCH_array[i][j] == _SEARCH_target) \ | |
269 | @{ (value) = i; goto found; @} \ | |
270 | (value) = -1; \ | |
271 | found:; \ | |
272 | @} while (0) | |
3ab51846 | 273 | @end smallexample |
14e33ee8 ZW |
274 | |
275 | This could also be written using a statement-expression: | |
c1f7febf | 276 | |
3ab51846 | 277 | @smallexample |
c1f7febf | 278 | #define SEARCH(array, target) \ |
310668e8 | 279 | (@{ \ |
c1f7febf RK |
280 | __label__ found; \ |
281 | typeof (target) _SEARCH_target = (target); \ | |
282 | typeof (*(array)) *_SEARCH_array = (array); \ | |
283 | int i, j; \ | |
284 | int value; \ | |
285 | for (i = 0; i < max; i++) \ | |
286 | for (j = 0; j < max; j++) \ | |
287 | if (_SEARCH_array[i][j] == _SEARCH_target) \ | |
310668e8 | 288 | @{ value = i; goto found; @} \ |
c1f7febf RK |
289 | value = -1; \ |
290 | found: \ | |
291 | value; \ | |
292 | @}) | |
3ab51846 | 293 | @end smallexample |
c1f7febf | 294 | |
14e33ee8 ZW |
295 | Local label declarations also make the labels they declare visible to |
296 | nested functions, if there are any. @xref{Nested Functions}, for details. | |
297 | ||
c1f7febf RK |
298 | @node Labels as Values |
299 | @section Labels as Values | |
300 | @cindex labels as values | |
301 | @cindex computed gotos | |
302 | @cindex goto with computed label | |
303 | @cindex address of a label | |
304 | ||
305 | You can get the address of a label defined in the current function | |
306 | (or a containing function) with the unary operator @samp{&&}. The | |
307 | value has type @code{void *}. This value is a constant and can be used | |
308 | wherever a constant of that type is valid. For example: | |
309 | ||
3ab51846 | 310 | @smallexample |
c1f7febf | 311 | void *ptr; |
0d893a63 | 312 | /* @r{@dots{}} */ |
c1f7febf | 313 | ptr = &&foo; |
3ab51846 | 314 | @end smallexample |
c1f7febf RK |
315 | |
316 | To use these values, you need to be able to jump to one. This is done | |
317 | with the computed goto statement@footnote{The analogous feature in | |
318 | Fortran is called an assigned goto, but that name seems inappropriate in | |
319 | C, where one can do more than simply store label addresses in label | |
320 | variables.}, @code{goto *@var{exp};}. For example, | |
321 | ||
3ab51846 | 322 | @smallexample |
c1f7febf | 323 | goto *ptr; |
3ab51846 | 324 | @end smallexample |
c1f7febf RK |
325 | |
326 | @noindent | |
327 | Any expression of type @code{void *} is allowed. | |
328 | ||
329 | One way of using these constants is in initializing a static array that | |
330 | will serve as a jump table: | |
331 | ||
3ab51846 | 332 | @smallexample |
c1f7febf | 333 | static void *array[] = @{ &&foo, &&bar, &&hack @}; |
3ab51846 | 334 | @end smallexample |
c1f7febf RK |
335 | |
336 | Then you can select a label with indexing, like this: | |
337 | ||
3ab51846 | 338 | @smallexample |
c1f7febf | 339 | goto *array[i]; |
3ab51846 | 340 | @end smallexample |
c1f7febf RK |
341 | |
342 | @noindent | |
343 | Note that this does not check whether the subscript is in bounds---array | |
344 | indexing in C never does that. | |
345 | ||
346 | Such an array of label values serves a purpose much like that of the | |
347 | @code{switch} statement. The @code{switch} statement is cleaner, so | |
348 | use that rather than an array unless the problem does not fit a | |
349 | @code{switch} statement very well. | |
350 | ||
351 | Another use of label values is in an interpreter for threaded code. | |
352 | The labels within the interpreter function can be stored in the | |
353 | threaded code for super-fast dispatching. | |
354 | ||
02f52e19 | 355 | You may not use this mechanism to jump to code in a different function. |
47620e09 | 356 | If you do that, totally unpredictable things will happen. The best way to |
c1f7febf RK |
357 | avoid this is to store the label address only in automatic variables and |
358 | never pass it as an argument. | |
359 | ||
47620e09 RH |
360 | An alternate way to write the above example is |
361 | ||
3ab51846 | 362 | @smallexample |
310668e8 JM |
363 | static const int array[] = @{ &&foo - &&foo, &&bar - &&foo, |
364 | &&hack - &&foo @}; | |
47620e09 | 365 | goto *(&&foo + array[i]); |
3ab51846 | 366 | @end smallexample |
47620e09 RH |
367 | |
368 | @noindent | |
369 | This is more friendly to code living in shared libraries, as it reduces | |
370 | the number of dynamic relocations that are needed, and by consequence, | |
371 | allows the data to be read-only. | |
372 | ||
2092ee7d JJ |
373 | The @code{&&foo} expressions for the same label might have different values |
374 | if the containing function is inlined or cloned. If a program relies on | |
375 | them being always the same, @code{__attribute__((__noinline__))} should | |
376 | be used to prevent inlining. If @code{&&foo} is used | |
377 | in a static variable initializer, inlining is forbidden. | |
378 | ||
c1f7febf RK |
379 | @node Nested Functions |
380 | @section Nested Functions | |
381 | @cindex nested functions | |
382 | @cindex downward funargs | |
383 | @cindex thunks | |
384 | ||
385 | A @dfn{nested function} is a function defined inside another function. | |
386 | (Nested functions are not supported for GNU C++.) The nested function's | |
387 | name is local to the block where it is defined. For example, here we | |
388 | define a nested function named @code{square}, and call it twice: | |
389 | ||
3ab51846 | 390 | @smallexample |
c1f7febf RK |
391 | @group |
392 | foo (double a, double b) | |
393 | @{ | |
394 | double square (double z) @{ return z * z; @} | |
395 | ||
396 | return square (a) + square (b); | |
397 | @} | |
398 | @end group | |
3ab51846 | 399 | @end smallexample |
c1f7febf RK |
400 | |
401 | The nested function can access all the variables of the containing | |
402 | function that are visible at the point of its definition. This is | |
403 | called @dfn{lexical scoping}. For example, here we show a nested | |
404 | function which uses an inherited variable named @code{offset}: | |
405 | ||
3ab51846 | 406 | @smallexample |
aee96fe9 | 407 | @group |
c1f7febf RK |
408 | bar (int *array, int offset, int size) |
409 | @{ | |
410 | int access (int *array, int index) | |
411 | @{ return array[index + offset]; @} | |
412 | int i; | |
0d893a63 | 413 | /* @r{@dots{}} */ |
c1f7febf | 414 | for (i = 0; i < size; i++) |
0d893a63 | 415 | /* @r{@dots{}} */ access (array, i) /* @r{@dots{}} */ |
c1f7febf | 416 | @} |
aee96fe9 | 417 | @end group |
3ab51846 | 418 | @end smallexample |
c1f7febf RK |
419 | |
420 | Nested function definitions are permitted within functions in the places | |
475b6e22 JM |
421 | where variable definitions are allowed; that is, in any block, mixed |
422 | with the other declarations and statements in the block. | |
c1f7febf RK |
423 | |
424 | It is possible to call the nested function from outside the scope of its | |
425 | name by storing its address or passing the address to another function: | |
426 | ||
3ab51846 | 427 | @smallexample |
c1f7febf RK |
428 | hack (int *array, int size) |
429 | @{ | |
430 | void store (int index, int value) | |
431 | @{ array[index] = value; @} | |
432 | ||
433 | intermediate (store, size); | |
434 | @} | |
3ab51846 | 435 | @end smallexample |
c1f7febf RK |
436 | |
437 | Here, the function @code{intermediate} receives the address of | |
438 | @code{store} as an argument. If @code{intermediate} calls @code{store}, | |
439 | the arguments given to @code{store} are used to store into @code{array}. | |
440 | But this technique works only so long as the containing function | |
441 | (@code{hack}, in this example) does not exit. | |
442 | ||
443 | If you try to call the nested function through its address after the | |
444 | containing function has exited, all hell will break loose. If you try | |
445 | to call it after a containing scope level has exited, and if it refers | |
446 | to some of the variables that are no longer in scope, you may be lucky, | |
447 | but it's not wise to take the risk. If, however, the nested function | |
448 | does not refer to anything that has gone out of scope, you should be | |
449 | safe. | |
450 | ||
9c34dbbf ZW |
451 | GCC implements taking the address of a nested function using a technique |
452 | called @dfn{trampolines}. A paper describing them is available as | |
453 | ||
454 | @noindent | |
b73b1546 | 455 | @uref{http://people.debian.org/~aaronl/Usenix88-lexic.pdf}. |
c1f7febf RK |
456 | |
457 | A nested function can jump to a label inherited from a containing | |
458 | function, provided the label was explicitly declared in the containing | |
459 | function (@pxref{Local Labels}). Such a jump returns instantly to the | |
460 | containing function, exiting the nested function which did the | |
461 | @code{goto} and any intermediate functions as well. Here is an example: | |
462 | ||
3ab51846 | 463 | @smallexample |
c1f7febf RK |
464 | @group |
465 | bar (int *array, int offset, int size) | |
466 | @{ | |
467 | __label__ failure; | |
468 | int access (int *array, int index) | |
469 | @{ | |
470 | if (index > size) | |
471 | goto failure; | |
472 | return array[index + offset]; | |
473 | @} | |
474 | int i; | |
0d893a63 | 475 | /* @r{@dots{}} */ |
c1f7febf | 476 | for (i = 0; i < size; i++) |
0d893a63 MK |
477 | /* @r{@dots{}} */ access (array, i) /* @r{@dots{}} */ |
478 | /* @r{@dots{}} */ | |
c1f7febf RK |
479 | return 0; |
480 | ||
481 | /* @r{Control comes here from @code{access} | |
482 | if it detects an error.} */ | |
483 | failure: | |
484 | return -1; | |
485 | @} | |
486 | @end group | |
3ab51846 | 487 | @end smallexample |
c1f7febf | 488 | |
28697267 DJ |
489 | A nested function always has no linkage. Declaring one with |
490 | @code{extern} or @code{static} is erroneous. If you need to declare the nested function | |
c1f7febf RK |
491 | before its definition, use @code{auto} (which is otherwise meaningless |
492 | for function declarations). | |
493 | ||
3ab51846 | 494 | @smallexample |
c1f7febf RK |
495 | bar (int *array, int offset, int size) |
496 | @{ | |
497 | __label__ failure; | |
498 | auto int access (int *, int); | |
0d893a63 | 499 | /* @r{@dots{}} */ |
c1f7febf RK |
500 | int access (int *array, int index) |
501 | @{ | |
502 | if (index > size) | |
503 | goto failure; | |
504 | return array[index + offset]; | |
505 | @} | |
0d893a63 | 506 | /* @r{@dots{}} */ |
c1f7febf | 507 | @} |
3ab51846 | 508 | @end smallexample |
c1f7febf RK |
509 | |
510 | @node Constructing Calls | |
511 | @section Constructing Function Calls | |
512 | @cindex constructing calls | |
513 | @cindex forwarding calls | |
514 | ||
515 | Using the built-in functions described below, you can record | |
516 | the arguments a function received, and call another function | |
517 | with the same arguments, without knowing the number or types | |
518 | of the arguments. | |
519 | ||
520 | You can also record the return value of that function call, | |
521 | and later return that value, without knowing what data type | |
522 | the function tried to return (as long as your caller expects | |
523 | that data type). | |
524 | ||
6429bc7c EB |
525 | However, these built-in functions may interact badly with some |
526 | sophisticated features or other extensions of the language. It | |
527 | is, therefore, not recommended to use them outside very simple | |
528 | functions acting as mere forwarders for their arguments. | |
529 | ||
84330467 JM |
530 | @deftypefn {Built-in Function} {void *} __builtin_apply_args () |
531 | This built-in function returns a pointer to data | |
c1f7febf RK |
532 | describing how to perform a call with the same arguments as were passed |
533 | to the current function. | |
534 | ||
535 | The function saves the arg pointer register, structure value address, | |
536 | and all registers that might be used to pass arguments to a function | |
537 | into a block of memory allocated on the stack. Then it returns the | |
538 | address of that block. | |
84330467 | 539 | @end deftypefn |
c1f7febf | 540 | |
84330467 JM |
541 | @deftypefn {Built-in Function} {void *} __builtin_apply (void (*@var{function})(), void *@var{arguments}, size_t @var{size}) |
542 | This built-in function invokes @var{function} | |
543 | with a copy of the parameters described by @var{arguments} | |
544 | and @var{size}. | |
c1f7febf RK |
545 | |
546 | The value of @var{arguments} should be the value returned by | |
547 | @code{__builtin_apply_args}. The argument @var{size} specifies the size | |
548 | of the stack argument data, in bytes. | |
549 | ||
84330467 | 550 | This function returns a pointer to data describing |
c1f7febf RK |
551 | how to return whatever value was returned by @var{function}. The data |
552 | is saved in a block of memory allocated on the stack. | |
553 | ||
554 | It is not always simple to compute the proper value for @var{size}. The | |
555 | value is used by @code{__builtin_apply} to compute the amount of data | |
556 | that should be pushed on the stack and copied from the incoming argument | |
557 | area. | |
84330467 | 558 | @end deftypefn |
c1f7febf | 559 | |
84330467 | 560 | @deftypefn {Built-in Function} {void} __builtin_return (void *@var{result}) |
c1f7febf RK |
561 | This built-in function returns the value described by @var{result} from |
562 | the containing function. You should specify, for @var{result}, a value | |
563 | returned by @code{__builtin_apply}. | |
84330467 | 564 | @end deftypefn |
c1f7febf | 565 | |
6ef5231b JJ |
566 | @deftypefn {Built-in Function} __builtin_va_arg_pack () |
567 | This built-in function represents all anonymous arguments of an inline | |
568 | function. It can be used only in inline functions which will be always | |
569 | inlined, never compiled as a separate function, such as those using | |
570 | @code{__attribute__ ((__always_inline__))} or | |
571 | @code{__attribute__ ((__gnu_inline__))} extern inline functions. | |
572 | It must be only passed as last argument to some other function | |
573 | with variable arguments. This is useful for writing small wrapper | |
574 | inlines for variable argument functions, when using preprocessor | |
575 | macros is undesirable. For example: | |
576 | @smallexample | |
577 | extern int myprintf (FILE *f, const char *format, ...); | |
578 | extern inline __attribute__ ((__gnu_inline__)) int | |
579 | myprintf (FILE *f, const char *format, ...) | |
580 | @{ | |
581 | int r = fprintf (f, "myprintf: "); | |
582 | if (r < 0) | |
583 | return r; | |
584 | int s = fprintf (f, format, __builtin_va_arg_pack ()); | |
585 | if (s < 0) | |
586 | return s; | |
587 | return r + s; | |
588 | @} | |
589 | @end smallexample | |
590 | @end deftypefn | |
591 | ||
ab0e176c JJ |
592 | @deftypefn {Built-in Function} __builtin_va_arg_pack_len () |
593 | This built-in function returns the number of anonymous arguments of | |
594 | an inline function. It can be used only in inline functions which | |
595 | will be always inlined, never compiled as a separate function, such | |
596 | as those using @code{__attribute__ ((__always_inline__))} or | |
597 | @code{__attribute__ ((__gnu_inline__))} extern inline functions. | |
598 | For example following will do link or runtime checking of open | |
599 | arguments for optimized code: | |
600 | @smallexample | |
601 | #ifdef __OPTIMIZE__ | |
602 | extern inline __attribute__((__gnu_inline__)) int | |
603 | myopen (const char *path, int oflag, ...) | |
604 | @{ | |
605 | if (__builtin_va_arg_pack_len () > 1) | |
606 | warn_open_too_many_arguments (); | |
607 | ||
608 | if (__builtin_constant_p (oflag)) | |
609 | @{ | |
610 | if ((oflag & O_CREAT) != 0 && __builtin_va_arg_pack_len () < 1) | |
611 | @{ | |
612 | warn_open_missing_mode (); | |
613 | return __open_2 (path, oflag); | |
614 | @} | |
615 | return open (path, oflag, __builtin_va_arg_pack ()); | |
616 | @} | |
617 | ||
618 | if (__builtin_va_arg_pack_len () < 1) | |
619 | return __open_2 (path, oflag); | |
620 | ||
621 | return open (path, oflag, __builtin_va_arg_pack ()); | |
622 | @} | |
623 | #endif | |
624 | @end smallexample | |
625 | @end deftypefn | |
626 | ||
c1f7febf RK |
627 | @node Typeof |
628 | @section Referring to a Type with @code{typeof} | |
629 | @findex typeof | |
630 | @findex sizeof | |
631 | @cindex macros, types of arguments | |
632 | ||
633 | Another way to refer to the type of an expression is with @code{typeof}. | |
634 | The syntax of using of this keyword looks like @code{sizeof}, but the | |
635 | construct acts semantically like a type name defined with @code{typedef}. | |
636 | ||
637 | There are two ways of writing the argument to @code{typeof}: with an | |
638 | expression or with a type. Here is an example with an expression: | |
639 | ||
3ab51846 | 640 | @smallexample |
c1f7febf | 641 | typeof (x[0](1)) |
3ab51846 | 642 | @end smallexample |
c1f7febf RK |
643 | |
644 | @noindent | |
89aed483 JM |
645 | This assumes that @code{x} is an array of pointers to functions; |
646 | the type described is that of the values of the functions. | |
c1f7febf RK |
647 | |
648 | Here is an example with a typename as the argument: | |
649 | ||
3ab51846 | 650 | @smallexample |
c1f7febf | 651 | typeof (int *) |
3ab51846 | 652 | @end smallexample |
c1f7febf RK |
653 | |
654 | @noindent | |
655 | Here the type described is that of pointers to @code{int}. | |
656 | ||
5490d604 | 657 | If you are writing a header file that must work when included in ISO C |
c1f7febf RK |
658 | programs, write @code{__typeof__} instead of @code{typeof}. |
659 | @xref{Alternate Keywords}. | |
660 | ||
661 | A @code{typeof}-construct can be used anywhere a typedef name could be | |
662 | used. For example, you can use it in a declaration, in a cast, or inside | |
663 | of @code{sizeof} or @code{typeof}. | |
664 | ||
95f79357 ZW |
665 | @code{typeof} is often useful in conjunction with the |
666 | statements-within-expressions feature. Here is how the two together can | |
667 | be used to define a safe ``maximum'' macro that operates on any | |
668 | arithmetic type and evaluates each of its arguments exactly once: | |
669 | ||
3ab51846 | 670 | @smallexample |
95f79357 ZW |
671 | #define max(a,b) \ |
672 | (@{ typeof (a) _a = (a); \ | |
673 | typeof (b) _b = (b); \ | |
674 | _a > _b ? _a : _b; @}) | |
3ab51846 | 675 | @end smallexample |
95f79357 | 676 | |
526278c9 VR |
677 | @cindex underscores in variables in macros |
678 | @cindex @samp{_} in variables in macros | |
679 | @cindex local variables in macros | |
680 | @cindex variables, local, in macros | |
681 | @cindex macros, local variables in | |
682 | ||
683 | The reason for using names that start with underscores for the local | |
684 | variables is to avoid conflicts with variable names that occur within the | |
685 | expressions that are substituted for @code{a} and @code{b}. Eventually we | |
686 | hope to design a new form of declaration syntax that allows you to declare | |
687 | variables whose scopes start only after their initializers; this will be a | |
688 | more reliable way to prevent such conflicts. | |
689 | ||
95f79357 ZW |
690 | @noindent |
691 | Some more examples of the use of @code{typeof}: | |
692 | ||
c1f7febf RK |
693 | @itemize @bullet |
694 | @item | |
695 | This declares @code{y} with the type of what @code{x} points to. | |
696 | ||
3ab51846 | 697 | @smallexample |
c1f7febf | 698 | typeof (*x) y; |
3ab51846 | 699 | @end smallexample |
c1f7febf RK |
700 | |
701 | @item | |
702 | This declares @code{y} as an array of such values. | |
703 | ||
3ab51846 | 704 | @smallexample |
c1f7febf | 705 | typeof (*x) y[4]; |
3ab51846 | 706 | @end smallexample |
c1f7febf RK |
707 | |
708 | @item | |
709 | This declares @code{y} as an array of pointers to characters: | |
710 | ||
3ab51846 | 711 | @smallexample |
c1f7febf | 712 | typeof (typeof (char *)[4]) y; |
3ab51846 | 713 | @end smallexample |
c1f7febf RK |
714 | |
715 | @noindent | |
716 | It is equivalent to the following traditional C declaration: | |
717 | ||
3ab51846 | 718 | @smallexample |
c1f7febf | 719 | char *y[4]; |
3ab51846 | 720 | @end smallexample |
c1f7febf RK |
721 | |
722 | To see the meaning of the declaration using @code{typeof}, and why it | |
962e6e00 | 723 | might be a useful way to write, rewrite it with these macros: |
c1f7febf | 724 | |
3ab51846 | 725 | @smallexample |
c1f7febf RK |
726 | #define pointer(T) typeof(T *) |
727 | #define array(T, N) typeof(T [N]) | |
3ab51846 | 728 | @end smallexample |
c1f7febf RK |
729 | |
730 | @noindent | |
731 | Now the declaration can be rewritten this way: | |
732 | ||
3ab51846 | 733 | @smallexample |
c1f7febf | 734 | array (pointer (char), 4) y; |
3ab51846 | 735 | @end smallexample |
c1f7febf RK |
736 | |
737 | @noindent | |
738 | Thus, @code{array (pointer (char), 4)} is the type of arrays of 4 | |
739 | pointers to @code{char}. | |
740 | @end itemize | |
741 | ||
95f79357 ZW |
742 | @emph{Compatibility Note:} In addition to @code{typeof}, GCC 2 supported |
743 | a more limited extension which permitted one to write | |
744 | ||
3ab51846 | 745 | @smallexample |
95f79357 | 746 | typedef @var{T} = @var{expr}; |
3ab51846 | 747 | @end smallexample |
95f79357 ZW |
748 | |
749 | @noindent | |
750 | with the effect of declaring @var{T} to have the type of the expression | |
751 | @var{expr}. This extension does not work with GCC 3 (versions between | |
752 | 3.0 and 3.2 will crash; 3.2.1 and later give an error). Code which | |
753 | relies on it should be rewritten to use @code{typeof}: | |
754 | ||
3ab51846 | 755 | @smallexample |
95f79357 | 756 | typedef typeof(@var{expr}) @var{T}; |
3ab51846 | 757 | @end smallexample |
95f79357 ZW |
758 | |
759 | @noindent | |
760 | This will work with all versions of GCC@. | |
761 | ||
c1f7febf RK |
762 | @node Conditionals |
763 | @section Conditionals with Omitted Operands | |
764 | @cindex conditional expressions, extensions | |
765 | @cindex omitted middle-operands | |
766 | @cindex middle-operands, omitted | |
767 | @cindex extensions, @code{?:} | |
768 | @cindex @code{?:} extensions | |
769 | ||
770 | The middle operand in a conditional expression may be omitted. Then | |
771 | if the first operand is nonzero, its value is the value of the conditional | |
772 | expression. | |
773 | ||
774 | Therefore, the expression | |
775 | ||
3ab51846 | 776 | @smallexample |
c1f7febf | 777 | x ? : y |
3ab51846 | 778 | @end smallexample |
c1f7febf RK |
779 | |
780 | @noindent | |
781 | has the value of @code{x} if that is nonzero; otherwise, the value of | |
782 | @code{y}. | |
783 | ||
784 | This example is perfectly equivalent to | |
785 | ||
3ab51846 | 786 | @smallexample |
c1f7febf | 787 | x ? x : y |
3ab51846 | 788 | @end smallexample |
c1f7febf RK |
789 | |
790 | @cindex side effect in ?: | |
791 | @cindex ?: side effect | |
792 | @noindent | |
793 | In this simple case, the ability to omit the middle operand is not | |
794 | especially useful. When it becomes useful is when the first operand does, | |
795 | or may (if it is a macro argument), contain a side effect. Then repeating | |
796 | the operand in the middle would perform the side effect twice. Omitting | |
797 | the middle operand uses the value already computed without the undesirable | |
798 | effects of recomputing it. | |
799 | ||
800 | @node Long Long | |
801 | @section Double-Word Integers | |
802 | @cindex @code{long long} data types | |
803 | @cindex double-word arithmetic | |
804 | @cindex multiprecision arithmetic | |
4b404517 JM |
805 | @cindex @code{LL} integer suffix |
806 | @cindex @code{ULL} integer suffix | |
c1f7febf | 807 | |
4b404517 JM |
808 | ISO C99 supports data types for integers that are at least 64 bits wide, |
809 | and as an extension GCC supports them in C89 mode and in C++. | |
810 | Simply write @code{long long int} for a signed integer, or | |
c1f7febf | 811 | @code{unsigned long long int} for an unsigned integer. To make an |
84330467 | 812 | integer constant of type @code{long long int}, add the suffix @samp{LL} |
c1f7febf | 813 | to the integer. To make an integer constant of type @code{unsigned long |
84330467 | 814 | long int}, add the suffix @samp{ULL} to the integer. |
c1f7febf RK |
815 | |
816 | You can use these types in arithmetic like any other integer types. | |
817 | Addition, subtraction, and bitwise boolean operations on these types | |
818 | are open-coded on all types of machines. Multiplication is open-coded | |
819 | if the machine supports fullword-to-doubleword a widening multiply | |
820 | instruction. Division and shifts are open-coded only on machines that | |
821 | provide special support. The operations that are not open-coded use | |
161d7b59 | 822 | special library routines that come with GCC@. |
c1f7febf RK |
823 | |
824 | There may be pitfalls when you use @code{long long} types for function | |
825 | arguments, unless you declare function prototypes. If a function | |
826 | expects type @code{int} for its argument, and you pass a value of type | |
827 | @code{long long int}, confusion will result because the caller and the | |
828 | subroutine will disagree about the number of bytes for the argument. | |
829 | Likewise, if the function expects @code{long long int} and you pass | |
830 | @code{int}. The best way to avoid such problems is to use prototypes. | |
831 | ||
832 | @node Complex | |
833 | @section Complex Numbers | |
834 | @cindex complex numbers | |
4b404517 JM |
835 | @cindex @code{_Complex} keyword |
836 | @cindex @code{__complex__} keyword | |
c1f7febf | 837 | |
4b404517 JM |
838 | ISO C99 supports complex floating data types, and as an extension GCC |
839 | supports them in C89 mode and in C++, and supports complex integer data | |
840 | types which are not part of ISO C99. You can declare complex types | |
841 | using the keyword @code{_Complex}. As an extension, the older GNU | |
842 | keyword @code{__complex__} is also supported. | |
c1f7febf | 843 | |
4b404517 | 844 | For example, @samp{_Complex double x;} declares @code{x} as a |
c1f7febf | 845 | variable whose real part and imaginary part are both of type |
4b404517 | 846 | @code{double}. @samp{_Complex short int y;} declares @code{y} to |
c1f7febf RK |
847 | have real and imaginary parts of type @code{short int}; this is not |
848 | likely to be useful, but it shows that the set of complex types is | |
849 | complete. | |
850 | ||
851 | To write a constant with a complex data type, use the suffix @samp{i} or | |
852 | @samp{j} (either one; they are equivalent). For example, @code{2.5fi} | |
4b404517 JM |
853 | has type @code{_Complex float} and @code{3i} has type |
854 | @code{_Complex int}. Such a constant always has a pure imaginary | |
c1f7febf | 855 | value, but you can form any complex value you like by adding one to a |
4b404517 JM |
856 | real constant. This is a GNU extension; if you have an ISO C99 |
857 | conforming C library (such as GNU libc), and want to construct complex | |
858 | constants of floating type, you should include @code{<complex.h>} and | |
859 | use the macros @code{I} or @code{_Complex_I} instead. | |
c1f7febf | 860 | |
4b404517 JM |
861 | @cindex @code{__real__} keyword |
862 | @cindex @code{__imag__} keyword | |
c1f7febf RK |
863 | To extract the real part of a complex-valued expression @var{exp}, write |
864 | @code{__real__ @var{exp}}. Likewise, use @code{__imag__} to | |
4b404517 JM |
865 | extract the imaginary part. This is a GNU extension; for values of |
866 | floating type, you should use the ISO C99 functions @code{crealf}, | |
867 | @code{creal}, @code{creall}, @code{cimagf}, @code{cimag} and | |
868 | @code{cimagl}, declared in @code{<complex.h>} and also provided as | |
161d7b59 | 869 | built-in functions by GCC@. |
c1f7febf | 870 | |
4b404517 | 871 | @cindex complex conjugation |
c1f7febf | 872 | The operator @samp{~} performs complex conjugation when used on a value |
4b404517 JM |
873 | with a complex type. This is a GNU extension; for values of |
874 | floating type, you should use the ISO C99 functions @code{conjf}, | |
875 | @code{conj} and @code{conjl}, declared in @code{<complex.h>} and also | |
161d7b59 | 876 | provided as built-in functions by GCC@. |
c1f7febf | 877 | |
f0523f02 | 878 | GCC can allocate complex automatic variables in a noncontiguous |
c1f7febf | 879 | fashion; it's even possible for the real part to be in a register while |
580fb356 JW |
880 | the imaginary part is on the stack (or vice-versa). Only the DWARF2 |
881 | debug info format can represent this, so use of DWARF2 is recommended. | |
882 | If you are using the stabs debug info format, GCC describes a noncontiguous | |
883 | complex variable as if it were two separate variables of noncomplex type. | |
c1f7febf RK |
884 | If the variable's actual name is @code{foo}, the two fictitious |
885 | variables are named @code{foo$real} and @code{foo$imag}. You can | |
886 | examine and set these two fictitious variables with your debugger. | |
887 | ||
c77cd3d1 UB |
888 | @node Floating Types |
889 | @section Additional Floating Types | |
890 | @cindex additional floating types | |
891 | @cindex @code{__float80} data type | |
892 | @cindex @code{__float128} data type | |
893 | @cindex @code{w} floating point suffix | |
894 | @cindex @code{q} floating point suffix | |
895 | @cindex @code{W} floating point suffix | |
896 | @cindex @code{Q} floating point suffix | |
897 | ||
898 | As an extension, the GNU C compiler supports additional floating | |
899 | types, @code{__float80} and @code{__float128} to support 80bit | |
900 | (@code{XFmode}) and 128 bit (@code{TFmode}) floating types. | |
901 | Support for additional types includes the arithmetic operators: | |
902 | add, subtract, multiply, divide; unary arithmetic operators; | |
903 | relational operators; equality operators; and conversions to and from | |
904 | integer and other floating types. Use a suffix @samp{w} or @samp{W} | |
905 | in a literal constant of type @code{__float80} and @samp{q} or @samp{Q} | |
906 | for @code{_float128}. You can declare complex types using the | |
907 | corresponding internal complex type, @code{XCmode} for @code{__float80} | |
908 | type and @code{TCmode} for @code{__float128} type: | |
909 | ||
910 | @smallexample | |
911 | typedef _Complex float __attribute__((mode(TC))) _Complex128; | |
912 | typedef _Complex float __attribute__((mode(XC))) _Complex80; | |
913 | @end smallexample | |
914 | ||
915 | Not all targets support additional floating point types. @code{__float80} | |
916 | is supported on i386, x86_64 and ia64 targets and target @code{__float128} | |
917 | is supported on x86_64 and ia64 targets. | |
918 | ||
9a8ce21f | 919 | @node Decimal Float |
85a92f7e JJ |
920 | @section Decimal Floating Types |
921 | @cindex decimal floating types | |
9a8ce21f JG |
922 | @cindex @code{_Decimal32} data type |
923 | @cindex @code{_Decimal64} data type | |
924 | @cindex @code{_Decimal128} data type | |
925 | @cindex @code{df} integer suffix | |
926 | @cindex @code{dd} integer suffix | |
927 | @cindex @code{dl} integer suffix | |
928 | @cindex @code{DF} integer suffix | |
929 | @cindex @code{DD} integer suffix | |
930 | @cindex @code{DL} integer suffix | |
931 | ||
85a92f7e JJ |
932 | As an extension, the GNU C compiler supports decimal floating types as |
933 | defined in the N1176 draft of ISO/IEC WDTR24732. Support for decimal | |
934 | floating types in GCC will evolve as the draft technical report changes. | |
935 | Calling conventions for any target might also change. Not all targets | |
936 | support decimal floating types. | |
9a8ce21f | 937 | |
85a92f7e JJ |
938 | The decimal floating types are @code{_Decimal32}, @code{_Decimal64}, and |
939 | @code{_Decimal128}. They use a radix of ten, unlike the floating types | |
940 | @code{float}, @code{double}, and @code{long double} whose radix is not | |
941 | specified by the C standard but is usually two. | |
942 | ||
943 | Support for decimal floating types includes the arithmetic operators | |
9a8ce21f JG |
944 | add, subtract, multiply, divide; unary arithmetic operators; |
945 | relational operators; equality operators; and conversions to and from | |
85a92f7e | 946 | integer and other floating types. Use a suffix @samp{df} or |
9a8ce21f JG |
947 | @samp{DF} in a literal constant of type @code{_Decimal32}, @samp{dd} |
948 | or @samp{DD} for @code{_Decimal64}, and @samp{dl} or @samp{DL} for | |
949 | @code{_Decimal128}. | |
950 | ||
85a92f7e JJ |
951 | GCC support of decimal float as specified by the draft technical report |
952 | is incomplete: | |
953 | ||
954 | @itemize @bullet | |
955 | @item | |
956 | Translation time data type (TTDT) is not supported. | |
957 | ||
85a92f7e JJ |
958 | @item |
959 | When the value of a decimal floating type cannot be represented in the | |
960 | integer type to which it is being converted, the result is undefined | |
961 | rather than the result value specified by the draft technical report. | |
962 | @end itemize | |
9a8ce21f JG |
963 | |
964 | Types @code{_Decimal32}, @code{_Decimal64}, and @code{_Decimal128} | |
965 | are supported by the DWARF2 debug information format. | |
966 | ||
6f4d7222 | 967 | @node Hex Floats |
6b42b9ea UD |
968 | @section Hex Floats |
969 | @cindex hex floats | |
c5c76735 | 970 | |
4b404517 | 971 | ISO C99 supports floating-point numbers written not only in the usual |
6f4d7222 | 972 | decimal notation, such as @code{1.55e1}, but also numbers such as |
4b404517 JM |
973 | @code{0x1.fp3} written in hexadecimal format. As a GNU extension, GCC |
974 | supports this in C89 mode (except in some cases when strictly | |
975 | conforming) and in C++. In that format the | |
84330467 | 976 | @samp{0x} hex introducer and the @samp{p} or @samp{P} exponent field are |
6f4d7222 | 977 | mandatory. The exponent is a decimal number that indicates the power of |
84330467 | 978 | 2 by which the significant part will be multiplied. Thus @samp{0x1.f} is |
aee96fe9 JM |
979 | @tex |
980 | $1 {15\over16}$, | |
981 | @end tex | |
982 | @ifnottex | |
983 | 1 15/16, | |
984 | @end ifnottex | |
985 | @samp{p3} multiplies it by 8, and the value of @code{0x1.fp3} | |
6f4d7222 UD |
986 | is the same as @code{1.55e1}. |
987 | ||
988 | Unlike for floating-point numbers in the decimal notation the exponent | |
989 | is always required in the hexadecimal notation. Otherwise the compiler | |
990 | would not be able to resolve the ambiguity of, e.g., @code{0x1.f}. This | |
84330467 | 991 | could mean @code{1.0f} or @code{1.9375} since @samp{f} is also the |
6f4d7222 UD |
992 | extension for floating-point constants of type @code{float}. |
993 | ||
0f996086 CF |
994 | @node Fixed-Point |
995 | @section Fixed-Point Types | |
996 | @cindex fixed-point types | |
997 | @cindex @code{_Fract} data type | |
998 | @cindex @code{_Accum} data type | |
999 | @cindex @code{_Sat} data type | |
1000 | @cindex @code{hr} fixed-suffix | |
1001 | @cindex @code{r} fixed-suffix | |
1002 | @cindex @code{lr} fixed-suffix | |
1003 | @cindex @code{llr} fixed-suffix | |
1004 | @cindex @code{uhr} fixed-suffix | |
1005 | @cindex @code{ur} fixed-suffix | |
1006 | @cindex @code{ulr} fixed-suffix | |
1007 | @cindex @code{ullr} fixed-suffix | |
1008 | @cindex @code{hk} fixed-suffix | |
1009 | @cindex @code{k} fixed-suffix | |
1010 | @cindex @code{lk} fixed-suffix | |
1011 | @cindex @code{llk} fixed-suffix | |
1012 | @cindex @code{uhk} fixed-suffix | |
1013 | @cindex @code{uk} fixed-suffix | |
1014 | @cindex @code{ulk} fixed-suffix | |
1015 | @cindex @code{ullk} fixed-suffix | |
1016 | @cindex @code{HR} fixed-suffix | |
1017 | @cindex @code{R} fixed-suffix | |
1018 | @cindex @code{LR} fixed-suffix | |
1019 | @cindex @code{LLR} fixed-suffix | |
1020 | @cindex @code{UHR} fixed-suffix | |
1021 | @cindex @code{UR} fixed-suffix | |
1022 | @cindex @code{ULR} fixed-suffix | |
1023 | @cindex @code{ULLR} fixed-suffix | |
1024 | @cindex @code{HK} fixed-suffix | |
1025 | @cindex @code{K} fixed-suffix | |
1026 | @cindex @code{LK} fixed-suffix | |
1027 | @cindex @code{LLK} fixed-suffix | |
1028 | @cindex @code{UHK} fixed-suffix | |
1029 | @cindex @code{UK} fixed-suffix | |
1030 | @cindex @code{ULK} fixed-suffix | |
1031 | @cindex @code{ULLK} fixed-suffix | |
1032 | ||
1033 | As an extension, the GNU C compiler supports fixed-point types as | |
1034 | defined in the N1169 draft of ISO/IEC DTR 18037. Support for fixed-point | |
1035 | types in GCC will evolve as the draft technical report changes. | |
1036 | Calling conventions for any target might also change. Not all targets | |
1037 | support fixed-point types. | |
1038 | ||
1039 | The fixed-point types are | |
1040 | @code{short _Fract}, | |
1041 | @code{_Fract}, | |
1042 | @code{long _Fract}, | |
1043 | @code{long long _Fract}, | |
1044 | @code{unsigned short _Fract}, | |
1045 | @code{unsigned _Fract}, | |
1046 | @code{unsigned long _Fract}, | |
1047 | @code{unsigned long long _Fract}, | |
1048 | @code{_Sat short _Fract}, | |
1049 | @code{_Sat _Fract}, | |
1050 | @code{_Sat long _Fract}, | |
1051 | @code{_Sat long long _Fract}, | |
1052 | @code{_Sat unsigned short _Fract}, | |
1053 | @code{_Sat unsigned _Fract}, | |
1054 | @code{_Sat unsigned long _Fract}, | |
1055 | @code{_Sat unsigned long long _Fract}, | |
1056 | @code{short _Accum}, | |
1057 | @code{_Accum}, | |
1058 | @code{long _Accum}, | |
1059 | @code{long long _Accum}, | |
1060 | @code{unsigned short _Accum}, | |
1061 | @code{unsigned _Accum}, | |
1062 | @code{unsigned long _Accum}, | |
1063 | @code{unsigned long long _Accum}, | |
1064 | @code{_Sat short _Accum}, | |
1065 | @code{_Sat _Accum}, | |
1066 | @code{_Sat long _Accum}, | |
1067 | @code{_Sat long long _Accum}, | |
1068 | @code{_Sat unsigned short _Accum}, | |
1069 | @code{_Sat unsigned _Accum}, | |
1070 | @code{_Sat unsigned long _Accum}, | |
1071 | @code{_Sat unsigned long long _Accum}. | |
1072 | Fixed-point data values contain fractional and optional integral parts. | |
1073 | The format of fixed-point data varies and depends on the target machine. | |
1074 | ||
1075 | Support for fixed-point types includes prefix and postfix increment | |
1076 | and decrement operators (@code{++}, @code{--}); unary arithmetic operators | |
1077 | (@code{+}, @code{-}, @code{!}); binary arithmetic operators (@code{+}, | |
1078 | @code{-}, @code{*}, @code{/}); binary shift operators (@code{<<}, @code{>>}); | |
1079 | relational operators (@code{<}, @code{<=}, @code{>=}, @code{>}); | |
1080 | equality operators (@code{==}, @code{!=}); assignment operators | |
1081 | (@code{+=}, @code{-=}, @code{*=}, @code{/=}, @code{<<=}, @code{>>=}); | |
1082 | and conversions to and from integer, floating-point, or fixed-point types. | |
1083 | ||
1084 | Use a suffix @samp{hr} or @samp{HR} in a literal constant of type | |
1085 | @code{short _Fract} and @code{_Sat short _Fract}, | |
1086 | @samp{r} or @samp{R} for @code{_Fract} and @code{_Sat _Fract}, | |
1087 | @samp{lr} or @samp{LR} for @code{long _Fract} and @code{_Sat long _Fract}, | |
1088 | @samp{llr} or @samp{LLR} for @code{long long _Fract} and | |
1089 | @code{_Sat long long _Fract}, | |
1090 | @samp{uhr} or @samp{UHR} for @code{unsigned short _Fract} and | |
1091 | @code{_Sat unsigned short _Fract}, | |
1092 | @samp{ur} or @samp{UR} for @code{unsigned _Fract} and | |
1093 | @code{_Sat unsigned _Fract}, | |
1094 | @samp{ulr} or @samp{ULR} for @code{unsigned long _Fract} and | |
1095 | @code{_Sat unsigned long _Fract}, | |
1096 | @samp{ullr} or @samp{ULLR} for @code{unsigned long long _Fract} | |
1097 | and @code{_Sat unsigned long long _Fract}, | |
1098 | @samp{hk} or @samp{HK} for @code{short _Accum} and @code{_Sat short _Accum}, | |
1099 | @samp{k} or @samp{K} for @code{_Accum} and @code{_Sat _Accum}, | |
1100 | @samp{lk} or @samp{LK} for @code{long _Accum} and @code{_Sat long _Accum}, | |
1101 | @samp{llk} or @samp{LLK} for @code{long long _Accum} and | |
1102 | @code{_Sat long long _Accum}, | |
1103 | @samp{uhk} or @samp{UHK} for @code{unsigned short _Accum} and | |
1104 | @code{_Sat unsigned short _Accum}, | |
1105 | @samp{uk} or @samp{UK} for @code{unsigned _Accum} and | |
1106 | @code{_Sat unsigned _Accum}, | |
1107 | @samp{ulk} or @samp{ULK} for @code{unsigned long _Accum} and | |
1108 | @code{_Sat unsigned long _Accum}, | |
1109 | and @samp{ullk} or @samp{ULLK} for @code{unsigned long long _Accum} | |
1110 | and @code{_Sat unsigned long long _Accum}. | |
1111 | ||
1112 | GCC support of fixed-point types as specified by the draft technical report | |
1113 | is incomplete: | |
1114 | ||
1115 | @itemize @bullet | |
1116 | @item | |
1117 | Pragmas to control overflow and rounding behaviors are not implemented. | |
1118 | @end itemize | |
1119 | ||
1120 | Fixed-point types are supported by the DWARF2 debug information format. | |
1121 | ||
c1f7febf RK |
1122 | @node Zero Length |
1123 | @section Arrays of Length Zero | |
1124 | @cindex arrays of length zero | |
1125 | @cindex zero-length arrays | |
1126 | @cindex length-zero arrays | |
ffc5c6a9 | 1127 | @cindex flexible array members |
c1f7febf | 1128 | |
161d7b59 | 1129 | Zero-length arrays are allowed in GNU C@. They are very useful as the |
584ef5fe | 1130 | last element of a structure which is really a header for a variable-length |
c1f7febf RK |
1131 | object: |
1132 | ||
3ab51846 | 1133 | @smallexample |
c1f7febf RK |
1134 | struct line @{ |
1135 | int length; | |
1136 | char contents[0]; | |
1137 | @}; | |
1138 | ||
584ef5fe RH |
1139 | struct line *thisline = (struct line *) |
1140 | malloc (sizeof (struct line) + this_length); | |
1141 | thisline->length = this_length; | |
3ab51846 | 1142 | @end smallexample |
c1f7febf | 1143 | |
3764f879 | 1144 | In ISO C90, you would have to give @code{contents} a length of 1, which |
c1f7febf RK |
1145 | means either you waste space or complicate the argument to @code{malloc}. |
1146 | ||
02f52e19 | 1147 | In ISO C99, you would use a @dfn{flexible array member}, which is |
584ef5fe RH |
1148 | slightly different in syntax and semantics: |
1149 | ||
1150 | @itemize @bullet | |
1151 | @item | |
1152 | Flexible array members are written as @code{contents[]} without | |
1153 | the @code{0}. | |
1154 | ||
1155 | @item | |
1156 | Flexible array members have incomplete type, and so the @code{sizeof} | |
1157 | operator may not be applied. As a quirk of the original implementation | |
1158 | of zero-length arrays, @code{sizeof} evaluates to zero. | |
1159 | ||
1160 | @item | |
1161 | Flexible array members may only appear as the last member of a | |
e7b6a0ee | 1162 | @code{struct} that is otherwise non-empty. |
2984fe64 JM |
1163 | |
1164 | @item | |
1165 | A structure containing a flexible array member, or a union containing | |
1166 | such a structure (possibly recursively), may not be a member of a | |
1167 | structure or an element of an array. (However, these uses are | |
1168 | permitted by GCC as extensions.) | |
ffc5c6a9 | 1169 | @end itemize |
a25f1211 | 1170 | |
ffc5c6a9 | 1171 | GCC versions before 3.0 allowed zero-length arrays to be statically |
e7b6a0ee DD |
1172 | initialized, as if they were flexible arrays. In addition to those |
1173 | cases that were useful, it also allowed initializations in situations | |
1174 | that would corrupt later data. Non-empty initialization of zero-length | |
1175 | arrays is now treated like any case where there are more initializer | |
1176 | elements than the array holds, in that a suitable warning about "excess | |
1177 | elements in array" is given, and the excess elements (all of them, in | |
1178 | this case) are ignored. | |
ffc5c6a9 RH |
1179 | |
1180 | Instead GCC allows static initialization of flexible array members. | |
1181 | This is equivalent to defining a new structure containing the original | |
1182 | structure followed by an array of sufficient size to contain the data. | |
e979f9e8 | 1183 | I.e.@: in the following, @code{f1} is constructed as if it were declared |
ffc5c6a9 | 1184 | like @code{f2}. |
a25f1211 | 1185 | |
3ab51846 | 1186 | @smallexample |
ffc5c6a9 RH |
1187 | struct f1 @{ |
1188 | int x; int y[]; | |
1189 | @} f1 = @{ 1, @{ 2, 3, 4 @} @}; | |
1190 | ||
1191 | struct f2 @{ | |
1192 | struct f1 f1; int data[3]; | |
1193 | @} f2 = @{ @{ 1 @}, @{ 2, 3, 4 @} @}; | |
3ab51846 | 1194 | @end smallexample |
584ef5fe | 1195 | |
ffc5c6a9 RH |
1196 | @noindent |
1197 | The convenience of this extension is that @code{f1} has the desired | |
1198 | type, eliminating the need to consistently refer to @code{f2.f1}. | |
1199 | ||
1200 | This has symmetry with normal static arrays, in that an array of | |
1201 | unknown size is also written with @code{[]}. | |
a25f1211 | 1202 | |
ffc5c6a9 RH |
1203 | Of course, this extension only makes sense if the extra data comes at |
1204 | the end of a top-level object, as otherwise we would be overwriting | |
1205 | data at subsequent offsets. To avoid undue complication and confusion | |
1206 | with initialization of deeply nested arrays, we simply disallow any | |
1207 | non-empty initialization except when the structure is the top-level | |
1208 | object. For example: | |
584ef5fe | 1209 | |
3ab51846 | 1210 | @smallexample |
ffc5c6a9 RH |
1211 | struct foo @{ int x; int y[]; @}; |
1212 | struct bar @{ struct foo z; @}; | |
1213 | ||
13ba36b4 JM |
1214 | struct foo a = @{ 1, @{ 2, 3, 4 @} @}; // @r{Valid.} |
1215 | struct bar b = @{ @{ 1, @{ 2, 3, 4 @} @} @}; // @r{Invalid.} | |
1216 | struct bar c = @{ @{ 1, @{ @} @} @}; // @r{Valid.} | |
1217 | struct foo d[1] = @{ @{ 1 @{ 2, 3, 4 @} @} @}; // @r{Invalid.} | |
3ab51846 | 1218 | @end smallexample |
4b606faf | 1219 | |
ba05abd3 GK |
1220 | @node Empty Structures |
1221 | @section Structures With No Members | |
1222 | @cindex empty structures | |
1223 | @cindex zero-size structures | |
1224 | ||
1225 | GCC permits a C structure to have no members: | |
1226 | ||
3ab51846 | 1227 | @smallexample |
ba05abd3 GK |
1228 | struct empty @{ |
1229 | @}; | |
3ab51846 | 1230 | @end smallexample |
ba05abd3 GK |
1231 | |
1232 | The structure will have size zero. In C++, empty structures are part | |
db0b376e MM |
1233 | of the language. G++ treats empty structures as if they had a single |
1234 | member of type @code{char}. | |
ba05abd3 | 1235 | |
c1f7febf RK |
1236 | @node Variable Length |
1237 | @section Arrays of Variable Length | |
1238 | @cindex variable-length arrays | |
1239 | @cindex arrays of variable length | |
4b404517 | 1240 | @cindex VLAs |
c1f7febf | 1241 | |
4b404517 JM |
1242 | Variable-length automatic arrays are allowed in ISO C99, and as an |
1243 | extension GCC accepts them in C89 mode and in C++. (However, GCC's | |
1244 | implementation of variable-length arrays does not yet conform in detail | |
1245 | to the ISO C99 standard.) These arrays are | |
c1f7febf RK |
1246 | declared like any other automatic arrays, but with a length that is not |
1247 | a constant expression. The storage is allocated at the point of | |
1248 | declaration and deallocated when the brace-level is exited. For | |
1249 | example: | |
1250 | ||
3ab51846 | 1251 | @smallexample |
c1f7febf RK |
1252 | FILE * |
1253 | concat_fopen (char *s1, char *s2, char *mode) | |
1254 | @{ | |
1255 | char str[strlen (s1) + strlen (s2) + 1]; | |
1256 | strcpy (str, s1); | |
1257 | strcat (str, s2); | |
1258 | return fopen (str, mode); | |
1259 | @} | |
3ab51846 | 1260 | @end smallexample |
c1f7febf RK |
1261 | |
1262 | @cindex scope of a variable length array | |
1263 | @cindex variable-length array scope | |
1264 | @cindex deallocating variable length arrays | |
1265 | Jumping or breaking out of the scope of the array name deallocates the | |
1266 | storage. Jumping into the scope is not allowed; you get an error | |
1267 | message for it. | |
1268 | ||
1269 | @cindex @code{alloca} vs variable-length arrays | |
1270 | You can use the function @code{alloca} to get an effect much like | |
1271 | variable-length arrays. The function @code{alloca} is available in | |
1272 | many other C implementations (but not in all). On the other hand, | |
1273 | variable-length arrays are more elegant. | |
1274 | ||
1275 | There are other differences between these two methods. Space allocated | |
1276 | with @code{alloca} exists until the containing @emph{function} returns. | |
1277 | The space for a variable-length array is deallocated as soon as the array | |
1278 | name's scope ends. (If you use both variable-length arrays and | |
1279 | @code{alloca} in the same function, deallocation of a variable-length array | |
1280 | will also deallocate anything more recently allocated with @code{alloca}.) | |
1281 | ||
1282 | You can also use variable-length arrays as arguments to functions: | |
1283 | ||
3ab51846 | 1284 | @smallexample |
c1f7febf RK |
1285 | struct entry |
1286 | tester (int len, char data[len][len]) | |
1287 | @{ | |
0d893a63 | 1288 | /* @r{@dots{}} */ |
c1f7febf | 1289 | @} |
3ab51846 | 1290 | @end smallexample |
c1f7febf RK |
1291 | |
1292 | The length of an array is computed once when the storage is allocated | |
1293 | and is remembered for the scope of the array in case you access it with | |
1294 | @code{sizeof}. | |
1295 | ||
1296 | If you want to pass the array first and the length afterward, you can | |
1297 | use a forward declaration in the parameter list---another GNU extension. | |
1298 | ||
3ab51846 | 1299 | @smallexample |
c1f7febf RK |
1300 | struct entry |
1301 | tester (int len; char data[len][len], int len) | |
1302 | @{ | |
0d893a63 | 1303 | /* @r{@dots{}} */ |
c1f7febf | 1304 | @} |
3ab51846 | 1305 | @end smallexample |
c1f7febf RK |
1306 | |
1307 | @cindex parameter forward declaration | |
1308 | The @samp{int len} before the semicolon is a @dfn{parameter forward | |
1309 | declaration}, and it serves the purpose of making the name @code{len} | |
1310 | known when the declaration of @code{data} is parsed. | |
1311 | ||
1312 | You can write any number of such parameter forward declarations in the | |
1313 | parameter list. They can be separated by commas or semicolons, but the | |
1314 | last one must end with a semicolon, which is followed by the ``real'' | |
1315 | parameter declarations. Each forward declaration must match a ``real'' | |
4b404517 JM |
1316 | declaration in parameter name and data type. ISO C99 does not support |
1317 | parameter forward declarations. | |
c1f7febf | 1318 | |
ccd96f0a NB |
1319 | @node Variadic Macros |
1320 | @section Macros with a Variable Number of Arguments. | |
c1f7febf RK |
1321 | @cindex variable number of arguments |
1322 | @cindex macro with variable arguments | |
1323 | @cindex rest argument (in macro) | |
ccd96f0a | 1324 | @cindex variadic macros |
c1f7febf | 1325 | |
ccd96f0a NB |
1326 | In the ISO C standard of 1999, a macro can be declared to accept a |
1327 | variable number of arguments much as a function can. The syntax for | |
1328 | defining the macro is similar to that of a function. Here is an | |
1329 | example: | |
c1f7febf | 1330 | |
478c9e72 | 1331 | @smallexample |
ccd96f0a | 1332 | #define debug(format, ...) fprintf (stderr, format, __VA_ARGS__) |
478c9e72 | 1333 | @end smallexample |
c1f7febf | 1334 | |
ccd96f0a NB |
1335 | Here @samp{@dots{}} is a @dfn{variable argument}. In the invocation of |
1336 | such a macro, it represents the zero or more tokens until the closing | |
1337 | parenthesis that ends the invocation, including any commas. This set of | |
1338 | tokens replaces the identifier @code{__VA_ARGS__} in the macro body | |
1339 | wherever it appears. See the CPP manual for more information. | |
1340 | ||
1341 | GCC has long supported variadic macros, and used a different syntax that | |
1342 | allowed you to give a name to the variable arguments just like any other | |
1343 | argument. Here is an example: | |
c1f7febf | 1344 | |
3ab51846 | 1345 | @smallexample |
ccd96f0a | 1346 | #define debug(format, args...) fprintf (stderr, format, args) |
3ab51846 | 1347 | @end smallexample |
c1f7febf | 1348 | |
ccd96f0a NB |
1349 | This is in all ways equivalent to the ISO C example above, but arguably |
1350 | more readable and descriptive. | |
c1f7febf | 1351 | |
ccd96f0a NB |
1352 | GNU CPP has two further variadic macro extensions, and permits them to |
1353 | be used with either of the above forms of macro definition. | |
1354 | ||
1355 | In standard C, you are not allowed to leave the variable argument out | |
1356 | entirely; but you are allowed to pass an empty argument. For example, | |
1357 | this invocation is invalid in ISO C, because there is no comma after | |
1358 | the string: | |
c1f7febf | 1359 | |
3ab51846 | 1360 | @smallexample |
ccd96f0a | 1361 | debug ("A message") |
3ab51846 | 1362 | @end smallexample |
c1f7febf | 1363 | |
ccd96f0a NB |
1364 | GNU CPP permits you to completely omit the variable arguments in this |
1365 | way. In the above examples, the compiler would complain, though since | |
1366 | the expansion of the macro still has the extra comma after the format | |
1367 | string. | |
1368 | ||
1369 | To help solve this problem, CPP behaves specially for variable arguments | |
1370 | used with the token paste operator, @samp{##}. If instead you write | |
c1f7febf | 1371 | |
478c9e72 | 1372 | @smallexample |
ccd96f0a | 1373 | #define debug(format, ...) fprintf (stderr, format, ## __VA_ARGS__) |
478c9e72 | 1374 | @end smallexample |
c1f7febf | 1375 | |
ccd96f0a NB |
1376 | and if the variable arguments are omitted or empty, the @samp{##} |
1377 | operator causes the preprocessor to remove the comma before it. If you | |
1378 | do provide some variable arguments in your macro invocation, GNU CPP | |
1379 | does not complain about the paste operation and instead places the | |
1380 | variable arguments after the comma. Just like any other pasted macro | |
1381 | argument, these arguments are not macro expanded. | |
1382 | ||
1383 | @node Escaped Newlines | |
1384 | @section Slightly Looser Rules for Escaped Newlines | |
1385 | @cindex escaped newlines | |
1386 | @cindex newlines (escaped) | |
1387 | ||
f458d1d5 ZW |
1388 | Recently, the preprocessor has relaxed its treatment of escaped |
1389 | newlines. Previously, the newline had to immediately follow a | |
e6cc3a24 ZW |
1390 | backslash. The current implementation allows whitespace in the form |
1391 | of spaces, horizontal and vertical tabs, and form feeds between the | |
ccd96f0a NB |
1392 | backslash and the subsequent newline. The preprocessor issues a |
1393 | warning, but treats it as a valid escaped newline and combines the two | |
1394 | lines to form a single logical line. This works within comments and | |
e6cc3a24 ZW |
1395 | tokens, as well as between tokens. Comments are @emph{not} treated as |
1396 | whitespace for the purposes of this relaxation, since they have not | |
1397 | yet been replaced with spaces. | |
c1f7febf RK |
1398 | |
1399 | @node Subscripting | |
1400 | @section Non-Lvalue Arrays May Have Subscripts | |
1401 | @cindex subscripting | |
1402 | @cindex arrays, non-lvalue | |
1403 | ||
1404 | @cindex subscripting and function values | |
207bf485 JM |
1405 | In ISO C99, arrays that are not lvalues still decay to pointers, and |
1406 | may be subscripted, although they may not be modified or used after | |
1407 | the next sequence point and the unary @samp{&} operator may not be | |
1408 | applied to them. As an extension, GCC allows such arrays to be | |
1409 | subscripted in C89 mode, though otherwise they do not decay to | |
1410 | pointers outside C99 mode. For example, | |
4b404517 | 1411 | this is valid in GNU C though not valid in C89: |
c1f7febf | 1412 | |
3ab51846 | 1413 | @smallexample |
c1f7febf RK |
1414 | @group |
1415 | struct foo @{int a[4];@}; | |
1416 | ||
1417 | struct foo f(); | |
1418 | ||
1419 | bar (int index) | |
1420 | @{ | |
1421 | return f().a[index]; | |
1422 | @} | |
1423 | @end group | |
3ab51846 | 1424 | @end smallexample |
c1f7febf RK |
1425 | |
1426 | @node Pointer Arith | |
1427 | @section Arithmetic on @code{void}- and Function-Pointers | |
1428 | @cindex void pointers, arithmetic | |
1429 | @cindex void, size of pointer to | |
1430 | @cindex function pointers, arithmetic | |
1431 | @cindex function, size of pointer to | |
1432 | ||
1433 | In GNU C, addition and subtraction operations are supported on pointers to | |
1434 | @code{void} and on pointers to functions. This is done by treating the | |
1435 | size of a @code{void} or of a function as 1. | |
1436 | ||
1437 | A consequence of this is that @code{sizeof} is also allowed on @code{void} | |
1438 | and on function types, and returns 1. | |
1439 | ||
84330467 JM |
1440 | @opindex Wpointer-arith |
1441 | The option @option{-Wpointer-arith} requests a warning if these extensions | |
c1f7febf RK |
1442 | are used. |
1443 | ||
1444 | @node Initializers | |
1445 | @section Non-Constant Initializers | |
1446 | @cindex initializers, non-constant | |
1447 | @cindex non-constant initializers | |
1448 | ||
4b404517 | 1449 | As in standard C++ and ISO C99, the elements of an aggregate initializer for an |
161d7b59 | 1450 | automatic variable are not required to be constant expressions in GNU C@. |
c1f7febf RK |
1451 | Here is an example of an initializer with run-time varying elements: |
1452 | ||
3ab51846 | 1453 | @smallexample |
c1f7febf RK |
1454 | foo (float f, float g) |
1455 | @{ | |
1456 | float beat_freqs[2] = @{ f-g, f+g @}; | |
0d893a63 | 1457 | /* @r{@dots{}} */ |
c1f7febf | 1458 | @} |
3ab51846 | 1459 | @end smallexample |
c1f7febf | 1460 | |
4b404517 JM |
1461 | @node Compound Literals |
1462 | @section Compound Literals | |
c1f7febf RK |
1463 | @cindex constructor expressions |
1464 | @cindex initializations in expressions | |
1465 | @cindex structures, constructor expression | |
1466 | @cindex expressions, constructor | |
4b404517 JM |
1467 | @cindex compound literals |
1468 | @c The GNU C name for what C99 calls compound literals was "constructor expressions". | |
c1f7febf | 1469 | |
4b404517 | 1470 | ISO C99 supports compound literals. A compound literal looks like |
c1f7febf RK |
1471 | a cast containing an initializer. Its value is an object of the |
1472 | type specified in the cast, containing the elements specified in | |
db3acfa5 JM |
1473 | the initializer; it is an lvalue. As an extension, GCC supports |
1474 | compound literals in C89 mode and in C++. | |
c1f7febf RK |
1475 | |
1476 | Usually, the specified type is a structure. Assume that | |
1477 | @code{struct foo} and @code{structure} are declared as shown: | |
1478 | ||
3ab51846 | 1479 | @smallexample |
c1f7febf | 1480 | struct foo @{int a; char b[2];@} structure; |
3ab51846 | 1481 | @end smallexample |
c1f7febf RK |
1482 | |
1483 | @noindent | |
4b404517 | 1484 | Here is an example of constructing a @code{struct foo} with a compound literal: |
c1f7febf | 1485 | |
3ab51846 | 1486 | @smallexample |
c1f7febf | 1487 | structure = ((struct foo) @{x + y, 'a', 0@}); |
3ab51846 | 1488 | @end smallexample |
c1f7febf RK |
1489 | |
1490 | @noindent | |
1491 | This is equivalent to writing the following: | |
1492 | ||
3ab51846 | 1493 | @smallexample |
c1f7febf RK |
1494 | @{ |
1495 | struct foo temp = @{x + y, 'a', 0@}; | |
1496 | structure = temp; | |
1497 | @} | |
3ab51846 | 1498 | @end smallexample |
c1f7febf | 1499 | |
4b404517 | 1500 | You can also construct an array. If all the elements of the compound literal |
c1f7febf | 1501 | are (made up of) simple constant expressions, suitable for use in |
db3acfa5 JM |
1502 | initializers of objects of static storage duration, then the compound |
1503 | literal can be coerced to a pointer to its first element and used in | |
1504 | such an initializer, as shown here: | |
c1f7febf | 1505 | |
3ab51846 | 1506 | @smallexample |
c1f7febf | 1507 | char **foo = (char *[]) @{ "x", "y", "z" @}; |
3ab51846 | 1508 | @end smallexample |
c1f7febf | 1509 | |
4b404517 JM |
1510 | Compound literals for scalar types and union types are is |
1511 | also allowed, but then the compound literal is equivalent | |
c1f7febf RK |
1512 | to a cast. |
1513 | ||
59c83dbf JJ |
1514 | As a GNU extension, GCC allows initialization of objects with static storage |
1515 | duration by compound literals (which is not possible in ISO C99, because | |
1516 | the initializer is not a constant). | |
1517 | It is handled as if the object was initialized only with the bracket | |
1eaf20ec | 1518 | enclosed list if the types of the compound literal and the object match. |
59c83dbf JJ |
1519 | The initializer list of the compound literal must be constant. |
1520 | If the object being initialized has array type of unknown size, the size is | |
ad47f1e5 | 1521 | determined by compound literal size. |
59c83dbf | 1522 | |
3ab51846 | 1523 | @smallexample |
59c83dbf JJ |
1524 | static struct foo x = (struct foo) @{1, 'a', 'b'@}; |
1525 | static int y[] = (int []) @{1, 2, 3@}; | |
1526 | static int z[] = (int [3]) @{1@}; | |
3ab51846 | 1527 | @end smallexample |
59c83dbf JJ |
1528 | |
1529 | @noindent | |
1530 | The above lines are equivalent to the following: | |
3ab51846 | 1531 | @smallexample |
59c83dbf JJ |
1532 | static struct foo x = @{1, 'a', 'b'@}; |
1533 | static int y[] = @{1, 2, 3@}; | |
ad47f1e5 | 1534 | static int z[] = @{1, 0, 0@}; |
3ab51846 | 1535 | @end smallexample |
59c83dbf | 1536 | |
4b404517 JM |
1537 | @node Designated Inits |
1538 | @section Designated Initializers | |
c1f7febf RK |
1539 | @cindex initializers with labeled elements |
1540 | @cindex labeled elements in initializers | |
1541 | @cindex case labels in initializers | |
4b404517 | 1542 | @cindex designated initializers |
c1f7febf | 1543 | |
26d4fec7 | 1544 | Standard C89 requires the elements of an initializer to appear in a fixed |
c1f7febf RK |
1545 | order, the same as the order of the elements in the array or structure |
1546 | being initialized. | |
1547 | ||
26d4fec7 JM |
1548 | In ISO C99 you can give the elements in any order, specifying the array |
1549 | indices or structure field names they apply to, and GNU C allows this as | |
1550 | an extension in C89 mode as well. This extension is not | |
c1f7febf RK |
1551 | implemented in GNU C++. |
1552 | ||
26d4fec7 | 1553 | To specify an array index, write |
c1f7febf RK |
1554 | @samp{[@var{index}] =} before the element value. For example, |
1555 | ||
3ab51846 | 1556 | @smallexample |
26d4fec7 | 1557 | int a[6] = @{ [4] = 29, [2] = 15 @}; |
3ab51846 | 1558 | @end smallexample |
c1f7febf RK |
1559 | |
1560 | @noindent | |
1561 | is equivalent to | |
1562 | ||
3ab51846 | 1563 | @smallexample |
c1f7febf | 1564 | int a[6] = @{ 0, 0, 15, 0, 29, 0 @}; |
3ab51846 | 1565 | @end smallexample |
c1f7febf RK |
1566 | |
1567 | @noindent | |
1568 | The index values must be constant expressions, even if the array being | |
1569 | initialized is automatic. | |
1570 | ||
26d4fec7 JM |
1571 | An alternative syntax for this which has been obsolete since GCC 2.5 but |
1572 | GCC still accepts is to write @samp{[@var{index}]} before the element | |
1573 | value, with no @samp{=}. | |
1574 | ||
c1f7febf | 1575 | To initialize a range of elements to the same value, write |
26d4fec7 JM |
1576 | @samp{[@var{first} ... @var{last}] = @var{value}}. This is a GNU |
1577 | extension. For example, | |
c1f7febf | 1578 | |
3ab51846 | 1579 | @smallexample |
c1f7febf | 1580 | int widths[] = @{ [0 ... 9] = 1, [10 ... 99] = 2, [100] = 3 @}; |
3ab51846 | 1581 | @end smallexample |
c1f7febf | 1582 | |
8b6a5902 JJ |
1583 | @noindent |
1584 | If the value in it has side-effects, the side-effects will happen only once, | |
1585 | not for each initialized field by the range initializer. | |
1586 | ||
c1f7febf RK |
1587 | @noindent |
1588 | Note that the length of the array is the highest value specified | |
1589 | plus one. | |
1590 | ||
1591 | In a structure initializer, specify the name of a field to initialize | |
26d4fec7 | 1592 | with @samp{.@var{fieldname} =} before the element value. For example, |
c1f7febf RK |
1593 | given the following structure, |
1594 | ||
3ab51846 | 1595 | @smallexample |
c1f7febf | 1596 | struct point @{ int x, y; @}; |
3ab51846 | 1597 | @end smallexample |
c1f7febf RK |
1598 | |
1599 | @noindent | |
1600 | the following initialization | |
1601 | ||
3ab51846 | 1602 | @smallexample |
26d4fec7 | 1603 | struct point p = @{ .y = yvalue, .x = xvalue @}; |
3ab51846 | 1604 | @end smallexample |
c1f7febf RK |
1605 | |
1606 | @noindent | |
1607 | is equivalent to | |
1608 | ||
3ab51846 | 1609 | @smallexample |
c1f7febf | 1610 | struct point p = @{ xvalue, yvalue @}; |
3ab51846 | 1611 | @end smallexample |
c1f7febf | 1612 | |
26d4fec7 JM |
1613 | Another syntax which has the same meaning, obsolete since GCC 2.5, is |
1614 | @samp{@var{fieldname}:}, as shown here: | |
c1f7febf | 1615 | |
3ab51846 | 1616 | @smallexample |
26d4fec7 | 1617 | struct point p = @{ y: yvalue, x: xvalue @}; |
3ab51846 | 1618 | @end smallexample |
c1f7febf | 1619 | |
4b404517 JM |
1620 | @cindex designators |
1621 | The @samp{[@var{index}]} or @samp{.@var{fieldname}} is known as a | |
1622 | @dfn{designator}. You can also use a designator (or the obsolete colon | |
1623 | syntax) when initializing a union, to specify which element of the union | |
1624 | should be used. For example, | |
c1f7febf | 1625 | |
3ab51846 | 1626 | @smallexample |
c1f7febf RK |
1627 | union foo @{ int i; double d; @}; |
1628 | ||
26d4fec7 | 1629 | union foo f = @{ .d = 4 @}; |
3ab51846 | 1630 | @end smallexample |
c1f7febf RK |
1631 | |
1632 | @noindent | |
1633 | will convert 4 to a @code{double} to store it in the union using | |
1634 | the second element. By contrast, casting 4 to type @code{union foo} | |
1635 | would store it into the union as the integer @code{i}, since it is | |
1636 | an integer. (@xref{Cast to Union}.) | |
1637 | ||
1638 | You can combine this technique of naming elements with ordinary C | |
1639 | initialization of successive elements. Each initializer element that | |
4b404517 | 1640 | does not have a designator applies to the next consecutive element of the |
c1f7febf RK |
1641 | array or structure. For example, |
1642 | ||
3ab51846 | 1643 | @smallexample |
c1f7febf | 1644 | int a[6] = @{ [1] = v1, v2, [4] = v4 @}; |
3ab51846 | 1645 | @end smallexample |
c1f7febf RK |
1646 | |
1647 | @noindent | |
1648 | is equivalent to | |
1649 | ||
3ab51846 | 1650 | @smallexample |
c1f7febf | 1651 | int a[6] = @{ 0, v1, v2, 0, v4, 0 @}; |
3ab51846 | 1652 | @end smallexample |
c1f7febf RK |
1653 | |
1654 | Labeling the elements of an array initializer is especially useful | |
1655 | when the indices are characters or belong to an @code{enum} type. | |
1656 | For example: | |
1657 | ||
3ab51846 | 1658 | @smallexample |
c1f7febf RK |
1659 | int whitespace[256] |
1660 | = @{ [' '] = 1, ['\t'] = 1, ['\h'] = 1, | |
1661 | ['\f'] = 1, ['\n'] = 1, ['\r'] = 1 @}; | |
3ab51846 | 1662 | @end smallexample |
c1f7febf | 1663 | |
4b404517 | 1664 | @cindex designator lists |
26d4fec7 | 1665 | You can also write a series of @samp{.@var{fieldname}} and |
4b404517 | 1666 | @samp{[@var{index}]} designators before an @samp{=} to specify a |
26d4fec7 JM |
1667 | nested subobject to initialize; the list is taken relative to the |
1668 | subobject corresponding to the closest surrounding brace pair. For | |
1669 | example, with the @samp{struct point} declaration above: | |
1670 | ||
478c9e72 | 1671 | @smallexample |
26d4fec7 | 1672 | struct point ptarray[10] = @{ [2].y = yv2, [2].x = xv2, [0].x = xv0 @}; |
478c9e72 | 1673 | @end smallexample |
26d4fec7 | 1674 | |
8b6a5902 JJ |
1675 | @noindent |
1676 | If the same field is initialized multiple times, it will have value from | |
1677 | the last initialization. If any such overridden initialization has | |
1678 | side-effect, it is unspecified whether the side-effect happens or not. | |
2dd76960 | 1679 | Currently, GCC will discard them and issue a warning. |
8b6a5902 | 1680 | |
c1f7febf RK |
1681 | @node Case Ranges |
1682 | @section Case Ranges | |
1683 | @cindex case ranges | |
1684 | @cindex ranges in case statements | |
1685 | ||
1686 | You can specify a range of consecutive values in a single @code{case} label, | |
1687 | like this: | |
1688 | ||
3ab51846 | 1689 | @smallexample |
c1f7febf | 1690 | case @var{low} ... @var{high}: |
3ab51846 | 1691 | @end smallexample |
c1f7febf RK |
1692 | |
1693 | @noindent | |
1694 | This has the same effect as the proper number of individual @code{case} | |
1695 | labels, one for each integer value from @var{low} to @var{high}, inclusive. | |
1696 | ||
1697 | This feature is especially useful for ranges of ASCII character codes: | |
1698 | ||
3ab51846 | 1699 | @smallexample |
c1f7febf | 1700 | case 'A' ... 'Z': |
3ab51846 | 1701 | @end smallexample |
c1f7febf RK |
1702 | |
1703 | @strong{Be careful:} Write spaces around the @code{...}, for otherwise | |
1704 | it may be parsed wrong when you use it with integer values. For example, | |
1705 | write this: | |
1706 | ||
3ab51846 | 1707 | @smallexample |
c1f7febf | 1708 | case 1 ... 5: |
3ab51846 | 1709 | @end smallexample |
c1f7febf RK |
1710 | |
1711 | @noindent | |
1712 | rather than this: | |
1713 | ||
3ab51846 | 1714 | @smallexample |
c1f7febf | 1715 | case 1...5: |
3ab51846 | 1716 | @end smallexample |
c1f7febf RK |
1717 | |
1718 | @node Cast to Union | |
1719 | @section Cast to a Union Type | |
1720 | @cindex cast to a union | |
1721 | @cindex union, casting to a | |
1722 | ||
1723 | A cast to union type is similar to other casts, except that the type | |
1724 | specified is a union type. You can specify the type either with | |
1725 | @code{union @var{tag}} or with a typedef name. A cast to union is actually | |
1726 | a constructor though, not a cast, and hence does not yield an lvalue like | |
4b404517 | 1727 | normal casts. (@xref{Compound Literals}.) |
c1f7febf RK |
1728 | |
1729 | The types that may be cast to the union type are those of the members | |
1730 | of the union. Thus, given the following union and variables: | |
1731 | ||
3ab51846 | 1732 | @smallexample |
c1f7febf RK |
1733 | union foo @{ int i; double d; @}; |
1734 | int x; | |
1735 | double y; | |
3ab51846 | 1736 | @end smallexample |
c1f7febf RK |
1737 | |
1738 | @noindent | |
aee96fe9 | 1739 | both @code{x} and @code{y} can be cast to type @code{union foo}. |
c1f7febf RK |
1740 | |
1741 | Using the cast as the right-hand side of an assignment to a variable of | |
1742 | union type is equivalent to storing in a member of the union: | |
1743 | ||
3ab51846 | 1744 | @smallexample |
c1f7febf | 1745 | union foo u; |
0d893a63 | 1746 | /* @r{@dots{}} */ |
c1f7febf RK |
1747 | u = (union foo) x @equiv{} u.i = x |
1748 | u = (union foo) y @equiv{} u.d = y | |
3ab51846 | 1749 | @end smallexample |
c1f7febf RK |
1750 | |
1751 | You can also use the union cast as a function argument: | |
1752 | ||
3ab51846 | 1753 | @smallexample |
c1f7febf | 1754 | void hack (union foo); |
0d893a63 | 1755 | /* @r{@dots{}} */ |
c1f7febf | 1756 | hack ((union foo) x); |
3ab51846 | 1757 | @end smallexample |
c1f7febf | 1758 | |
4b404517 JM |
1759 | @node Mixed Declarations |
1760 | @section Mixed Declarations and Code | |
1761 | @cindex mixed declarations and code | |
1762 | @cindex declarations, mixed with code | |
1763 | @cindex code, mixed with declarations | |
1764 | ||
1765 | ISO C99 and ISO C++ allow declarations and code to be freely mixed | |
1766 | within compound statements. As an extension, GCC also allows this in | |
1767 | C89 mode. For example, you could do: | |
1768 | ||
3ab51846 | 1769 | @smallexample |
4b404517 | 1770 | int i; |
0d893a63 | 1771 | /* @r{@dots{}} */ |
4b404517 JM |
1772 | i++; |
1773 | int j = i + 2; | |
3ab51846 | 1774 | @end smallexample |
4b404517 JM |
1775 | |
1776 | Each identifier is visible from where it is declared until the end of | |
1777 | the enclosing block. | |
1778 | ||
c1f7febf RK |
1779 | @node Function Attributes |
1780 | @section Declaring Attributes of Functions | |
1781 | @cindex function attributes | |
1782 | @cindex declaring attributes of functions | |
1783 | @cindex functions that never return | |
6e9a3221 | 1784 | @cindex functions that return more than once |
c1f7febf RK |
1785 | @cindex functions that have no side effects |
1786 | @cindex functions in arbitrary sections | |
2a59078d | 1787 | @cindex functions that behave like malloc |
c1f7febf RK |
1788 | @cindex @code{volatile} applied to function |
1789 | @cindex @code{const} applied to function | |
26f6672d | 1790 | @cindex functions with @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon} style arguments |
b34c7881 | 1791 | @cindex functions with non-null pointer arguments |
c1f7febf RK |
1792 | @cindex functions that are passed arguments in registers on the 386 |
1793 | @cindex functions that pop the argument stack on the 386 | |
1794 | @cindex functions that do not pop the argument stack on the 386 | |
1795 | ||
1796 | In GNU C, you declare certain things about functions called in your program | |
1797 | which help the compiler optimize function calls and check your code more | |
1798 | carefully. | |
1799 | ||
1800 | The keyword @code{__attribute__} allows you to specify special | |
1801 | attributes when making a declaration. This keyword is followed by an | |
9162542e | 1802 | attribute specification inside double parentheses. The following |
eacecf96 | 1803 | attributes are currently defined for functions on all targets: |
837edd5f GK |
1804 | @code{aligned}, @code{alloc_size}, @code{noreturn}, |
1805 | @code{returns_twice}, @code{noinline}, @code{always_inline}, | |
1806 | @code{flatten}, @code{pure}, @code{const}, @code{nothrow}, | |
1807 | @code{sentinel}, @code{format}, @code{format_arg}, | |
51bc54a6 DM |
1808 | @code{no_instrument_function}, @code{section}, @code{constructor}, |
1809 | @code{destructor}, @code{used}, @code{unused}, @code{deprecated}, | |
1810 | @code{weak}, @code{malloc}, @code{alias}, @code{warn_unused_result}, | |
d752cfdb | 1811 | @code{nonnull}, @code{gnu_inline}, @code{externally_visible}, |
d2af6a68 JJ |
1812 | @code{hot}, @code{cold}, @code{artificial}, @code{error} |
1813 | and @code{warning}. | |
837edd5f GK |
1814 | Several other attributes are defined for functions on particular |
1815 | target systems. Other attributes, including @code{section} are | |
1816 | supported for variables declarations (@pxref{Variable Attributes}) and | |
1817 | for types (@pxref{Type Attributes}). | |
c1f7febf RK |
1818 | |
1819 | You may also specify attributes with @samp{__} preceding and following | |
1820 | each keyword. This allows you to use them in header files without | |
1821 | being concerned about a possible macro of the same name. For example, | |
1822 | you may use @code{__noreturn__} instead of @code{noreturn}. | |
1823 | ||
2c5e91d2 JM |
1824 | @xref{Attribute Syntax}, for details of the exact syntax for using |
1825 | attributes. | |
1826 | ||
c1f7febf | 1827 | @table @code |
8a36672b | 1828 | @c Keep this table alphabetized by attribute name. Treat _ as space. |
c1f7febf | 1829 | |
c8619b90 NS |
1830 | @item alias ("@var{target}") |
1831 | @cindex @code{alias} attribute | |
1832 | The @code{alias} attribute causes the declaration to be emitted as an | |
1833 | alias for another symbol, which must be specified. For instance, | |
c1f7febf RK |
1834 | |
1835 | @smallexample | |
c8619b90 NS |
1836 | void __f () @{ /* @r{Do something.} */; @} |
1837 | void f () __attribute__ ((weak, alias ("__f"))); | |
c1f7febf RK |
1838 | @end smallexample |
1839 | ||
a9b0b825 | 1840 | defines @samp{f} to be a weak alias for @samp{__f}. In C++, the |
52eb57df RH |
1841 | mangled name for the target must be used. It is an error if @samp{__f} |
1842 | is not defined in the same translation unit. | |
c8619b90 NS |
1843 | |
1844 | Not all target machines support this attribute. | |
9162542e | 1845 | |
837edd5f GK |
1846 | @item aligned (@var{alignment}) |
1847 | @cindex @code{aligned} attribute | |
1848 | This attribute specifies a minimum alignment for the function, | |
1849 | measured in bytes. | |
1850 | ||
1851 | You cannot use this attribute to decrease the alignment of a function, | |
1852 | only to increase it. However, when you explicitly specify a function | |
1853 | alignment this will override the effect of the | |
1854 | @option{-falign-functions} (@pxref{Optimize Options}) option for this | |
1855 | function. | |
1856 | ||
1857 | Note that the effectiveness of @code{aligned} attributes may be | |
1858 | limited by inherent limitations in your linker. On many systems, the | |
1859 | linker is only able to arrange for functions to be aligned up to a | |
1860 | certain maximum alignment. (For some linkers, the maximum supported | |
1861 | alignment may be very very small.) See your linker documentation for | |
1862 | further information. | |
1863 | ||
1864 | The @code{aligned} attribute can also be used for variables and fields | |
1865 | (@pxref{Variable Attributes}.) | |
1866 | ||
51bc54a6 DM |
1867 | @item alloc_size |
1868 | @cindex @code{alloc_size} attribute | |
1869 | The @code{alloc_size} attribute is used to tell the compiler that the | |
1870 | function return value points to memory, where the size is given by | |
1871 | one or two of the functions parameters. GCC uses this | |
1872 | information to improve the correctness of @code{__builtin_object_size}. | |
1873 | ||
1874 | The function parameter(s) denoting the allocated size are specified by | |
1875 | one or two integer arguments supplied to the attribute. The allocated size | |
1876 | is either the value of the single function argument specified or the product | |
1877 | of the two function arguments specified. Argument numbering starts at | |
1878 | one. | |
1879 | ||
1880 | For instance, | |
1881 | ||
1882 | @smallexample | |
1883 | void* my_calloc(size_t, size_t) __attribute__((alloc_size(1,2))) | |
1c42f5c6 | 1884 | void my_realloc(void*, size_t) __attribute__((alloc_size(2))) |
51bc54a6 DM |
1885 | @end smallexample |
1886 | ||
1887 | declares that my_calloc will return memory of the size given by | |
1888 | the product of parameter 1 and 2 and that my_realloc will return memory | |
1889 | of the size given by parameter 2. | |
1890 | ||
6aa77e6c | 1891 | @item always_inline |
c8619b90 | 1892 | @cindex @code{always_inline} function attribute |
6aa77e6c AH |
1893 | Generally, functions are not inlined unless optimization is specified. |
1894 | For functions declared inline, this attribute inlines the function even | |
1895 | if no optimization level was specified. | |
1896 | ||
4eb7fd83 JJ |
1897 | @item gnu_inline |
1898 | @cindex @code{gnu_inline} function attribute | |
da1c7394 ILT |
1899 | This attribute should be used with a function which is also declared |
1900 | with the @code{inline} keyword. It directs GCC to treat the function | |
1901 | as if it were defined in gnu89 mode even when compiling in C99 or | |
1902 | gnu99 mode. | |
1903 | ||
1904 | If the function is declared @code{extern}, then this definition of the | |
1905 | function is used only for inlining. In no case is the function | |
1906 | compiled as a standalone function, not even if you take its address | |
1907 | explicitly. Such an address becomes an external reference, as if you | |
1908 | had only declared the function, and had not defined it. This has | |
1909 | almost the effect of a macro. The way to use this is to put a | |
1910 | function definition in a header file with this attribute, and put | |
1911 | another copy of the function, without @code{extern}, in a library | |
1912 | file. The definition in the header file will cause most calls to the | |
1913 | function to be inlined. If any uses of the function remain, they will | |
1914 | refer to the single copy in the library. Note that the two | |
1915 | definitions of the functions need not be precisely the same, although | |
1916 | if they do not have the same effect your program may behave oddly. | |
1917 | ||
3a47c4e4 AO |
1918 | In C, if the function is neither @code{extern} nor @code{static}, then |
1919 | the function is compiled as a standalone function, as well as being | |
da1c7394 ILT |
1920 | inlined where possible. |
1921 | ||
1922 | This is how GCC traditionally handled functions declared | |
1923 | @code{inline}. Since ISO C99 specifies a different semantics for | |
1924 | @code{inline}, this function attribute is provided as a transition | |
1925 | measure and as a useful feature in its own right. This attribute is | |
1926 | available in GCC 4.1.3 and later. It is available if either of the | |
1927 | preprocessor macros @code{__GNUC_GNU_INLINE__} or | |
1928 | @code{__GNUC_STDC_INLINE__} are defined. @xref{Inline,,An Inline | |
1929 | Function is As Fast As a Macro}. | |
4eb7fd83 | 1930 | |
3a47c4e4 AO |
1931 | In C++, this attribute does not depend on @code{extern} in any way, |
1932 | but it still requires the @code{inline} keyword to enable its special | |
1933 | behavior. | |
1934 | ||
d752cfdb JJ |
1935 | @cindex @code{artificial} function attribute |
1936 | @item artificial | |
1937 | This attribute is useful for small inline wrappers which if possible | |
1938 | should appear during debugging as a unit, depending on the debug | |
1939 | info format it will either mean marking the function as artificial | |
1940 | or using the caller location for all instructions within the inlined | |
1941 | body. | |
1942 | ||
0691d1d4 RG |
1943 | @cindex @code{flatten} function attribute |
1944 | @item flatten | |
1945 | Generally, inlining into a function is limited. For a function marked with | |
1946 | this attribute, every call inside this function will be inlined, if possible. | |
1947 | Whether the function itself is considered for inlining depends on its size and | |
1948 | the current inlining parameters. The @code{flatten} attribute only works | |
1949 | reliably in unit-at-a-time mode. | |
1950 | ||
d2af6a68 JJ |
1951 | @item error ("@var{message}") |
1952 | @cindex @code{error} function attribute | |
1953 | If this attribute is used on a function declaration and a call to such a function | |
1954 | is not eliminated through dead code elimination or other optimizations, an error | |
1955 | which will include @var{message} will be diagnosed. This is useful | |
1956 | for compile time checking, especially together with @code{__builtin_constant_p} | |
1957 | and inline functions where checking the inline function arguments is not | |
1958 | possible through @code{extern char [(condition) ? 1 : -1];} tricks. | |
1959 | While it is possible to leave the function undefined and thus invoke | |
1960 | a link failure, when using this attribute the problem will be diagnosed | |
1961 | earlier and with exact location of the call even in presence of inline | |
1962 | functions or when not emitting debugging information. | |
1963 | ||
1964 | @item warning ("@var{message}") | |
1965 | @cindex @code{warning} function attribute | |
1966 | If this attribute is used on a function declaration and a call to such a function | |
1967 | is not eliminated through dead code elimination or other optimizations, a warning | |
1968 | which will include @var{message} will be diagnosed. This is useful | |
1969 | for compile time checking, especially together with @code{__builtin_constant_p} | |
1970 | and inline functions. While it is possible to define the function with | |
1971 | a message in @code{.gnu.warning*} section, when using this attribute the problem | |
1972 | will be diagnosed earlier and with exact location of the call even in presence | |
1973 | of inline functions or when not emitting debugging information. | |
1974 | ||
c8619b90 NS |
1975 | @item cdecl |
1976 | @cindex functions that do pop the argument stack on the 386 | |
1977 | @opindex mrtd | |
1978 | On the Intel 386, the @code{cdecl} attribute causes the compiler to | |
1979 | assume that the calling function will pop off the stack space used to | |
1980 | pass arguments. This is | |
1981 | useful to override the effects of the @option{-mrtd} switch. | |
2a8f6b90 | 1982 | |
2a8f6b90 | 1983 | @item const |
c8619b90 | 1984 | @cindex @code{const} function attribute |
2a8f6b90 JH |
1985 | Many functions do not examine any values except their arguments, and |
1986 | have no effects except the return value. Basically this is just slightly | |
50c177f7 | 1987 | more strict class than the @code{pure} attribute below, since function is not |
2a59078d | 1988 | allowed to read global memory. |
2a8f6b90 JH |
1989 | |
1990 | @cindex pointer arguments | |
1991 | Note that a function that has pointer arguments and examines the data | |
1992 | pointed to must @emph{not} be declared @code{const}. Likewise, a | |
1993 | function that calls a non-@code{const} function usually must not be | |
1994 | @code{const}. It does not make sense for a @code{const} function to | |
1995 | return @code{void}. | |
1996 | ||
f0523f02 | 1997 | The attribute @code{const} is not implemented in GCC versions earlier |
c1f7febf RK |
1998 | than 2.5. An alternative way to declare that a function has no side |
1999 | effects, which works in the current version and in some older versions, | |
2000 | is as follows: | |
2001 | ||
2002 | @smallexample | |
2003 | typedef int intfn (); | |
2004 | ||
2005 | extern const intfn square; | |
2006 | @end smallexample | |
2007 | ||
2008 | This approach does not work in GNU C++ from 2.6.0 on, since the language | |
2009 | specifies that the @samp{const} must be attached to the return value. | |
2010 | ||
c8619b90 NS |
2011 | @item constructor |
2012 | @itemx destructor | |
fc8600f9 MM |
2013 | @itemx constructor (@var{priority}) |
2014 | @itemx destructor (@var{priority}) | |
c8619b90 NS |
2015 | @cindex @code{constructor} function attribute |
2016 | @cindex @code{destructor} function attribute | |
2017 | The @code{constructor} attribute causes the function to be called | |
2018 | automatically before execution enters @code{main ()}. Similarly, the | |
2019 | @code{destructor} attribute causes the function to be called | |
2020 | automatically after @code{main ()} has completed or @code{exit ()} has | |
2021 | been called. Functions with these attributes are useful for | |
2022 | initializing data that will be used implicitly during the execution of | |
2023 | the program. | |
2024 | ||
fc8600f9 MM |
2025 | You may provide an optional integer priority to control the order in |
2026 | which constructor and destructor functions are run. A constructor | |
2027 | with a smaller priority number runs before a constructor with a larger | |
2028 | priority number; the opposite relationship holds for destructors. So, | |
2029 | if you have a constructor that allocates a resource and a destructor | |
2030 | that deallocates the same resource, both functions typically have the | |
2031 | same priority. The priorities for constructor and destructor | |
2032 | functions are the same as those specified for namespace-scope C++ | |
2033 | objects (@pxref{C++ Attributes}). | |
2034 | ||
c8619b90 NS |
2035 | These attributes are not currently implemented for Objective-C@. |
2036 | ||
2037 | @item deprecated | |
2038 | @cindex @code{deprecated} attribute. | |
2039 | The @code{deprecated} attribute results in a warning if the function | |
2040 | is used anywhere in the source file. This is useful when identifying | |
2041 | functions that are expected to be removed in a future version of a | |
2042 | program. The warning also includes the location of the declaration | |
2043 | of the deprecated function, to enable users to easily find further | |
2044 | information about why the function is deprecated, or what they should | |
2045 | do instead. Note that the warnings only occurs for uses: | |
2046 | ||
2047 | @smallexample | |
2048 | int old_fn () __attribute__ ((deprecated)); | |
2049 | int old_fn (); | |
2050 | int (*fn_ptr)() = old_fn; | |
2051 | @end smallexample | |
2052 | ||
2053 | results in a warning on line 3 but not line 2. | |
2054 | ||
2055 | The @code{deprecated} attribute can also be used for variables and | |
2056 | types (@pxref{Variable Attributes}, @pxref{Type Attributes}.) | |
2057 | ||
2058 | @item dllexport | |
2059 | @cindex @code{__declspec(dllexport)} | |
b2ca3702 MM |
2060 | On Microsoft Windows targets and Symbian OS targets the |
2061 | @code{dllexport} attribute causes the compiler to provide a global | |
2062 | pointer to a pointer in a DLL, so that it can be referenced with the | |
2063 | @code{dllimport} attribute. On Microsoft Windows targets, the pointer | |
2064 | name is formed by combining @code{_imp__} and the function or variable | |
2065 | name. | |
2066 | ||
2067 | You can use @code{__declspec(dllexport)} as a synonym for | |
2068 | @code{__attribute__ ((dllexport))} for compatibility with other | |
2069 | compilers. | |
2070 | ||
2071 | On systems that support the @code{visibility} attribute, this | |
3a687f8b MM |
2072 | attribute also implies ``default'' visibility. It is an error to |
2073 | explicitly specify any other visibility. | |
c8619b90 | 2074 | |
b2ca3702 MM |
2075 | Currently, the @code{dllexport} attribute is ignored for inlined |
2076 | functions, unless the @option{-fkeep-inline-functions} flag has been | |
2077 | used. The attribute is also ignored for undefined symbols. | |
c8619b90 | 2078 | |
8a36672b JM |
2079 | When applied to C++ classes, the attribute marks defined non-inlined |
2080 | member functions and static data members as exports. Static consts | |
c8619b90 NS |
2081 | initialized in-class are not marked unless they are also defined |
2082 | out-of-class. | |
2083 | ||
b55e3aad | 2084 | For Microsoft Windows targets there are alternative methods for |
b2ca3702 | 2085 | including the symbol in the DLL's export table such as using a |
b55e3aad NC |
2086 | @file{.def} file with an @code{EXPORTS} section or, with GNU ld, using |
2087 | the @option{--export-all} linker flag. | |
c8619b90 NS |
2088 | |
2089 | @item dllimport | |
2090 | @cindex @code{__declspec(dllimport)} | |
b2ca3702 | 2091 | On Microsoft Windows and Symbian OS targets, the @code{dllimport} |
b55e3aad | 2092 | attribute causes the compiler to reference a function or variable via |
b2ca3702 | 2093 | a global pointer to a pointer that is set up by the DLL exporting the |
3a687f8b MM |
2094 | symbol. The attribute implies @code{extern}. On Microsoft Windows |
2095 | targets, the pointer name is formed by combining @code{_imp__} and the | |
2096 | function or variable name. | |
b2ca3702 MM |
2097 | |
2098 | You can use @code{__declspec(dllimport)} as a synonym for | |
2099 | @code{__attribute__ ((dllimport))} for compatibility with other | |
2100 | compilers. | |
c8619b90 | 2101 | |
3a687f8b MM |
2102 | On systems that support the @code{visibility} attribute, this |
2103 | attribute also implies ``default'' visibility. It is an error to | |
2104 | explicitly specify any other visibility. | |
2105 | ||
8a36672b | 2106 | Currently, the attribute is ignored for inlined functions. If the |
c8619b90 NS |
2107 | attribute is applied to a symbol @emph{definition}, an error is reported. |
2108 | If a symbol previously declared @code{dllimport} is later defined, the | |
2109 | attribute is ignored in subsequent references, and a warning is emitted. | |
2110 | The attribute is also overridden by a subsequent declaration as | |
2111 | @code{dllexport}. | |
2112 | ||
2113 | When applied to C++ classes, the attribute marks non-inlined | |
2114 | member functions and static data members as imports. However, the | |
2115 | attribute is ignored for virtual methods to allow creation of vtables | |
2116 | using thunks. | |
2117 | ||
b2ca3702 | 2118 | On the SH Symbian OS target the @code{dllimport} attribute also has |
78466c0e | 2119 | another affect---it can cause the vtable and run-time type information |
b2ca3702 MM |
2120 | for a class to be exported. This happens when the class has a |
2121 | dllimport'ed constructor or a non-inline, non-pure virtual function | |
2122 | and, for either of those two conditions, the class also has a inline | |
2123 | constructor or destructor and has a key function that is defined in | |
2124 | the current translation unit. | |
b55e3aad NC |
2125 | |
2126 | For Microsoft Windows based targets the use of the @code{dllimport} | |
2127 | attribute on functions is not necessary, but provides a small | |
8a36672b | 2128 | performance benefit by eliminating a thunk in the DLL@. The use of the |
b55e3aad | 2129 | @code{dllimport} attribute on imported variables was required on older |
b2ca3702 | 2130 | versions of the GNU linker, but can now be avoided by passing the |
8a36672b | 2131 | @option{--enable-auto-import} switch to the GNU linker. As with |
b2ca3702 | 2132 | functions, using the attribute for a variable eliminates a thunk in |
8a36672b | 2133 | the DLL@. |
b2ca3702 | 2134 | |
d32034a7 DS |
2135 | One drawback to using this attribute is that a pointer to a |
2136 | @emph{variable} marked as @code{dllimport} cannot be used as a constant | |
2137 | address. However, a pointer to a @emph{function} with the | |
2138 | @code{dllimport} attribute can be used as a constant initializer; in | |
2139 | this case, the address of a stub function in the import lib is | |
2140 | referenced. On Microsoft Windows targets, the attribute can be disabled | |
b2ca3702 | 2141 | for functions by setting the @option{-mnop-fun-dllimport} flag. |
c8619b90 NS |
2142 | |
2143 | @item eightbit_data | |
2144 | @cindex eight bit data on the H8/300, H8/300H, and H8S | |
2145 | Use this attribute on the H8/300, H8/300H, and H8S to indicate that the specified | |
2146 | variable should be placed into the eight bit data section. | |
2147 | The compiler will generate more efficient code for certain operations | |
2148 | on data in the eight bit data area. Note the eight bit data area is limited to | |
2149 | 256 bytes of data. | |
2150 | ||
2151 | You must use GAS and GLD from GNU binutils version 2.7 or later for | |
2152 | this attribute to work correctly. | |
2153 | ||
0d4a78eb BS |
2154 | @item exception_handler |
2155 | @cindex exception handler functions on the Blackfin processor | |
2156 | Use this attribute on the Blackfin to indicate that the specified function | |
2157 | is an exception handler. The compiler will generate function entry and | |
2158 | exit sequences suitable for use in an exception handler when this | |
2159 | attribute is present. | |
2160 | ||
c8619b90 NS |
2161 | @item far |
2162 | @cindex functions which handle memory bank switching | |
2163 | On 68HC11 and 68HC12 the @code{far} attribute causes the compiler to | |
2164 | use a calling convention that takes care of switching memory banks when | |
2165 | entering and leaving a function. This calling convention is also the | |
2166 | default when using the @option{-mlong-calls} option. | |
2167 | ||
2168 | On 68HC12 the compiler will use the @code{call} and @code{rtc} instructions | |
2169 | to call and return from a function. | |
2170 | ||
2171 | On 68HC11 the compiler will generate a sequence of instructions | |
2172 | to invoke a board-specific routine to switch the memory bank and call the | |
8a36672b | 2173 | real function. The board-specific routine simulates a @code{call}. |
c8619b90 | 2174 | At the end of a function, it will jump to a board-specific routine |
8a36672b | 2175 | instead of using @code{rts}. The board-specific return routine simulates |
c8619b90 NS |
2176 | the @code{rtc}. |
2177 | ||
2178 | @item fastcall | |
2179 | @cindex functions that pop the argument stack on the 386 | |
2180 | On the Intel 386, the @code{fastcall} attribute causes the compiler to | |
2f84b963 RG |
2181 | pass the first argument (if of integral type) in the register ECX and |
2182 | the second argument (if of integral type) in the register EDX@. Subsequent | |
2183 | and other typed arguments are passed on the stack. The called function will | |
2184 | pop the arguments off the stack. If the number of arguments is variable all | |
c8619b90 | 2185 | arguments are pushed on the stack. |
c1f7febf RK |
2186 | |
2187 | @item format (@var{archetype}, @var{string-index}, @var{first-to-check}) | |
2188 | @cindex @code{format} function attribute | |
84330467 | 2189 | @opindex Wformat |
bb72a084 | 2190 | The @code{format} attribute specifies that a function takes @code{printf}, |
26f6672d JM |
2191 | @code{scanf}, @code{strftime} or @code{strfmon} style arguments which |
2192 | should be type-checked against a format string. For example, the | |
2193 | declaration: | |
c1f7febf RK |
2194 | |
2195 | @smallexample | |
2196 | extern int | |
2197 | my_printf (void *my_object, const char *my_format, ...) | |
2198 | __attribute__ ((format (printf, 2, 3))); | |
2199 | @end smallexample | |
2200 | ||
2201 | @noindent | |
2202 | causes the compiler to check the arguments in calls to @code{my_printf} | |
2203 | for consistency with the @code{printf} style format string argument | |
2204 | @code{my_format}. | |
2205 | ||
2206 | The parameter @var{archetype} determines how the format string is | |
6590fc9f KT |
2207 | interpreted, and should be @code{printf}, @code{scanf}, @code{strftime}, |
2208 | @code{gnu_printf}, @code{gnu_scanf}, @code{gnu_strftime} or | |
2209 | @code{strfmon}. (You can also use @code{__printf__}, | |
2210 | @code{__scanf__}, @code{__strftime__} or @code{__strfmon__}.) On | |
2211 | MinGW targets, @code{ms_printf}, @code{ms_scanf}, and | |
2212 | @code{ms_strftime} are also present. | |
2213 | @var{archtype} values such as @code{printf} refer to the formats accepted | |
2214 | by the system's C run-time library, while @code{gnu_} values always refer | |
2215 | to the formats accepted by the GNU C Library. On Microsoft Windows | |
2216 | targets, @code{ms_} values refer to the formats accepted by the | |
2217 | @file{msvcrt.dll} library. | |
2218 | The parameter @var{string-index} | |
2219 | specifies which argument is the format string argument (starting | |
2220 | from 1), while @var{first-to-check} is the number of the first | |
2221 | argument to check against the format string. For functions | |
2222 | where the arguments are not available to be checked (such as | |
c1f7febf | 2223 | @code{vprintf}), specify the third parameter as zero. In this case the |
b722c82c JM |
2224 | compiler only checks the format string for consistency. For |
2225 | @code{strftime} formats, the third parameter is required to be zero. | |
f57a2e3a BE |
2226 | Since non-static C++ methods have an implicit @code{this} argument, the |
2227 | arguments of such methods should be counted from two, not one, when | |
2228 | giving values for @var{string-index} and @var{first-to-check}. | |
c1f7febf RK |
2229 | |
2230 | In the example above, the format string (@code{my_format}) is the second | |
2231 | argument of the function @code{my_print}, and the arguments to check | |
2232 | start with the third argument, so the correct parameters for the format | |
2233 | attribute are 2 and 3. | |
2234 | ||
84330467 | 2235 | @opindex ffreestanding |
e6e931b7 | 2236 | @opindex fno-builtin |
c1f7febf | 2237 | The @code{format} attribute allows you to identify your own functions |
f0523f02 | 2238 | which take format strings as arguments, so that GCC can check the |
b722c82c | 2239 | calls to these functions for errors. The compiler always (unless |
e6e931b7 | 2240 | @option{-ffreestanding} or @option{-fno-builtin} is used) checks formats |
b722c82c | 2241 | for the standard library functions @code{printf}, @code{fprintf}, |
bb72a084 | 2242 | @code{sprintf}, @code{scanf}, @code{fscanf}, @code{sscanf}, @code{strftime}, |
c1f7febf | 2243 | @code{vprintf}, @code{vfprintf} and @code{vsprintf} whenever such |
84330467 | 2244 | warnings are requested (using @option{-Wformat}), so there is no need to |
b722c82c JM |
2245 | modify the header file @file{stdio.h}. In C99 mode, the functions |
2246 | @code{snprintf}, @code{vsnprintf}, @code{vscanf}, @code{vfscanf} and | |
26f6672d | 2247 | @code{vsscanf} are also checked. Except in strictly conforming C |
b4c984fb KG |
2248 | standard modes, the X/Open function @code{strfmon} is also checked as |
2249 | are @code{printf_unlocked} and @code{fprintf_unlocked}. | |
b722c82c | 2250 | @xref{C Dialect Options,,Options Controlling C Dialect}. |
c1f7febf | 2251 | |
a2bec818 DJ |
2252 | The target may provide additional types of format checks. |
2253 | @xref{Target Format Checks,,Format Checks Specific to Particular | |
2254 | Target Machines}. | |
2255 | ||
c1f7febf RK |
2256 | @item format_arg (@var{string-index}) |
2257 | @cindex @code{format_arg} function attribute | |
84330467 | 2258 | @opindex Wformat-nonliteral |
26f6672d JM |
2259 | The @code{format_arg} attribute specifies that a function takes a format |
2260 | string for a @code{printf}, @code{scanf}, @code{strftime} or | |
2261 | @code{strfmon} style function and modifies it (for example, to translate | |
2262 | it into another language), so the result can be passed to a | |
2263 | @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon} style | |
2264 | function (with the remaining arguments to the format function the same | |
2265 | as they would have been for the unmodified string). For example, the | |
2266 | declaration: | |
c1f7febf RK |
2267 | |
2268 | @smallexample | |
2269 | extern char * | |
2270 | my_dgettext (char *my_domain, const char *my_format) | |
2271 | __attribute__ ((format_arg (2))); | |
2272 | @end smallexample | |
2273 | ||
2274 | @noindent | |
26f6672d JM |
2275 | causes the compiler to check the arguments in calls to a @code{printf}, |
2276 | @code{scanf}, @code{strftime} or @code{strfmon} type function, whose | |
2277 | format string argument is a call to the @code{my_dgettext} function, for | |
2278 | consistency with the format string argument @code{my_format}. If the | |
2279 | @code{format_arg} attribute had not been specified, all the compiler | |
2280 | could tell in such calls to format functions would be that the format | |
2281 | string argument is not constant; this would generate a warning when | |
84330467 | 2282 | @option{-Wformat-nonliteral} is used, but the calls could not be checked |
26f6672d | 2283 | without the attribute. |
c1f7febf RK |
2284 | |
2285 | The parameter @var{string-index} specifies which argument is the format | |
f57a2e3a BE |
2286 | string argument (starting from one). Since non-static C++ methods have |
2287 | an implicit @code{this} argument, the arguments of such methods should | |
2288 | be counted from two. | |
c1f7febf RK |
2289 | |
2290 | The @code{format-arg} attribute allows you to identify your own | |
f0523f02 | 2291 | functions which modify format strings, so that GCC can check the |
26f6672d JM |
2292 | calls to @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon} |
2293 | type function whose operands are a call to one of your own function. | |
2294 | The compiler always treats @code{gettext}, @code{dgettext}, and | |
2295 | @code{dcgettext} in this manner except when strict ISO C support is | |
84330467 | 2296 | requested by @option{-ansi} or an appropriate @option{-std} option, or |
e6e931b7 JM |
2297 | @option{-ffreestanding} or @option{-fno-builtin} |
2298 | is used. @xref{C Dialect Options,,Options | |
26f6672d | 2299 | Controlling C Dialect}. |
c1f7febf | 2300 | |
c8619b90 | 2301 | @item function_vector |
561642fa | 2302 | @cindex calling functions through the function vector on H8/300, M16C, M32C and SH2A processors |
c8619b90 NS |
2303 | Use this attribute on the H8/300, H8/300H, and H8S to indicate that the specified |
2304 | function should be called through the function vector. Calling a | |
2305 | function through the function vector will reduce code size, however; | |
2306 | the function vector has a limited size (maximum 128 entries on the H8/300 | |
2307 | and 64 entries on the H8/300H and H8S) and shares space with the interrupt vector. | |
b34c7881 | 2308 | |
561642fa AP |
2309 | In SH2A target, this attribute declares a function to be called using the |
2310 | TBR relative addressing mode. The argument to this attribute is the entry | |
2311 | number of the same function in a vector table containing all the TBR | |
2312 | relative addressable functions. For the successful jump, register TBR | |
2313 | should contain the start address of this TBR relative vector table. | |
2314 | In the startup routine of the user application, user needs to care of this | |
2315 | TBR register initialization. The TBR relative vector table can have at | |
2316 | max 256 function entries. The jumps to these functions will be generated | |
2317 | using a SH2A specific, non delayed branch instruction JSR/N @@(disp8,TBR). | |
c8619b90 NS |
2318 | You must use GAS and GLD from GNU binutils version 2.7 or later for |
2319 | this attribute to work correctly. | |
b34c7881 | 2320 | |
561642fa AP |
2321 | Please refer the example of M16C target, to see the use of this |
2322 | attribute while declaring a function, | |
2323 | ||
2324 | In an application, for a function being called once, this attribute will | |
2325 | save at least 8 bytes of code; and if other successive calls are being | |
2326 | made to the same function, it will save 2 bytes of code per each of these | |
2327 | calls. | |
2328 | ||
5abd2125 JS |
2329 | On M16C/M32C targets, the @code{function_vector} attribute declares a |
2330 | special page subroutine call function. Use of this attribute reduces | |
2331 | the code size by 2 bytes for each call generated to the | |
2332 | subroutine. The argument to the attribute is the vector number entry | |
2333 | from the special page vector table which contains the 16 low-order | |
2334 | bits of the subroutine's entry address. Each vector table has special | |
2335 | page number (18 to 255) which are used in @code{jsrs} instruction. | |
2336 | Jump addresses of the routines are generated by adding 0x0F0000 (in | |
2337 | case of M16C targets) or 0xFF0000 (in case of M32C targets), to the 2 | |
2338 | byte addresses set in the vector table. Therefore you need to ensure | |
2339 | that all the special page vector routines should get mapped within the | |
2340 | address range 0x0F0000 to 0x0FFFFF (for M16C) and 0xFF0000 to 0xFFFFFF | |
2341 | (for M32C). | |
2342 | ||
2343 | In the following example 2 bytes will be saved for each call to | |
2344 | function @code{foo}. | |
2345 | ||
2346 | @smallexample | |
2347 | void foo (void) __attribute__((function_vector(0x18))); | |
2348 | void foo (void) | |
2349 | @{ | |
2350 | @} | |
2351 | ||
2352 | void bar (void) | |
2353 | @{ | |
2354 | foo(); | |
2355 | @} | |
2356 | @end smallexample | |
2357 | ||
2358 | If functions are defined in one file and are called in another file, | |
2359 | then be sure to write this declaration in both files. | |
2360 | ||
2361 | This attribute is ignored for R8C target. | |
2362 | ||
c8619b90 NS |
2363 | @item interrupt |
2364 | @cindex interrupt handler functions | |
cd985f66 | 2365 | Use this attribute on the ARM, AVR, CRX, M32C, M32R/D, m68k, |
2bccb817 KH |
2366 | and Xstormy16 ports to indicate that the specified function is an |
2367 | interrupt handler. The compiler will generate function entry and exit | |
2368 | sequences suitable for use in an interrupt handler when this attribute | |
2369 | is present. | |
b34c7881 | 2370 | |
2bccb817 | 2371 | Note, interrupt handlers for the Blackfin, H8/300, H8/300H, H8S, and |
0d4a78eb | 2372 | SH processors can be specified via the @code{interrupt_handler} attribute. |
b34c7881 | 2373 | |
c8619b90 | 2374 | Note, on the AVR, interrupts will be enabled inside the function. |
9162542e | 2375 | |
c8619b90 NS |
2376 | Note, for the ARM, you can specify the kind of interrupt to be handled by |
2377 | adding an optional parameter to the interrupt attribute like this: | |
e23bd218 IR |
2378 | |
2379 | @smallexample | |
c8619b90 | 2380 | void f () __attribute__ ((interrupt ("IRQ"))); |
e23bd218 IR |
2381 | @end smallexample |
2382 | ||
c8619b90 | 2383 | Permissible values for this parameter are: IRQ, FIQ, SWI, ABORT and UNDEF@. |
e23bd218 | 2384 | |
7a085dce | 2385 | On ARMv7-M the interrupt type is ignored, and the attribute means the function |
5b3e6663 PB |
2386 | may be called with a word aligned stack pointer. |
2387 | ||
c8619b90 | 2388 | @item interrupt_handler |
0d4a78eb BS |
2389 | @cindex interrupt handler functions on the Blackfin, m68k, H8/300 and SH processors |
2390 | Use this attribute on the Blackfin, m68k, H8/300, H8/300H, H8S, and SH to | |
2391 | indicate that the specified function is an interrupt handler. The compiler | |
2392 | will generate function entry and exit sequences suitable for use in an | |
2393 | interrupt handler when this attribute is present. | |
2394 | ||
a4242737 KH |
2395 | @item interrupt_thread |
2396 | @cindex interrupt thread functions on fido | |
2397 | Use this attribute on fido, a subarchitecture of the m68k, to indicate | |
2398 | that the specified function is an interrupt handler that is designed | |
2399 | to run as a thread. The compiler omits generate prologue/epilogue | |
2400 | sequences and replaces the return instruction with a @code{sleep} | |
2401 | instruction. This attribute is available only on fido. | |
2402 | ||
0d4a78eb BS |
2403 | @item kspisusp |
2404 | @cindex User stack pointer in interrupts on the Blackfin | |
2405 | When used together with @code{interrupt_handler}, @code{exception_handler} | |
2406 | or @code{nmi_handler}, code will be generated to load the stack pointer | |
2407 | from the USP register in the function prologue. | |
72954a4f | 2408 | |
4af797b5 JZ |
2409 | @item l1_text |
2410 | @cindex @code{l1_text} function attribute | |
2411 | This attribute specifies a function to be placed into L1 Instruction | |
0ee2ea09 | 2412 | SRAM@. The function will be put into a specific section named @code{.l1.text}. |
4af797b5 JZ |
2413 | With @option{-mfdpic}, function calls with a such function as the callee |
2414 | or caller will use inlined PLT. | |
2415 | ||
c8619b90 NS |
2416 | @item long_call/short_call |
2417 | @cindex indirect calls on ARM | |
2418 | This attribute specifies how a particular function is called on | |
2419 | ARM@. Both attributes override the @option{-mlong-calls} (@pxref{ARM Options}) | |
2420 | command line switch and @code{#pragma long_calls} settings. The | |
87c365a4 NS |
2421 | @code{long_call} attribute indicates that the function might be far |
2422 | away from the call site and require a different (more expensive) | |
2423 | calling sequence. The @code{short_call} attribute always places | |
c8619b90 NS |
2424 | the offset to the function from the call site into the @samp{BL} |
2425 | instruction directly. | |
72954a4f | 2426 | |
c8619b90 NS |
2427 | @item longcall/shortcall |
2428 | @cindex functions called via pointer on the RS/6000 and PowerPC | |
87c365a4 NS |
2429 | On the Blackfin, RS/6000 and PowerPC, the @code{longcall} attribute |
2430 | indicates that the function might be far away from the call site and | |
2431 | require a different (more expensive) calling sequence. The | |
2432 | @code{shortcall} attribute indicates that the function is always close | |
2433 | enough for the shorter calling sequence to be used. These attributes | |
2434 | override both the @option{-mlongcall} switch and, on the RS/6000 and | |
2435 | PowerPC, the @code{#pragma longcall} setting. | |
72954a4f | 2436 | |
c8619b90 NS |
2437 | @xref{RS/6000 and PowerPC Options}, for more information on whether long |
2438 | calls are necessary. | |
c1f7febf | 2439 | |
cd3a59b3 | 2440 | @item long_call/near/far |
4dbdb061 | 2441 | @cindex indirect calls on MIPS |
cd3a59b3 SL |
2442 | These attributes specify how a particular function is called on MIPS@. |
2443 | The attributes override the @option{-mlong-calls} (@pxref{MIPS Options}) | |
2444 | command-line switch. The @code{long_call} and @code{far} attributes are | |
2445 | synonyms, and cause the compiler to always call | |
4dbdb061 | 2446 | the function by first loading its address into a register, and then using |
cd3a59b3 SL |
2447 | the contents of that register. The @code{near} attribute has the opposite |
2448 | effect; it specifies that non-PIC calls should be made using the more | |
2449 | efficient @code{jal} instruction. | |
4dbdb061 | 2450 | |
140592a0 AG |
2451 | @item malloc |
2452 | @cindex @code{malloc} attribute | |
2453 | The @code{malloc} attribute is used to tell the compiler that a function | |
928a5ba9 JM |
2454 | may be treated as if any non-@code{NULL} pointer it returns cannot |
2455 | alias any other pointer valid when the function returns. | |
140592a0 | 2456 | This will often improve optimization. |
928a5ba9 JM |
2457 | Standard functions with this property include @code{malloc} and |
2458 | @code{calloc}. @code{realloc}-like functions have this property as | |
2459 | long as the old pointer is never referred to (including comparing it | |
2460 | to the new pointer) after the function returns a non-@code{NULL} | |
2461 | value. | |
140592a0 | 2462 | |
f9e4a411 SL |
2463 | @item mips16/nomips16 |
2464 | @cindex @code{mips16} attribute | |
2465 | @cindex @code{nomips16} attribute | |
2466 | ||
2467 | On MIPS targets, you can use the @code{mips16} and @code{nomips16} | |
2468 | function attributes to locally select or turn off MIPS16 code generation. | |
2469 | A function with the @code{mips16} attribute is emitted as MIPS16 code, | |
2470 | while MIPS16 code generation is disabled for functions with the | |
2471 | @code{nomips16} attribute. These attributes override the | |
2472 | @option{-mips16} and @option{-mno-mips16} options on the command line | |
2473 | (@pxref{MIPS Options}). | |
2474 | ||
2475 | When compiling files containing mixed MIPS16 and non-MIPS16 code, the | |
2476 | preprocessor symbol @code{__mips16} reflects the setting on the command line, | |
2477 | not that within individual functions. Mixed MIPS16 and non-MIPS16 code | |
2478 | may interact badly with some GCC extensions such as @code{__builtin_apply} | |
2479 | (@pxref{Constructing Calls}). | |
2480 | ||
c8619b90 NS |
2481 | @item model (@var{model-name}) |
2482 | @cindex function addressability on the M32R/D | |
2483 | @cindex variable addressability on the IA-64 | |
2484 | ||
2485 | On the M32R/D, use this attribute to set the addressability of an | |
2486 | object, and of the code generated for a function. The identifier | |
2487 | @var{model-name} is one of @code{small}, @code{medium}, or | |
2488 | @code{large}, representing each of the code models. | |
2489 | ||
2490 | Small model objects live in the lower 16MB of memory (so that their | |
2491 | addresses can be loaded with the @code{ld24} instruction), and are | |
2492 | callable with the @code{bl} instruction. | |
2493 | ||
2494 | Medium model objects may live anywhere in the 32-bit address space (the | |
2495 | compiler will generate @code{seth/add3} instructions to load their addresses), | |
2496 | and are callable with the @code{bl} instruction. | |
2497 | ||
2498 | Large model objects may live anywhere in the 32-bit address space (the | |
2499 | compiler will generate @code{seth/add3} instructions to load their addresses), | |
2500 | and may not be reachable with the @code{bl} instruction (the compiler will | |
2501 | generate the much slower @code{seth/add3/jl} instruction sequence). | |
2502 | ||
2503 | On IA-64, use this attribute to set the addressability of an object. | |
2504 | At present, the only supported identifier for @var{model-name} is | |
2505 | @code{small}, indicating addressability via ``small'' (22-bit) | |
2506 | addresses (so that their addresses can be loaded with the @code{addl} | |
2507 | instruction). Caveat: such addressing is by definition not position | |
2508 | independent and hence this attribute must not be used for objects | |
2509 | defined by shared libraries. | |
2510 | ||
7c800926 | 2511 | @item ms_abi/sysv_abi |
f9ac6b1e | 2512 | @cindex @code{ms_abi} attribute |
7c800926 KT |
2513 | @cindex @code{sysv_abi} attribute |
2514 | ||
2515 | On 64-bit x86_65-*-* targets, you can use an ABI attribute to indicate | |
2516 | which calling convention should be used for a function. The @code{ms_abi} | |
2517 | attribute tells the compiler to use the Microsoft ABI, while the | |
2518 | @code{sysv_abi} attribute tells the compiler to use the ABI used on | |
2519 | GNU/Linux and other systems. The default is to use the Microsoft ABI | |
2520 | when targeting Windows. On all other systems, the default is the AMD ABI. | |
2521 | ||
2522 | Note, This feature is currently sorried out for Windows targets trying to | |
2523 | ||
c8619b90 NS |
2524 | @item naked |
2525 | @cindex function without a prologue/epilogue code | |
8d8da227 | 2526 | Use this attribute on the ARM, AVR, IP2K and SPU ports to indicate that |
85d9c13c | 2527 | the specified function does not need prologue/epilogue sequences generated by |
007e61c2 PB |
2528 | the compiler. It is up to the programmer to provide these sequences. The |
2529 | only statements that can be safely included in naked functions are | |
2530 | @code{asm} statements that do not have operands. All other statements, | |
2531 | including declarations of local variables, @code{if} statements, and so | |
2532 | forth, should be avoided. Naked functions should be used to implement the | |
2533 | body of an assembly function, while allowing the compiler to construct | |
2534 | the requisite function declaration for the assembler. | |
c8619b90 NS |
2535 | |
2536 | @item near | |
2537 | @cindex functions which do not handle memory bank switching on 68HC11/68HC12 | |
2538 | On 68HC11 and 68HC12 the @code{near} attribute causes the compiler to | |
2539 | use the normal calling convention based on @code{jsr} and @code{rts}. | |
2540 | This attribute can be used to cancel the effect of the @option{-mlong-calls} | |
2541 | option. | |
2542 | ||
0d4a78eb BS |
2543 | @item nesting |
2544 | @cindex Allow nesting in an interrupt handler on the Blackfin processor. | |
2545 | Use this attribute together with @code{interrupt_handler}, | |
2546 | @code{exception_handler} or @code{nmi_handler} to indicate that the function | |
2547 | entry code should enable nested interrupts or exceptions. | |
2548 | ||
2549 | @item nmi_handler | |
2550 | @cindex NMI handler functions on the Blackfin processor | |
2551 | Use this attribute on the Blackfin to indicate that the specified function | |
2552 | is an NMI handler. The compiler will generate function entry and | |
2553 | exit sequences suitable for use in an NMI handler when this | |
2554 | attribute is present. | |
2555 | ||
c8619b90 NS |
2556 | @item no_instrument_function |
2557 | @cindex @code{no_instrument_function} function attribute | |
2558 | @opindex finstrument-functions | |
2559 | If @option{-finstrument-functions} is given, profiling function calls will | |
2560 | be generated at entry and exit of most user-compiled functions. | |
2561 | Functions with this attribute will not be so instrumented. | |
2562 | ||
2563 | @item noinline | |
2564 | @cindex @code{noinline} function attribute | |
2565 | This function attribute prevents a function from being considered for | |
2566 | inlining. | |
ccd2a21e HPN |
2567 | @c Don't enumerate the optimizations by name here; we try to be |
2568 | @c future-compatible with this mechanism. | |
2569 | If the function does not have side-effects, there are optimizations | |
2570 | other than inlining that causes function calls to be optimized away, | |
2571 | although the function call is live. To keep such calls from being | |
2572 | optimized away, put | |
2573 | @smallexample | |
2574 | asm (""); | |
2575 | @end smallexample | |
2576 | (@pxref{Extended Asm}) in the called function, to serve as a special | |
2577 | side-effect. | |
c8619b90 NS |
2578 | |
2579 | @item nonnull (@var{arg-index}, @dots{}) | |
2580 | @cindex @code{nonnull} function attribute | |
2581 | The @code{nonnull} attribute specifies that some function parameters should | |
2582 | be non-null pointers. For instance, the declaration: | |
c1f7febf RK |
2583 | |
2584 | @smallexample | |
c8619b90 NS |
2585 | extern void * |
2586 | my_memcpy (void *dest, const void *src, size_t len) | |
6ccde948 | 2587 | __attribute__((nonnull (1, 2))); |
c1f7febf RK |
2588 | @end smallexample |
2589 | ||
c8619b90 NS |
2590 | @noindent |
2591 | causes the compiler to check that, in calls to @code{my_memcpy}, | |
2592 | arguments @var{dest} and @var{src} are non-null. If the compiler | |
2593 | determines that a null pointer is passed in an argument slot marked | |
2594 | as non-null, and the @option{-Wnonnull} option is enabled, a warning | |
2595 | is issued. The compiler may also choose to make optimizations based | |
2596 | on the knowledge that certain function arguments will not be null. | |
af3e86c2 | 2597 | |
c8619b90 NS |
2598 | If no argument index list is given to the @code{nonnull} attribute, |
2599 | all pointer arguments are marked as non-null. To illustrate, the | |
2600 | following declaration is equivalent to the previous example: | |
47bd70b5 JJ |
2601 | |
2602 | @smallexample | |
c8619b90 NS |
2603 | extern void * |
2604 | my_memcpy (void *dest, const void *src, size_t len) | |
6ccde948 | 2605 | __attribute__((nonnull)); |
47bd70b5 JJ |
2606 | @end smallexample |
2607 | ||
c8619b90 NS |
2608 | @item noreturn |
2609 | @cindex @code{noreturn} function attribute | |
2610 | A few standard library functions, such as @code{abort} and @code{exit}, | |
2611 | cannot return. GCC knows this automatically. Some programs define | |
2612 | their own functions that never return. You can declare them | |
2613 | @code{noreturn} to tell the compiler this fact. For example, | |
9e8aab55 | 2614 | |
c8619b90 NS |
2615 | @smallexample |
2616 | @group | |
2617 | void fatal () __attribute__ ((noreturn)); | |
d5c4db17 | 2618 | |
c8619b90 NS |
2619 | void |
2620 | fatal (/* @r{@dots{}} */) | |
2621 | @{ | |
2622 | /* @r{@dots{}} */ /* @r{Print error message.} */ /* @r{@dots{}} */ | |
2623 | exit (1); | |
2624 | @} | |
2625 | @end group | |
2626 | @end smallexample | |
9e8aab55 | 2627 | |
c8619b90 NS |
2628 | The @code{noreturn} keyword tells the compiler to assume that |
2629 | @code{fatal} cannot return. It can then optimize without regard to what | |
2630 | would happen if @code{fatal} ever did return. This makes slightly | |
2631 | better code. More importantly, it helps avoid spurious warnings of | |
2632 | uninitialized variables. | |
9e8aab55 | 2633 | |
c8619b90 NS |
2634 | The @code{noreturn} keyword does not affect the exceptional path when that |
2635 | applies: a @code{noreturn}-marked function may still return to the caller | |
2e9522f4 | 2636 | by throwing an exception or calling @code{longjmp}. |
9e8aab55 | 2637 | |
c8619b90 NS |
2638 | Do not assume that registers saved by the calling function are |
2639 | restored before calling the @code{noreturn} function. | |
47bd70b5 | 2640 | |
c8619b90 NS |
2641 | It does not make sense for a @code{noreturn} function to have a return |
2642 | type other than @code{void}. | |
c1f7febf | 2643 | |
c8619b90 NS |
2644 | The attribute @code{noreturn} is not implemented in GCC versions |
2645 | earlier than 2.5. An alternative way to declare that a function does | |
2646 | not return, which works in the current version and in some older | |
2647 | versions, is as follows: | |
5d34c8e9 | 2648 | |
c8619b90 NS |
2649 | @smallexample |
2650 | typedef void voidfn (); | |
c1f7febf | 2651 | |
c8619b90 NS |
2652 | volatile voidfn fatal; |
2653 | @end smallexample | |
e91f04de | 2654 | |
a1e73046 PC |
2655 | This approach does not work in GNU C++. |
2656 | ||
c8619b90 NS |
2657 | @item nothrow |
2658 | @cindex @code{nothrow} function attribute | |
2659 | The @code{nothrow} attribute is used to inform the compiler that a | |
2660 | function cannot throw an exception. For example, most functions in | |
2661 | the standard C library can be guaranteed not to throw an exception | |
2662 | with the notable exceptions of @code{qsort} and @code{bsearch} that | |
2663 | take function pointer arguments. The @code{nothrow} attribute is not | |
3f3174b6 | 2664 | implemented in GCC versions earlier than 3.3. |
c1f7febf | 2665 | |
c8619b90 NS |
2666 | @item pure |
2667 | @cindex @code{pure} function attribute | |
2668 | Many functions have no effects except the return value and their | |
2669 | return value depends only on the parameters and/or global variables. | |
2670 | Such a function can be subject | |
2671 | to common subexpression elimination and loop optimization just as an | |
2672 | arithmetic operator would be. These functions should be declared | |
2673 | with the attribute @code{pure}. For example, | |
a5c76ee6 | 2674 | |
c8619b90 NS |
2675 | @smallexample |
2676 | int square (int) __attribute__ ((pure)); | |
2677 | @end smallexample | |
c1f7febf | 2678 | |
c8619b90 NS |
2679 | @noindent |
2680 | says that the hypothetical function @code{square} is safe to call | |
2681 | fewer times than the program says. | |
c27ba912 | 2682 | |
c8619b90 NS |
2683 | Some of common examples of pure functions are @code{strlen} or @code{memcmp}. |
2684 | Interesting non-pure functions are functions with infinite loops or those | |
2685 | depending on volatile memory or other system resource, that may change between | |
2686 | two consecutive calls (such as @code{feof} in a multithreading environment). | |
c1f7febf | 2687 | |
c8619b90 NS |
2688 | The attribute @code{pure} is not implemented in GCC versions earlier |
2689 | than 2.96. | |
c1f7febf | 2690 | |
52bf96d2 JH |
2691 | @item hot |
2692 | @cindex @code{hot} function attribute | |
2693 | The @code{hot} attribute is used to inform the compiler that a function is a | |
2694 | hot spot of the compiled program. The function is optimized more aggressively | |
2695 | and on many target it is placed into special subsection of the text section so | |
2696 | all hot functions appears close together improving locality. | |
2697 | ||
2698 | When profile feedback is available, via @option{-fprofile-use}, hot functions | |
2699 | are automatically detected and this attribute is ignored. | |
2700 | ||
2701 | The @code{hot} attribute is not implemented in GCC versions earlier than 4.3. | |
2702 | ||
2703 | @item cold | |
2704 | @cindex @code{cold} function attribute | |
2705 | The @code{cold} attribute is used to inform the compiler that a function is | |
2706 | unlikely executed. The function is optimized for size rather than speed and on | |
2707 | many targets it is placed into special subsection of the text section so all | |
2708 | cold functions appears close together improving code locality of non-cold parts | |
2709 | of program. The paths leading to call of cold functions within code are marked | |
44c7bd63 | 2710 | as unlikely by the branch prediction mechanism. It is thus useful to mark |
52bf96d2 JH |
2711 | functions used to handle unlikely conditions, such as @code{perror}, as cold to |
2712 | improve optimization of hot functions that do call marked functions in rare | |
2713 | occasions. | |
2714 | ||
2715 | When profile feedback is available, via @option{-fprofile-use}, hot functions | |
2716 | are automatically detected and this attribute is ignored. | |
2717 | ||
2718 | The @code{hot} attribute is not implemented in GCC versions earlier than 4.3. | |
2719 | ||
c8619b90 NS |
2720 | @item regparm (@var{number}) |
2721 | @cindex @code{regparm} attribute | |
2722 | @cindex functions that are passed arguments in registers on the 386 | |
2723 | On the Intel 386, the @code{regparm} attribute causes the compiler to | |
2f84b963 RG |
2724 | pass arguments number one to @var{number} if they are of integral type |
2725 | in registers EAX, EDX, and ECX instead of on the stack. Functions that | |
2726 | take a variable number of arguments will continue to be passed all of their | |
c8619b90 | 2727 | arguments on the stack. |
6d3d9133 | 2728 | |
c8619b90 NS |
2729 | Beware that on some ELF systems this attribute is unsuitable for |
2730 | global functions in shared libraries with lazy binding (which is the | |
2731 | default). Lazy binding will send the first call via resolving code in | |
2732 | the loader, which might assume EAX, EDX and ECX can be clobbered, as | |
2733 | per the standard calling conventions. Solaris 8 is affected by this. | |
2734 | GNU systems with GLIBC 2.1 or higher, and FreeBSD, are believed to be | |
2735 | safe since the loaders there save all registers. (Lazy binding can be | |
2736 | disabled with the linker or the loader if desired, to avoid the | |
2737 | problem.) | |
6d3d9133 | 2738 | |
2f84b963 RG |
2739 | @item sseregparm |
2740 | @cindex @code{sseregparm} attribute | |
2741 | On the Intel 386 with SSE support, the @code{sseregparm} attribute | |
56829cae | 2742 | causes the compiler to pass up to 3 floating point arguments in |
2f84b963 RG |
2743 | SSE registers instead of on the stack. Functions that take a |
2744 | variable number of arguments will continue to pass all of their | |
2745 | floating point arguments on the stack. | |
2746 | ||
33932946 SH |
2747 | @item force_align_arg_pointer |
2748 | @cindex @code{force_align_arg_pointer} attribute | |
2749 | On the Intel x86, the @code{force_align_arg_pointer} attribute may be | |
2750 | applied to individual function definitions, generating an alternate | |
2751 | prologue and epilogue that realigns the runtime stack. This supports | |
2752 | mixing legacy codes that run with a 4-byte aligned stack with modern | |
2753 | codes that keep a 16-byte stack for SSE compatibility. The alternate | |
2754 | prologue and epilogue are slower and bigger than the regular ones, and | |
2755 | the alternate prologue requires a scratch register; this lowers the | |
2756 | number of registers available if used in conjunction with the | |
2757 | @code{regparm} attribute. The @code{force_align_arg_pointer} | |
2758 | attribute is incompatible with nested functions; this is considered a | |
2759 | hard error. | |
2760 | ||
561642fa AP |
2761 | @item resbank |
2762 | @cindex @code{resbank} attribute | |
2763 | On the SH2A target, this attribute enables the high-speed register | |
2764 | saving and restoration using a register bank for @code{interrupt_handler} | |
2765 | routines. Saving to the bank is performed automatcially after the CPU | |
2766 | accepts an interrupt that uses a register bank. | |
2767 | ||
2768 | The nineteen 32-bit registers comprising general register R0 to R14, | |
2769 | control register GBR, and system registers MACH, MACL, and PR and the | |
2770 | vector table address offset are saved into a register bank. Register | |
2771 | banks are stacked in first-in last-out (FILO) sequence. Restoration | |
2772 | from the bank is executed by issuing a RESBANK instruction. | |
2773 | ||
6e9a3221 AN |
2774 | @item returns_twice |
2775 | @cindex @code{returns_twice} attribute | |
2776 | The @code{returns_twice} attribute tells the compiler that a function may | |
2777 | return more than one time. The compiler will ensure that all registers | |
2778 | are dead before calling such a function and will emit a warning about | |
2779 | the variables that may be clobbered after the second return from the | |
2780 | function. Examples of such functions are @code{setjmp} and @code{vfork}. | |
2781 | The @code{longjmp}-like counterpart of such function, if any, might need | |
2782 | to be marked with the @code{noreturn} attribute. | |
2783 | ||
c8619b90 | 2784 | @item saveall |
0d4a78eb BS |
2785 | @cindex save all registers on the Blackfin, H8/300, H8/300H, and H8S |
2786 | Use this attribute on the Blackfin, H8/300, H8/300H, and H8S to indicate that | |
c8619b90 NS |
2787 | all registers except the stack pointer should be saved in the prologue |
2788 | regardless of whether they are used or not. | |
6d3d9133 | 2789 | |
c8619b90 NS |
2790 | @item section ("@var{section-name}") |
2791 | @cindex @code{section} function attribute | |
2792 | Normally, the compiler places the code it generates in the @code{text} section. | |
2793 | Sometimes, however, you need additional sections, or you need certain | |
2794 | particular functions to appear in special sections. The @code{section} | |
2795 | attribute specifies that a function lives in a particular section. | |
2796 | For example, the declaration: | |
6d3d9133 NC |
2797 | |
2798 | @smallexample | |
c8619b90 | 2799 | extern void foobar (void) __attribute__ ((section ("bar"))); |
6d3d9133 NC |
2800 | @end smallexample |
2801 | ||
c8619b90 NS |
2802 | @noindent |
2803 | puts the function @code{foobar} in the @code{bar} section. | |
6d3d9133 | 2804 | |
c8619b90 NS |
2805 | Some file formats do not support arbitrary sections so the @code{section} |
2806 | attribute is not available on all platforms. | |
2807 | If you need to map the entire contents of a module to a particular | |
2808 | section, consider using the facilities of the linker instead. | |
2809 | ||
3d091dac KG |
2810 | @item sentinel |
2811 | @cindex @code{sentinel} function attribute | |
254986c7 KG |
2812 | This function attribute ensures that a parameter in a function call is |
2813 | an explicit @code{NULL}. The attribute is only valid on variadic | |
2814 | functions. By default, the sentinel is located at position zero, the | |
2815 | last parameter of the function call. If an optional integer position | |
2816 | argument P is supplied to the attribute, the sentinel must be located at | |
2817 | position P counting backwards from the end of the argument list. | |
2818 | ||
2819 | @smallexample | |
2820 | __attribute__ ((sentinel)) | |
2821 | is equivalent to | |
2822 | __attribute__ ((sentinel(0))) | |
2823 | @end smallexample | |
2824 | ||
2825 | The attribute is automatically set with a position of 0 for the built-in | |
2826 | functions @code{execl} and @code{execlp}. The built-in function | |
254ea84c | 2827 | @code{execle} has the attribute set with a position of 1. |
254986c7 KG |
2828 | |
2829 | A valid @code{NULL} in this context is defined as zero with any pointer | |
2830 | type. If your system defines the @code{NULL} macro with an integer type | |
2831 | then you need to add an explicit cast. GCC replaces @code{stddef.h} | |
2832 | with a copy that redefines NULL appropriately. | |
2833 | ||
2834 | The warnings for missing or incorrect sentinels are enabled with | |
2835 | @option{-Wformat}. | |
3d091dac | 2836 | |
c8619b90 NS |
2837 | @item short_call |
2838 | See long_call/short_call. | |
2839 | ||
2840 | @item shortcall | |
2841 | See longcall/shortcall. | |
2842 | ||
2843 | @item signal | |
2844 | @cindex signal handler functions on the AVR processors | |
2845 | Use this attribute on the AVR to indicate that the specified | |
2846 | function is a signal handler. The compiler will generate function | |
2847 | entry and exit sequences suitable for use in a signal handler when this | |
2848 | attribute is present. Interrupts will be disabled inside the function. | |
b93e3893 AO |
2849 | |
2850 | @item sp_switch | |
88ab0d1c | 2851 | Use this attribute on the SH to indicate an @code{interrupt_handler} |
b93e3893 AO |
2852 | function should switch to an alternate stack. It expects a string |
2853 | argument that names a global variable holding the address of the | |
2854 | alternate stack. | |
2855 | ||
2856 | @smallexample | |
2857 | void *alt_stack; | |
aee96fe9 JM |
2858 | void f () __attribute__ ((interrupt_handler, |
2859 | sp_switch ("alt_stack"))); | |
b93e3893 AO |
2860 | @end smallexample |
2861 | ||
c8619b90 NS |
2862 | @item stdcall |
2863 | @cindex functions that pop the argument stack on the 386 | |
2864 | On the Intel 386, the @code{stdcall} attribute causes the compiler to | |
2865 | assume that the called function will pop off the stack space used to | |
2866 | pass arguments, unless it takes a variable number of arguments. | |
c1f7febf RK |
2867 | |
2868 | @item tiny_data | |
dbacaa98 KH |
2869 | @cindex tiny data section on the H8/300H and H8S |
2870 | Use this attribute on the H8/300H and H8S to indicate that the specified | |
c1f7febf RK |
2871 | variable should be placed into the tiny data section. |
2872 | The compiler will generate more efficient code for loads and stores | |
2873 | on data in the tiny data section. Note the tiny data area is limited to | |
2874 | slightly under 32kbytes of data. | |
845da534 | 2875 | |
c8619b90 NS |
2876 | @item trap_exit |
2877 | Use this attribute on the SH for an @code{interrupt_handler} to return using | |
2878 | @code{trapa} instead of @code{rte}. This attribute expects an integer | |
2879 | argument specifying the trap number to be used. | |
845da534 | 2880 | |
c8619b90 NS |
2881 | @item unused |
2882 | @cindex @code{unused} attribute. | |
2883 | This attribute, attached to a function, means that the function is meant | |
2884 | to be possibly unused. GCC will not produce a warning for this | |
2885 | function. | |
a32767e4 | 2886 | |
c8619b90 NS |
2887 | @item used |
2888 | @cindex @code{used} attribute. | |
2889 | This attribute, attached to a function, means that code must be emitted | |
2890 | for the function even if it appears that the function is not referenced. | |
2891 | This is useful, for example, when the function is referenced only in | |
2892 | inline assembly. | |
5936c7e7 | 2893 | |
812b587e SE |
2894 | @item version_id |
2895 | @cindex @code{version_id} attribute on IA64 HP-UX | |
2896 | This attribute, attached to a global variable or function, renames a | |
2897 | symbol to contain a version string, thus allowing for function level | |
2898 | versioning. HP-UX system header files may use version level functioning | |
2899 | for some system calls. | |
2900 | ||
2901 | @smallexample | |
2902 | extern int foo () __attribute__((version_id ("20040821"))); | |
2903 | @end smallexample | |
2904 | ||
2905 | Calls to @var{foo} will be mapped to calls to @var{foo@{20040821@}}. | |
2906 | ||
c8619b90 NS |
2907 | @item visibility ("@var{visibility_type}") |
2908 | @cindex @code{visibility} attribute | |
46bdbc00 GK |
2909 | This attribute affects the linkage of the declaration to which it is attached. |
2910 | There are four supported @var{visibility_type} values: default, | |
2911 | hidden, protected or internal visibility. | |
5936c7e7 | 2912 | |
c8619b90 NS |
2913 | @smallexample |
2914 | void __attribute__ ((visibility ("protected"))) | |
2915 | f () @{ /* @r{Do something.} */; @} | |
2916 | int i __attribute__ ((visibility ("hidden"))); | |
2917 | @end smallexample | |
5936c7e7 | 2918 | |
46bdbc00 GK |
2919 | The possible values of @var{visibility_type} correspond to the |
2920 | visibility settings in the ELF gABI. | |
5936c7e7 | 2921 | |
c8619b90 | 2922 | @table @dfn |
63c5b495 | 2923 | @c keep this list of visibilities in alphabetical order. |
6b6cb52e | 2924 | |
c8619b90 | 2925 | @item default |
46bdbc00 GK |
2926 | Default visibility is the normal case for the object file format. |
2927 | This value is available for the visibility attribute to override other | |
2928 | options that may change the assumed visibility of entities. | |
2929 | ||
2930 | On ELF, default visibility means that the declaration is visible to other | |
2931 | modules and, in shared libraries, means that the declared entity may be | |
2932 | overridden. | |
2933 | ||
2934 | On Darwin, default visibility means that the declaration is visible to | |
2935 | other modules. | |
2936 | ||
2937 | Default visibility corresponds to ``external linkage'' in the language. | |
6b6cb52e | 2938 | |
c8619b90 | 2939 | @item hidden |
46bdbc00 GK |
2940 | Hidden visibility indicates that the entity declared will have a new |
2941 | form of linkage, which we'll call ``hidden linkage''. Two | |
2942 | declarations of an object with hidden linkage refer to the same object | |
2943 | if they are in the same shared object. | |
6b6cb52e | 2944 | |
c8619b90 NS |
2945 | @item internal |
2946 | Internal visibility is like hidden visibility, but with additional | |
46bdbc00 GK |
2947 | processor specific semantics. Unless otherwise specified by the |
2948 | psABI, GCC defines internal visibility to mean that a function is | |
2949 | @emph{never} called from another module. Compare this with hidden | |
2950 | functions which, while they cannot be referenced directly by other | |
2951 | modules, can be referenced indirectly via function pointers. By | |
2952 | indicating that a function cannot be called from outside the module, | |
2953 | GCC may for instance omit the load of a PIC register since it is known | |
2954 | that the calling function loaded the correct value. | |
6b6cb52e | 2955 | |
c8619b90 | 2956 | @item protected |
46bdbc00 GK |
2957 | Protected visibility is like default visibility except that it |
2958 | indicates that references within the defining module will bind to the | |
2959 | definition in that module. That is, the declared entity cannot be | |
2960 | overridden by another module. | |
6b6cb52e | 2961 | |
c8619b90 | 2962 | @end table |
6b6cb52e | 2963 | |
46bdbc00 GK |
2964 | All visibilities are supported on many, but not all, ELF targets |
2965 | (supported when the assembler supports the @samp{.visibility} | |
2966 | pseudo-op). Default visibility is supported everywhere. Hidden | |
2967 | visibility is supported on Darwin targets. | |
2968 | ||
2969 | The visibility attribute should be applied only to declarations which | |
2970 | would otherwise have external linkage. The attribute should be applied | |
2971 | consistently, so that the same entity should not be declared with | |
2972 | different settings of the attribute. | |
2973 | ||
2974 | In C++, the visibility attribute applies to types as well as functions | |
b9e75696 JM |
2975 | and objects, because in C++ types have linkage. A class must not have |
2976 | greater visibility than its non-static data member types and bases, | |
2977 | and class members default to the visibility of their class. Also, a | |
b70f0f48 JM |
2978 | declaration without explicit visibility is limited to the visibility |
2979 | of its type. | |
46bdbc00 GK |
2980 | |
2981 | In C++, you can mark member functions and static member variables of a | |
2982 | class with the visibility attribute. This is useful if if you know a | |
2983 | particular method or static member variable should only be used from | |
2984 | one shared object; then you can mark it hidden while the rest of the | |
2985 | class has default visibility. Care must be taken to avoid breaking | |
b70f0f48 JM |
2986 | the One Definition Rule; for example, it is usually not useful to mark |
2987 | an inline method as hidden without marking the whole class as hidden. | |
6b6cb52e | 2988 | |
b9e75696 JM |
2989 | A C++ namespace declaration can also have the visibility attribute. |
2990 | This attribute applies only to the particular namespace body, not to | |
2991 | other definitions of the same namespace; it is equivalent to using | |
2992 | @samp{#pragma GCC visibility} before and after the namespace | |
2993 | definition (@pxref{Visibility Pragmas}). | |
2994 | ||
2995 | In C++, if a template argument has limited visibility, this | |
2996 | restriction is implicitly propagated to the template instantiation. | |
2997 | Otherwise, template instantiations and specializations default to the | |
2998 | visibility of their template. | |
2999 | ||
b70f0f48 JM |
3000 | If both the template and enclosing class have explicit visibility, the |
3001 | visibility from the template is used. | |
3002 | ||
c8619b90 NS |
3003 | @item warn_unused_result |
3004 | @cindex @code{warn_unused_result} attribute | |
3005 | The @code{warn_unused_result} attribute causes a warning to be emitted | |
3006 | if a caller of the function with this attribute does not use its | |
3007 | return value. This is useful for functions where not checking | |
3008 | the result is either a security problem or always a bug, such as | |
3009 | @code{realloc}. | |
6b6cb52e | 3010 | |
c8619b90 NS |
3011 | @smallexample |
3012 | int fn () __attribute__ ((warn_unused_result)); | |
3013 | int foo () | |
3014 | @{ | |
3015 | if (fn () < 0) return -1; | |
3016 | fn (); | |
3017 | return 0; | |
3018 | @} | |
3019 | @end smallexample | |
6b6cb52e | 3020 | |
c8619b90 | 3021 | results in warning on line 5. |
6b6cb52e | 3022 | |
c8619b90 NS |
3023 | @item weak |
3024 | @cindex @code{weak} attribute | |
3025 | The @code{weak} attribute causes the declaration to be emitted as a weak | |
3026 | symbol rather than a global. This is primarily useful in defining | |
3027 | library functions which can be overridden in user code, though it can | |
3028 | also be used with non-function declarations. Weak symbols are supported | |
3029 | for ELF targets, and also for a.out targets when using the GNU assembler | |
3030 | and linker. | |
6b6cb52e | 3031 | |
a0203ca7 AO |
3032 | @item weakref |
3033 | @itemx weakref ("@var{target}") | |
3034 | @cindex @code{weakref} attribute | |
3035 | The @code{weakref} attribute marks a declaration as a weak reference. | |
3036 | Without arguments, it should be accompanied by an @code{alias} attribute | |
3037 | naming the target symbol. Optionally, the @var{target} may be given as | |
3038 | an argument to @code{weakref} itself. In either case, @code{weakref} | |
3039 | implicitly marks the declaration as @code{weak}. Without a | |
3040 | @var{target}, given as an argument to @code{weakref} or to @code{alias}, | |
3041 | @code{weakref} is equivalent to @code{weak}. | |
3042 | ||
3043 | @smallexample | |
a9b0b825 | 3044 | static int x() __attribute__ ((weakref ("y"))); |
a0203ca7 | 3045 | /* is equivalent to... */ |
a9b0b825 | 3046 | static int x() __attribute__ ((weak, weakref, alias ("y"))); |
a0203ca7 | 3047 | /* and to... */ |
a9b0b825 GK |
3048 | static int x() __attribute__ ((weakref)); |
3049 | static int x() __attribute__ ((alias ("y"))); | |
a0203ca7 AO |
3050 | @end smallexample |
3051 | ||
3052 | A weak reference is an alias that does not by itself require a | |
3053 | definition to be given for the target symbol. If the target symbol is | |
3054 | only referenced through weak references, then the becomes a @code{weak} | |
3055 | undefined symbol. If it is directly referenced, however, then such | |
3056 | strong references prevail, and a definition will be required for the | |
3057 | symbol, not necessarily in the same translation unit. | |
3058 | ||
3059 | The effect is equivalent to moving all references to the alias to a | |
3060 | separate translation unit, renaming the alias to the aliased symbol, | |
3061 | declaring it as weak, compiling the two separate translation units and | |
3062 | performing a reloadable link on them. | |
3063 | ||
a9b0b825 GK |
3064 | At present, a declaration to which @code{weakref} is attached can |
3065 | only be @code{static}. | |
3066 | ||
ce91e74c JH |
3067 | @item externally_visible |
3068 | @cindex @code{externally_visible} attribute. | |
3069 | This attribute, attached to a global variable or function nullify | |
3070 | effect of @option{-fwhole-program} command line option, so the object | |
3071 | remain visible outside the current compilation unit | |
3072 | ||
c1f7febf RK |
3073 | @end table |
3074 | ||
3075 | You can specify multiple attributes in a declaration by separating them | |
3076 | by commas within the double parentheses or by immediately following an | |
3077 | attribute declaration with another attribute declaration. | |
3078 | ||
3079 | @cindex @code{#pragma}, reason for not using | |
3080 | @cindex pragma, reason for not using | |
9f1bbeaa JM |
3081 | Some people object to the @code{__attribute__} feature, suggesting that |
3082 | ISO C's @code{#pragma} should be used instead. At the time | |
3083 | @code{__attribute__} was designed, there were two reasons for not doing | |
3084 | this. | |
c1f7febf RK |
3085 | |
3086 | @enumerate | |
3087 | @item | |
3088 | It is impossible to generate @code{#pragma} commands from a macro. | |
3089 | ||
3090 | @item | |
3091 | There is no telling what the same @code{#pragma} might mean in another | |
3092 | compiler. | |
3093 | @end enumerate | |
3094 | ||
9f1bbeaa JM |
3095 | These two reasons applied to almost any application that might have been |
3096 | proposed for @code{#pragma}. It was basically a mistake to use | |
3097 | @code{#pragma} for @emph{anything}. | |
3098 | ||
3099 | The ISO C99 standard includes @code{_Pragma}, which now allows pragmas | |
3100 | to be generated from macros. In addition, a @code{#pragma GCC} | |
3101 | namespace is now in use for GCC-specific pragmas. However, it has been | |
3102 | found convenient to use @code{__attribute__} to achieve a natural | |
3103 | attachment of attributes to their corresponding declarations, whereas | |
3104 | @code{#pragma GCC} is of use for constructs that do not naturally form | |
3105 | part of the grammar. @xref{Other Directives,,Miscellaneous | |
48795525 | 3106 | Preprocessing Directives, cpp, The GNU C Preprocessor}. |
c1f7febf | 3107 | |
2c5e91d2 JM |
3108 | @node Attribute Syntax |
3109 | @section Attribute Syntax | |
3110 | @cindex attribute syntax | |
3111 | ||
3112 | This section describes the syntax with which @code{__attribute__} may be | |
3113 | used, and the constructs to which attribute specifiers bind, for the C | |
161d7b59 | 3114 | language. Some details may vary for C++ and Objective-C@. Because of |
2c5e91d2 JM |
3115 | infelicities in the grammar for attributes, some forms described here |
3116 | may not be successfully parsed in all cases. | |
3117 | ||
91d231cb JM |
3118 | There are some problems with the semantics of attributes in C++. For |
3119 | example, there are no manglings for attributes, although they may affect | |
3120 | code generation, so problems may arise when attributed types are used in | |
3121 | conjunction with templates or overloading. Similarly, @code{typeid} | |
3122 | does not distinguish between types with different attributes. Support | |
3123 | for attributes in C++ may be restricted in future to attributes on | |
3124 | declarations only, but not on nested declarators. | |
3125 | ||
2c5e91d2 JM |
3126 | @xref{Function Attributes}, for details of the semantics of attributes |
3127 | applying to functions. @xref{Variable Attributes}, for details of the | |
3128 | semantics of attributes applying to variables. @xref{Type Attributes}, | |
3129 | for details of the semantics of attributes applying to structure, union | |
3130 | and enumerated types. | |
3131 | ||
3132 | An @dfn{attribute specifier} is of the form | |
3133 | @code{__attribute__ ((@var{attribute-list}))}. An @dfn{attribute list} | |
3134 | is a possibly empty comma-separated sequence of @dfn{attributes}, where | |
3135 | each attribute is one of the following: | |
3136 | ||
3137 | @itemize @bullet | |
3138 | @item | |
3139 | Empty. Empty attributes are ignored. | |
3140 | ||
3141 | @item | |
3142 | A word (which may be an identifier such as @code{unused}, or a reserved | |
3143 | word such as @code{const}). | |
3144 | ||
3145 | @item | |
3146 | A word, followed by, in parentheses, parameters for the attribute. | |
3147 | These parameters take one of the following forms: | |
3148 | ||
3149 | @itemize @bullet | |
3150 | @item | |
3151 | An identifier. For example, @code{mode} attributes use this form. | |
3152 | ||
3153 | @item | |
3154 | An identifier followed by a comma and a non-empty comma-separated list | |
3155 | of expressions. For example, @code{format} attributes use this form. | |
3156 | ||
3157 | @item | |
3158 | A possibly empty comma-separated list of expressions. For example, | |
3159 | @code{format_arg} attributes use this form with the list being a single | |
3160 | integer constant expression, and @code{alias} attributes use this form | |
3161 | with the list being a single string constant. | |
3162 | @end itemize | |
3163 | @end itemize | |
3164 | ||
3165 | An @dfn{attribute specifier list} is a sequence of one or more attribute | |
3166 | specifiers, not separated by any other tokens. | |
3167 | ||
50fc59e7 | 3168 | In GNU C, an attribute specifier list may appear after the colon following a |
2c5e91d2 JM |
3169 | label, other than a @code{case} or @code{default} label. The only |
3170 | attribute it makes sense to use after a label is @code{unused}. This | |
3171 | feature is intended for code generated by programs which contains labels | |
3172 | that may be unused but which is compiled with @option{-Wall}. It would | |
3173 | not normally be appropriate to use in it human-written code, though it | |
3174 | could be useful in cases where the code that jumps to the label is | |
8a36672b | 3175 | contained within an @code{#ifdef} conditional. GNU C++ does not permit |
50fc59e7 NS |
3176 | such placement of attribute lists, as it is permissible for a |
3177 | declaration, which could begin with an attribute list, to be labelled in | |
8a36672b | 3178 | C++. Declarations cannot be labelled in C90 or C99, so the ambiguity |
50fc59e7 | 3179 | does not arise there. |
2c5e91d2 JM |
3180 | |
3181 | An attribute specifier list may appear as part of a @code{struct}, | |
3182 | @code{union} or @code{enum} specifier. It may go either immediately | |
3183 | after the @code{struct}, @code{union} or @code{enum} keyword, or after | |
b9e75696 | 3184 | the closing brace. The former syntax is preferred. |
2c5e91d2 JM |
3185 | Where attribute specifiers follow the closing brace, they are considered |
3186 | to relate to the structure, union or enumerated type defined, not to any | |
3187 | enclosing declaration the type specifier appears in, and the type | |
3188 | defined is not complete until after the attribute specifiers. | |
3189 | @c Otherwise, there would be the following problems: a shift/reduce | |
4fe9b91c | 3190 | @c conflict between attributes binding the struct/union/enum and |
2c5e91d2 JM |
3191 | @c binding to the list of specifiers/qualifiers; and "aligned" |
3192 | @c attributes could use sizeof for the structure, but the size could be | |
3193 | @c changed later by "packed" attributes. | |
3194 | ||
3195 | Otherwise, an attribute specifier appears as part of a declaration, | |
3196 | counting declarations of unnamed parameters and type names, and relates | |
3197 | to that declaration (which may be nested in another declaration, for | |
91d231cb JM |
3198 | example in the case of a parameter declaration), or to a particular declarator |
3199 | within a declaration. Where an | |
ff867905 JM |
3200 | attribute specifier is applied to a parameter declared as a function or |
3201 | an array, it should apply to the function or array rather than the | |
3202 | pointer to which the parameter is implicitly converted, but this is not | |
3203 | yet correctly implemented. | |
2c5e91d2 JM |
3204 | |
3205 | Any list of specifiers and qualifiers at the start of a declaration may | |
3206 | contain attribute specifiers, whether or not such a list may in that | |
3207 | context contain storage class specifiers. (Some attributes, however, | |
3208 | are essentially in the nature of storage class specifiers, and only make | |
3209 | sense where storage class specifiers may be used; for example, | |
3210 | @code{section}.) There is one necessary limitation to this syntax: the | |
3211 | first old-style parameter declaration in a function definition cannot | |
3212 | begin with an attribute specifier, because such an attribute applies to | |
3213 | the function instead by syntax described below (which, however, is not | |
3214 | yet implemented in this case). In some other cases, attribute | |
3215 | specifiers are permitted by this grammar but not yet supported by the | |
3216 | compiler. All attribute specifiers in this place relate to the | |
c771326b | 3217 | declaration as a whole. In the obsolescent usage where a type of |
2c5e91d2 JM |
3218 | @code{int} is implied by the absence of type specifiers, such a list of |
3219 | specifiers and qualifiers may be an attribute specifier list with no | |
3220 | other specifiers or qualifiers. | |
3221 | ||
7dcb0442 JM |
3222 | At present, the first parameter in a function prototype must have some |
3223 | type specifier which is not an attribute specifier; this resolves an | |
3224 | ambiguity in the interpretation of @code{void f(int | |
3225 | (__attribute__((foo)) x))}, but is subject to change. At present, if | |
3226 | the parentheses of a function declarator contain only attributes then | |
3227 | those attributes are ignored, rather than yielding an error or warning | |
3228 | or implying a single parameter of type int, but this is subject to | |
3229 | change. | |
3230 | ||
2c5e91d2 JM |
3231 | An attribute specifier list may appear immediately before a declarator |
3232 | (other than the first) in a comma-separated list of declarators in a | |
3233 | declaration of more than one identifier using a single list of | |
4b01f8d8 | 3234 | specifiers and qualifiers. Such attribute specifiers apply |
9c34dbbf ZW |
3235 | only to the identifier before whose declarator they appear. For |
3236 | example, in | |
3237 | ||
3238 | @smallexample | |
3239 | __attribute__((noreturn)) void d0 (void), | |
3240 | __attribute__((format(printf, 1, 2))) d1 (const char *, ...), | |
3241 | d2 (void) | |
3242 | @end smallexample | |
3243 | ||
3244 | @noindent | |
3245 | the @code{noreturn} attribute applies to all the functions | |
4b01f8d8 | 3246 | declared; the @code{format} attribute only applies to @code{d1}. |
2c5e91d2 JM |
3247 | |
3248 | An attribute specifier list may appear immediately before the comma, | |
3249 | @code{=} or semicolon terminating the declaration of an identifier other | |
770a9950 JM |
3250 | than a function definition. Such attribute specifiers apply |
3251 | to the declared object or function. Where an | |
9c34dbbf | 3252 | assembler name for an object or function is specified (@pxref{Asm |
770a9950 JM |
3253 | Labels}), the attribute must follow the @code{asm} |
3254 | specification. | |
2c5e91d2 JM |
3255 | |
3256 | An attribute specifier list may, in future, be permitted to appear after | |
3257 | the declarator in a function definition (before any old-style parameter | |
3258 | declarations or the function body). | |
3259 | ||
0e03329a JM |
3260 | Attribute specifiers may be mixed with type qualifiers appearing inside |
3261 | the @code{[]} of a parameter array declarator, in the C99 construct by | |
3262 | which such qualifiers are applied to the pointer to which the array is | |
3263 | implicitly converted. Such attribute specifiers apply to the pointer, | |
3264 | not to the array, but at present this is not implemented and they are | |
3265 | ignored. | |
3266 | ||
2c5e91d2 JM |
3267 | An attribute specifier list may appear at the start of a nested |
3268 | declarator. At present, there are some limitations in this usage: the | |
91d231cb JM |
3269 | attributes correctly apply to the declarator, but for most individual |
3270 | attributes the semantics this implies are not implemented. | |
3271 | When attribute specifiers follow the @code{*} of a pointer | |
4b01f8d8 | 3272 | declarator, they may be mixed with any type qualifiers present. |
91d231cb | 3273 | The following describes the formal semantics of this syntax. It will make the |
2c5e91d2 JM |
3274 | most sense if you are familiar with the formal specification of |
3275 | declarators in the ISO C standard. | |
3276 | ||
3277 | Consider (as in C99 subclause 6.7.5 paragraph 4) a declaration @code{T | |
3278 | D1}, where @code{T} contains declaration specifiers that specify a type | |
3279 | @var{Type} (such as @code{int}) and @code{D1} is a declarator that | |
3280 | contains an identifier @var{ident}. The type specified for @var{ident} | |
3281 | for derived declarators whose type does not include an attribute | |
3282 | specifier is as in the ISO C standard. | |
3283 | ||
3284 | If @code{D1} has the form @code{( @var{attribute-specifier-list} D )}, | |
3285 | and the declaration @code{T D} specifies the type | |
3286 | ``@var{derived-declarator-type-list} @var{Type}'' for @var{ident}, then | |
3287 | @code{T D1} specifies the type ``@var{derived-declarator-type-list} | |
3288 | @var{attribute-specifier-list} @var{Type}'' for @var{ident}. | |
3289 | ||
3290 | If @code{D1} has the form @code{* | |
3291 | @var{type-qualifier-and-attribute-specifier-list} D}, and the | |
3292 | declaration @code{T D} specifies the type | |
3293 | ``@var{derived-declarator-type-list} @var{Type}'' for @var{ident}, then | |
3294 | @code{T D1} specifies the type ``@var{derived-declarator-type-list} | |
3295 | @var{type-qualifier-and-attribute-specifier-list} @var{Type}'' for | |
3296 | @var{ident}. | |
3297 | ||
f282ffb3 | 3298 | For example, |
9c34dbbf ZW |
3299 | |
3300 | @smallexample | |
3301 | void (__attribute__((noreturn)) ****f) (void); | |
3302 | @end smallexample | |
3303 | ||
3304 | @noindent | |
3305 | specifies the type ``pointer to pointer to pointer to pointer to | |
3306 | non-returning function returning @code{void}''. As another example, | |
3307 | ||
3308 | @smallexample | |
3309 | char *__attribute__((aligned(8))) *f; | |
3310 | @end smallexample | |
3311 | ||
3312 | @noindent | |
3313 | specifies the type ``pointer to 8-byte-aligned pointer to @code{char}''. | |
91d231cb JM |
3314 | Note again that this does not work with most attributes; for example, |
3315 | the usage of @samp{aligned} and @samp{noreturn} attributes given above | |
3316 | is not yet supported. | |
3317 | ||
3318 | For compatibility with existing code written for compiler versions that | |
3319 | did not implement attributes on nested declarators, some laxity is | |
3320 | allowed in the placing of attributes. If an attribute that only applies | |
3321 | to types is applied to a declaration, it will be treated as applying to | |
3322 | the type of that declaration. If an attribute that only applies to | |
3323 | declarations is applied to the type of a declaration, it will be treated | |
3324 | as applying to that declaration; and, for compatibility with code | |
3325 | placing the attributes immediately before the identifier declared, such | |
3326 | an attribute applied to a function return type will be treated as | |
3327 | applying to the function type, and such an attribute applied to an array | |
3328 | element type will be treated as applying to the array type. If an | |
3329 | attribute that only applies to function types is applied to a | |
3330 | pointer-to-function type, it will be treated as applying to the pointer | |
3331 | target type; if such an attribute is applied to a function return type | |
3332 | that is not a pointer-to-function type, it will be treated as applying | |
3333 | to the function type. | |
2c5e91d2 | 3334 | |
c1f7febf RK |
3335 | @node Function Prototypes |
3336 | @section Prototypes and Old-Style Function Definitions | |
3337 | @cindex function prototype declarations | |
3338 | @cindex old-style function definitions | |
3339 | @cindex promotion of formal parameters | |
3340 | ||
5490d604 | 3341 | GNU C extends ISO C to allow a function prototype to override a later |
c1f7febf RK |
3342 | old-style non-prototype definition. Consider the following example: |
3343 | ||
3ab51846 | 3344 | @smallexample |
c1f7febf | 3345 | /* @r{Use prototypes unless the compiler is old-fashioned.} */ |
d863830b | 3346 | #ifdef __STDC__ |
c1f7febf RK |
3347 | #define P(x) x |
3348 | #else | |
3349 | #define P(x) () | |
3350 | #endif | |
3351 | ||
3352 | /* @r{Prototype function declaration.} */ | |
3353 | int isroot P((uid_t)); | |
3354 | ||
3355 | /* @r{Old-style function definition.} */ | |
3356 | int | |
12bcfaa1 | 3357 | isroot (x) /* @r{??? lossage here ???} */ |
c1f7febf RK |
3358 | uid_t x; |
3359 | @{ | |
3360 | return x == 0; | |
3361 | @} | |
3ab51846 | 3362 | @end smallexample |
c1f7febf | 3363 | |
5490d604 | 3364 | Suppose the type @code{uid_t} happens to be @code{short}. ISO C does |
c1f7febf RK |
3365 | not allow this example, because subword arguments in old-style |
3366 | non-prototype definitions are promoted. Therefore in this example the | |
3367 | function definition's argument is really an @code{int}, which does not | |
3368 | match the prototype argument type of @code{short}. | |
3369 | ||
5490d604 | 3370 | This restriction of ISO C makes it hard to write code that is portable |
c1f7febf RK |
3371 | to traditional C compilers, because the programmer does not know |
3372 | whether the @code{uid_t} type is @code{short}, @code{int}, or | |
3373 | @code{long}. Therefore, in cases like these GNU C allows a prototype | |
3374 | to override a later old-style definition. More precisely, in GNU C, a | |
3375 | function prototype argument type overrides the argument type specified | |
3376 | by a later old-style definition if the former type is the same as the | |
3377 | latter type before promotion. Thus in GNU C the above example is | |
3378 | equivalent to the following: | |
3379 | ||
3ab51846 | 3380 | @smallexample |
c1f7febf RK |
3381 | int isroot (uid_t); |
3382 | ||
3383 | int | |
3384 | isroot (uid_t x) | |
3385 | @{ | |
3386 | return x == 0; | |
3387 | @} | |
3ab51846 | 3388 | @end smallexample |
c1f7febf | 3389 | |
9c34dbbf | 3390 | @noindent |
c1f7febf RK |
3391 | GNU C++ does not support old-style function definitions, so this |
3392 | extension is irrelevant. | |
3393 | ||
3394 | @node C++ Comments | |
3395 | @section C++ Style Comments | |
3396 | @cindex // | |
3397 | @cindex C++ comments | |
3398 | @cindex comments, C++ style | |
3399 | ||
3400 | In GNU C, you may use C++ style comments, which start with @samp{//} and | |
3401 | continue until the end of the line. Many other C implementations allow | |
f458d1d5 ZW |
3402 | such comments, and they are included in the 1999 C standard. However, |
3403 | C++ style comments are not recognized if you specify an @option{-std} | |
3404 | option specifying a version of ISO C before C99, or @option{-ansi} | |
3405 | (equivalent to @option{-std=c89}). | |
c1f7febf RK |
3406 | |
3407 | @node Dollar Signs | |
3408 | @section Dollar Signs in Identifier Names | |
3409 | @cindex $ | |
3410 | @cindex dollar signs in identifier names | |
3411 | @cindex identifier names, dollar signs in | |
3412 | ||
79188db9 RK |
3413 | In GNU C, you may normally use dollar signs in identifier names. |
3414 | This is because many traditional C implementations allow such identifiers. | |
3415 | However, dollar signs in identifiers are not supported on a few target | |
3416 | machines, typically because the target assembler does not allow them. | |
c1f7febf RK |
3417 | |
3418 | @node Character Escapes | |
3419 | @section The Character @key{ESC} in Constants | |
3420 | ||
3421 | You can use the sequence @samp{\e} in a string or character constant to | |
3422 | stand for the ASCII character @key{ESC}. | |
3423 | ||
3424 | @node Alignment | |
3425 | @section Inquiring on Alignment of Types or Variables | |
3426 | @cindex alignment | |
3427 | @cindex type alignment | |
3428 | @cindex variable alignment | |
3429 | ||
3430 | The keyword @code{__alignof__} allows you to inquire about how an object | |
3431 | is aligned, or the minimum alignment usually required by a type. Its | |
3432 | syntax is just like @code{sizeof}. | |
3433 | ||
3434 | For example, if the target machine requires a @code{double} value to be | |
3435 | aligned on an 8-byte boundary, then @code{__alignof__ (double)} is 8. | |
3436 | This is true on many RISC machines. On more traditional machine | |
3437 | designs, @code{__alignof__ (double)} is 4 or even 2. | |
3438 | ||
3439 | Some machines never actually require alignment; they allow reference to any | |
64c18e57 | 3440 | data type even at an odd address. For these machines, @code{__alignof__} |
2a31c49c HPN |
3441 | reports the smallest alignment that GCC will give the data type, usually as |
3442 | mandated by the target ABI. | |
c1f7febf | 3443 | |
5372b3fb NB |
3444 | If the operand of @code{__alignof__} is an lvalue rather than a type, |
3445 | its value is the required alignment for its type, taking into account | |
3446 | any minimum alignment specified with GCC's @code{__attribute__} | |
3447 | extension (@pxref{Variable Attributes}). For example, after this | |
3448 | declaration: | |
c1f7febf | 3449 | |
3ab51846 | 3450 | @smallexample |
c1f7febf | 3451 | struct foo @{ int x; char y; @} foo1; |
3ab51846 | 3452 | @end smallexample |
c1f7febf RK |
3453 | |
3454 | @noindent | |
5372b3fb NB |
3455 | the value of @code{__alignof__ (foo1.y)} is 1, even though its actual |
3456 | alignment is probably 2 or 4, the same as @code{__alignof__ (int)}. | |
c1f7febf | 3457 | |
9d27bffe SS |
3458 | It is an error to ask for the alignment of an incomplete type. |
3459 | ||
c1f7febf RK |
3460 | @node Variable Attributes |
3461 | @section Specifying Attributes of Variables | |
3462 | @cindex attribute of variables | |
3463 | @cindex variable attributes | |
3464 | ||
3465 | The keyword @code{__attribute__} allows you to specify special | |
3466 | attributes of variables or structure fields. This keyword is followed | |
905e8651 RH |
3467 | by an attribute specification inside double parentheses. Some |
3468 | attributes are currently defined generically for variables. | |
3469 | Other attributes are defined for variables on particular target | |
3470 | systems. Other attributes are available for functions | |
3471 | (@pxref{Function Attributes}) and for types (@pxref{Type Attributes}). | |
3472 | Other front ends might define more attributes | |
3473 | (@pxref{C++ Extensions,,Extensions to the C++ Language}). | |
c1f7febf RK |
3474 | |
3475 | You may also specify attributes with @samp{__} preceding and following | |
3476 | each keyword. This allows you to use them in header files without | |
3477 | being concerned about a possible macro of the same name. For example, | |
3478 | you may use @code{__aligned__} instead of @code{aligned}. | |
3479 | ||
2c5e91d2 JM |
3480 | @xref{Attribute Syntax}, for details of the exact syntax for using |
3481 | attributes. | |
3482 | ||
c1f7febf RK |
3483 | @table @code |
3484 | @cindex @code{aligned} attribute | |
3485 | @item aligned (@var{alignment}) | |
3486 | This attribute specifies a minimum alignment for the variable or | |
3487 | structure field, measured in bytes. For example, the declaration: | |
3488 | ||
3489 | @smallexample | |
3490 | int x __attribute__ ((aligned (16))) = 0; | |
3491 | @end smallexample | |
3492 | ||
3493 | @noindent | |
3494 | causes the compiler to allocate the global variable @code{x} on a | |
3495 | 16-byte boundary. On a 68040, this could be used in conjunction with | |
3496 | an @code{asm} expression to access the @code{move16} instruction which | |
3497 | requires 16-byte aligned operands. | |
3498 | ||
3499 | You can also specify the alignment of structure fields. For example, to | |
3500 | create a double-word aligned @code{int} pair, you could write: | |
3501 | ||
3502 | @smallexample | |
3503 | struct foo @{ int x[2] __attribute__ ((aligned (8))); @}; | |
3504 | @end smallexample | |
3505 | ||
3506 | @noindent | |
3507 | This is an alternative to creating a union with a @code{double} member | |
3508 | that forces the union to be double-word aligned. | |
3509 | ||
c1f7febf RK |
3510 | As in the preceding examples, you can explicitly specify the alignment |
3511 | (in bytes) that you wish the compiler to use for a given variable or | |
3512 | structure field. Alternatively, you can leave out the alignment factor | |
3513 | and just ask the compiler to align a variable or field to the maximum | |
3514 | useful alignment for the target machine you are compiling for. For | |
3515 | example, you could write: | |
3516 | ||
3517 | @smallexample | |
3518 | short array[3] __attribute__ ((aligned)); | |
3519 | @end smallexample | |
3520 | ||
3521 | Whenever you leave out the alignment factor in an @code{aligned} attribute | |
3522 | specification, the compiler automatically sets the alignment for the declared | |
3523 | variable or field to the largest alignment which is ever used for any data | |
3524 | type on the target machine you are compiling for. Doing this can often make | |
3525 | copy operations more efficient, because the compiler can use whatever | |
3526 | instructions copy the biggest chunks of memory when performing copies to | |
3527 | or from the variables or fields that you have aligned this way. | |
3528 | ||
e9f9692b MW |
3529 | When used on a struct, or struct member, the @code{aligned} attribute can |
3530 | only increase the alignment; in order to decrease it, the @code{packed} | |
3531 | attribute must be specified as well. When used as part of a typedef, the | |
3532 | @code{aligned} attribute can both increase and decrease alignment, and | |
3533 | specifying the @code{packed} attribute will generate a warning. | |
c1f7febf RK |
3534 | |
3535 | Note that the effectiveness of @code{aligned} attributes may be limited | |
3536 | by inherent limitations in your linker. On many systems, the linker is | |
3537 | only able to arrange for variables to be aligned up to a certain maximum | |
3538 | alignment. (For some linkers, the maximum supported alignment may | |
3539 | be very very small.) If your linker is only able to align variables | |
3540 | up to a maximum of 8 byte alignment, then specifying @code{aligned(16)} | |
3541 | in an @code{__attribute__} will still only provide you with 8 byte | |
3542 | alignment. See your linker documentation for further information. | |
3543 | ||
837edd5f GK |
3544 | The @code{aligned} attribute can also be used for functions |
3545 | (@pxref{Function Attributes}.) | |
3546 | ||
0bfa5f65 RH |
3547 | @item cleanup (@var{cleanup_function}) |
3548 | @cindex @code{cleanup} attribute | |
3549 | The @code{cleanup} attribute runs a function when the variable goes | |
3550 | out of scope. This attribute can only be applied to auto function | |
3551 | scope variables; it may not be applied to parameters or variables | |
3552 | with static storage duration. The function must take one parameter, | |
3553 | a pointer to a type compatible with the variable. The return value | |
3554 | of the function (if any) is ignored. | |
3555 | ||
3556 | If @option{-fexceptions} is enabled, then @var{cleanup_function} | |
3557 | will be run during the stack unwinding that happens during the | |
3558 | processing of the exception. Note that the @code{cleanup} attribute | |
3559 | does not allow the exception to be caught, only to perform an action. | |
3560 | It is undefined what happens if @var{cleanup_function} does not | |
3561 | return normally. | |
3562 | ||
905e8651 RH |
3563 | @item common |
3564 | @itemx nocommon | |
3565 | @cindex @code{common} attribute | |
3566 | @cindex @code{nocommon} attribute | |
3567 | @opindex fcommon | |
3568 | @opindex fno-common | |
3569 | The @code{common} attribute requests GCC to place a variable in | |
3570 | ``common'' storage. The @code{nocommon} attribute requests the | |
78466c0e | 3571 | opposite---to allocate space for it directly. |
905e8651 | 3572 | |
daf2f129 | 3573 | These attributes override the default chosen by the |
905e8651 RH |
3574 | @option{-fno-common} and @option{-fcommon} flags respectively. |
3575 | ||
3576 | @item deprecated | |
3577 | @cindex @code{deprecated} attribute | |
3578 | The @code{deprecated} attribute results in a warning if the variable | |
3579 | is used anywhere in the source file. This is useful when identifying | |
3580 | variables that are expected to be removed in a future version of a | |
3581 | program. The warning also includes the location of the declaration | |
3582 | of the deprecated variable, to enable users to easily find further | |
3583 | information about why the variable is deprecated, or what they should | |
64c18e57 | 3584 | do instead. Note that the warning only occurs for uses: |
905e8651 RH |
3585 | |
3586 | @smallexample | |
3587 | extern int old_var __attribute__ ((deprecated)); | |
3588 | extern int old_var; | |
3589 | int new_fn () @{ return old_var; @} | |
3590 | @end smallexample | |
3591 | ||
3592 | results in a warning on line 3 but not line 2. | |
3593 | ||
3594 | The @code{deprecated} attribute can also be used for functions and | |
3595 | types (@pxref{Function Attributes}, @pxref{Type Attributes}.) | |
3596 | ||
c1f7febf RK |
3597 | @item mode (@var{mode}) |
3598 | @cindex @code{mode} attribute | |
3599 | This attribute specifies the data type for the declaration---whichever | |
3600 | type corresponds to the mode @var{mode}. This in effect lets you | |
3601 | request an integer or floating point type according to its width. | |
3602 | ||
3603 | You may also specify a mode of @samp{byte} or @samp{__byte__} to | |
3604 | indicate the mode corresponding to a one-byte integer, @samp{word} or | |
3605 | @samp{__word__} for the mode of a one-word integer, and @samp{pointer} | |
3606 | or @samp{__pointer__} for the mode used to represent pointers. | |
3607 | ||
c1f7febf RK |
3608 | @item packed |
3609 | @cindex @code{packed} attribute | |
3610 | The @code{packed} attribute specifies that a variable or structure field | |
3611 | should have the smallest possible alignment---one byte for a variable, | |
3612 | and one bit for a field, unless you specify a larger value with the | |
3613 | @code{aligned} attribute. | |
3614 | ||
3615 | Here is a structure in which the field @code{x} is packed, so that it | |
3616 | immediately follows @code{a}: | |
3617 | ||
3ab51846 | 3618 | @smallexample |
c1f7febf RK |
3619 | struct foo |
3620 | @{ | |
3621 | char a; | |
3622 | int x[2] __attribute__ ((packed)); | |
3623 | @}; | |
3ab51846 | 3624 | @end smallexample |
c1f7febf | 3625 | |
84330467 | 3626 | @item section ("@var{section-name}") |
c1f7febf RK |
3627 | @cindex @code{section} variable attribute |
3628 | Normally, the compiler places the objects it generates in sections like | |
3629 | @code{data} and @code{bss}. Sometimes, however, you need additional sections, | |
3630 | or you need certain particular variables to appear in special sections, | |
3631 | for example to map to special hardware. The @code{section} | |
3632 | attribute specifies that a variable (or function) lives in a particular | |
3633 | section. For example, this small program uses several specific section names: | |
3634 | ||
3635 | @smallexample | |
3636 | struct duart a __attribute__ ((section ("DUART_A"))) = @{ 0 @}; | |
3637 | struct duart b __attribute__ ((section ("DUART_B"))) = @{ 0 @}; | |
3638 | char stack[10000] __attribute__ ((section ("STACK"))) = @{ 0 @}; | |
3639 | int init_data __attribute__ ((section ("INITDATA"))) = 0; | |
3640 | ||
3641 | main() | |
3642 | @{ | |
12bcfaa1 | 3643 | /* @r{Initialize stack pointer} */ |
c1f7febf RK |
3644 | init_sp (stack + sizeof (stack)); |
3645 | ||
12bcfaa1 | 3646 | /* @r{Initialize initialized data} */ |
c1f7febf RK |
3647 | memcpy (&init_data, &data, &edata - &data); |
3648 | ||
12bcfaa1 | 3649 | /* @r{Turn on the serial ports} */ |
c1f7febf RK |
3650 | init_duart (&a); |
3651 | init_duart (&b); | |
3652 | @} | |
3653 | @end smallexample | |
3654 | ||
3655 | @noindent | |
3656 | Use the @code{section} attribute with an @emph{initialized} definition | |
f0523f02 | 3657 | of a @emph{global} variable, as shown in the example. GCC issues |
c1f7febf RK |
3658 | a warning and otherwise ignores the @code{section} attribute in |
3659 | uninitialized variable declarations. | |
3660 | ||
3661 | You may only use the @code{section} attribute with a fully initialized | |
3662 | global definition because of the way linkers work. The linker requires | |
3663 | each object be defined once, with the exception that uninitialized | |
3664 | variables tentatively go in the @code{common} (or @code{bss}) section | |
84330467 JM |
3665 | and can be multiply ``defined''. You can force a variable to be |
3666 | initialized with the @option{-fno-common} flag or the @code{nocommon} | |
c1f7febf RK |
3667 | attribute. |
3668 | ||
3669 | Some file formats do not support arbitrary sections so the @code{section} | |
3670 | attribute is not available on all platforms. | |
3671 | If you need to map the entire contents of a module to a particular | |
3672 | section, consider using the facilities of the linker instead. | |
3673 | ||
593d3a34 MK |
3674 | @item shared |
3675 | @cindex @code{shared} variable attribute | |
95fef11f | 3676 | On Microsoft Windows, in addition to putting variable definitions in a named |
02f52e19 | 3677 | section, the section can also be shared among all running copies of an |
161d7b59 | 3678 | executable or DLL@. For example, this small program defines shared data |
84330467 | 3679 | by putting it in a named section @code{shared} and marking the section |
593d3a34 MK |
3680 | shareable: |
3681 | ||
3682 | @smallexample | |
3683 | int foo __attribute__((section ("shared"), shared)) = 0; | |
3684 | ||
3685 | int | |
3686 | main() | |
3687 | @{ | |
12bcfaa1 JM |
3688 | /* @r{Read and write foo. All running |
3689 | copies see the same value.} */ | |
593d3a34 MK |
3690 | return 0; |
3691 | @} | |
3692 | @end smallexample | |
3693 | ||
3694 | @noindent | |
3695 | You may only use the @code{shared} attribute along with @code{section} | |
02f52e19 | 3696 | attribute with a fully initialized global definition because of the way |
593d3a34 MK |
3697 | linkers work. See @code{section} attribute for more information. |
3698 | ||
95fef11f | 3699 | The @code{shared} attribute is only available on Microsoft Windows@. |
593d3a34 | 3700 | |
905e8651 RH |
3701 | @item tls_model ("@var{tls_model}") |
3702 | @cindex @code{tls_model} attribute | |
3703 | The @code{tls_model} attribute sets thread-local storage model | |
3704 | (@pxref{Thread-Local}) of a particular @code{__thread} variable, | |
4ec7afd7 | 3705 | overriding @option{-ftls-model=} command line switch on a per-variable |
905e8651 RH |
3706 | basis. |
3707 | The @var{tls_model} argument should be one of @code{global-dynamic}, | |
3708 | @code{local-dynamic}, @code{initial-exec} or @code{local-exec}. | |
3709 | ||
3710 | Not all targets support this attribute. | |
3711 | ||
c1f7febf RK |
3712 | @item unused |
3713 | This attribute, attached to a variable, means that the variable is meant | |
f0523f02 | 3714 | to be possibly unused. GCC will not produce a warning for this |
c1f7febf RK |
3715 | variable. |
3716 | ||
5f79d643 RM |
3717 | @item used |
3718 | This attribute, attached to a variable, means that the variable must be | |
3719 | emitted even if it appears that the variable is not referenced. | |
3720 | ||
1b9191d2 AH |
3721 | @item vector_size (@var{bytes}) |
3722 | This attribute specifies the vector size for the variable, measured in | |
3723 | bytes. For example, the declaration: | |
3724 | ||
3725 | @smallexample | |
3726 | int foo __attribute__ ((vector_size (16))); | |
3727 | @end smallexample | |
3728 | ||
3729 | @noindent | |
3730 | causes the compiler to set the mode for @code{foo}, to be 16 bytes, | |
3731 | divided into @code{int} sized units. Assuming a 32-bit int (a vector of | |
3732 | 4 units of 4 bytes), the corresponding mode of @code{foo} will be V4SI@. | |
3733 | ||
3734 | This attribute is only applicable to integral and float scalars, | |
3735 | although arrays, pointers, and function return values are allowed in | |
3736 | conjunction with this construct. | |
3737 | ||
3738 | Aggregates with this attribute are invalid, even if they are of the same | |
3739 | size as a corresponding scalar. For example, the declaration: | |
3740 | ||
3741 | @smallexample | |
ad706f54 | 3742 | struct S @{ int a; @}; |
1b9191d2 AH |
3743 | struct S __attribute__ ((vector_size (16))) foo; |
3744 | @end smallexample | |
3745 | ||
3746 | @noindent | |
3747 | is invalid even if the size of the structure is the same as the size of | |
3748 | the @code{int}. | |
3749 | ||
a20f6f00 DS |
3750 | @item selectany |
3751 | The @code{selectany} attribute causes an initialized global variable to | |
3752 | have link-once semantics. When multiple definitions of the variable are | |
3753 | encountered by the linker, the first is selected and the remainder are | |
3754 | discarded. Following usage by the Microsoft compiler, the linker is told | |
3755 | @emph{not} to warn about size or content differences of the multiple | |
3756 | definitions. | |
3757 | ||
3758 | Although the primary usage of this attribute is for POD types, the | |
3759 | attribute can also be applied to global C++ objects that are initialized | |
3760 | by a constructor. In this case, the static initialization and destruction | |
3761 | code for the object is emitted in each translation defining the object, | |
3762 | but the calls to the constructor and destructor are protected by a | |
0ac11108 | 3763 | link-once guard variable. |
a20f6f00 DS |
3764 | |
3765 | The @code{selectany} attribute is only available on Microsoft Windows | |
3766 | targets. You can use @code{__declspec (selectany)} as a synonym for | |
3767 | @code{__attribute__ ((selectany))} for compatibility with other | |
3768 | compilers. | |
3769 | ||
c1f7febf | 3770 | @item weak |
38bb2b65 | 3771 | The @code{weak} attribute is described in @ref{Function Attributes}. |
6b6cb52e DS |
3772 | |
3773 | @item dllimport | |
38bb2b65 | 3774 | The @code{dllimport} attribute is described in @ref{Function Attributes}. |
6b6cb52e | 3775 | |
9baf8aea | 3776 | @item dllexport |
38bb2b65 | 3777 | The @code{dllexport} attribute is described in @ref{Function Attributes}. |
6b6cb52e | 3778 | |
905e8651 RH |
3779 | @end table |
3780 | ||
4af797b5 JZ |
3781 | @subsection Blackfin Variable Attributes |
3782 | ||
3783 | Three attributes are currently defined for the Blackfin. | |
3784 | ||
3785 | @table @code | |
3786 | @item l1_data | |
3787 | @item l1_data_A | |
3788 | @item l1_data_B | |
3789 | @cindex @code{l1_data} variable attribute | |
3790 | @cindex @code{l1_data_A} variable attribute | |
3791 | @cindex @code{l1_data_B} variable attribute | |
3792 | Use these attributes on the Blackfin to place the variable into L1 Data SRAM. | |
3793 | Variables with @code{l1_data} attribute will be put into the specific section | |
3794 | named @code{.l1.data}. Those with @code{l1_data_A} attribute will be put into | |
3795 | the specific section named @code{.l1.data.A}. Those with @code{l1_data_B} | |
3796 | attribute will be put into the specific section named @code{.l1.data.B}. | |
3797 | @end table | |
3798 | ||
905e8651 | 3799 | @subsection M32R/D Variable Attributes |
845da534 | 3800 | |
8a36672b | 3801 | One attribute is currently defined for the M32R/D@. |
905e8651 RH |
3802 | |
3803 | @table @code | |
845da534 DE |
3804 | @item model (@var{model-name}) |
3805 | @cindex variable addressability on the M32R/D | |
3806 | Use this attribute on the M32R/D to set the addressability of an object. | |
3807 | The identifier @var{model-name} is one of @code{small}, @code{medium}, | |
3808 | or @code{large}, representing each of the code models. | |
3809 | ||
3810 | Small model objects live in the lower 16MB of memory (so that their | |
3811 | addresses can be loaded with the @code{ld24} instruction). | |
3812 | ||
02f52e19 | 3813 | Medium and large model objects may live anywhere in the 32-bit address space |
845da534 DE |
3814 | (the compiler will generate @code{seth/add3} instructions to load their |
3815 | addresses). | |
905e8651 | 3816 | @end table |
845da534 | 3817 | |
1ccbef77 | 3818 | @anchor{i386 Variable Attributes} |
fe77449a DR |
3819 | @subsection i386 Variable Attributes |
3820 | ||
3821 | Two attributes are currently defined for i386 configurations: | |
3822 | @code{ms_struct} and @code{gcc_struct} | |
3823 | ||
905e8651 | 3824 | @table @code |
fe77449a DR |
3825 | @item ms_struct |
3826 | @itemx gcc_struct | |
905e8651 RH |
3827 | @cindex @code{ms_struct} attribute |
3828 | @cindex @code{gcc_struct} attribute | |
fe77449a DR |
3829 | |
3830 | If @code{packed} is used on a structure, or if bit-fields are used | |
3831 | it may be that the Microsoft ABI packs them differently | |
3832 | than GCC would normally pack them. Particularly when moving packed | |
3833 | data between functions compiled with GCC and the native Microsoft compiler | |
3834 | (either via function call or as data in a file), it may be necessary to access | |
3835 | either format. | |
3836 | ||
95fef11f | 3837 | Currently @option{-m[no-]ms-bitfields} is provided for the Microsoft Windows X86 |
fe77449a | 3838 | compilers to match the native Microsoft compiler. |
0ac11108 EC |
3839 | |
3840 | The Microsoft structure layout algorithm is fairly simple with the exception | |
3841 | of the bitfield packing: | |
3842 | ||
3843 | The padding and alignment of members of structures and whether a bit field | |
3844 | can straddle a storage-unit boundary | |
3845 | ||
3846 | @enumerate | |
3847 | @item Structure members are stored sequentially in the order in which they are | |
3848 | declared: the first member has the lowest memory address and the last member | |
3849 | the highest. | |
3850 | ||
3851 | @item Every data object has an alignment-requirement. The alignment-requirement | |
3852 | for all data except structures, unions, and arrays is either the size of the | |
3853 | object or the current packing size (specified with either the aligned attribute | |
3854 | or the pack pragma), whichever is less. For structures, unions, and arrays, | |
3855 | the alignment-requirement is the largest alignment-requirement of its members. | |
3856 | Every object is allocated an offset so that: | |
3857 | ||
3858 | offset % alignment-requirement == 0 | |
3859 | ||
3860 | @item Adjacent bit fields are packed into the same 1-, 2-, or 4-byte allocation | |
3861 | unit if the integral types are the same size and if the next bit field fits | |
3862 | into the current allocation unit without crossing the boundary imposed by the | |
3863 | common alignment requirements of the bit fields. | |
3864 | @end enumerate | |
3865 | ||
3866 | Handling of zero-length bitfields: | |
3867 | ||
3868 | MSVC interprets zero-length bitfields in the following ways: | |
3869 | ||
3870 | @enumerate | |
3871 | @item If a zero-length bitfield is inserted between two bitfields that would | |
3872 | normally be coalesced, the bitfields will not be coalesced. | |
3873 | ||
3874 | For example: | |
3875 | ||
3876 | @smallexample | |
3877 | struct | |
3878 | @{ | |
3879 | unsigned long bf_1 : 12; | |
3880 | unsigned long : 0; | |
3881 | unsigned long bf_2 : 12; | |
3882 | @} t1; | |
3883 | @end smallexample | |
3884 | ||
3885 | The size of @code{t1} would be 8 bytes with the zero-length bitfield. If the | |
3886 | zero-length bitfield were removed, @code{t1}'s size would be 4 bytes. | |
3887 | ||
3888 | @item If a zero-length bitfield is inserted after a bitfield, @code{foo}, and the | |
3889 | alignment of the zero-length bitfield is greater than the member that follows it, | |
3890 | @code{bar}, @code{bar} will be aligned as the type of the zero-length bitfield. | |
3891 | ||
3892 | For example: | |
3893 | ||
3894 | @smallexample | |
3895 | struct | |
3896 | @{ | |
3897 | char foo : 4; | |
3898 | short : 0; | |
3899 | char bar; | |
3900 | @} t2; | |
3901 | ||
3902 | struct | |
3903 | @{ | |
3904 | char foo : 4; | |
3905 | short : 0; | |
3906 | double bar; | |
3907 | @} t3; | |
3908 | @end smallexample | |
3909 | ||
3910 | For @code{t2}, @code{bar} will be placed at offset 2, rather than offset 1. | |
3911 | Accordingly, the size of @code{t2} will be 4. For @code{t3}, the zero-length | |
3912 | bitfield will not affect the alignment of @code{bar} or, as a result, the size | |
3913 | of the structure. | |
3914 | ||
3915 | Taking this into account, it is important to note the following: | |
3916 | ||
3917 | @enumerate | |
3918 | @item If a zero-length bitfield follows a normal bitfield, the type of the | |
3919 | zero-length bitfield may affect the alignment of the structure as whole. For | |
3920 | example, @code{t2} has a size of 4 bytes, since the zero-length bitfield follows a | |
3921 | normal bitfield, and is of type short. | |
3922 | ||
3923 | @item Even if a zero-length bitfield is not followed by a normal bitfield, it may | |
3924 | still affect the alignment of the structure: | |
3925 | ||
3926 | @smallexample | |
3927 | struct | |
3928 | @{ | |
3929 | char foo : 6; | |
3930 | long : 0; | |
3931 | @} t4; | |
3932 | @end smallexample | |
3933 | ||
3934 | Here, @code{t4} will take up 4 bytes. | |
3935 | @end enumerate | |
3936 | ||
3937 | @item Zero-length bitfields following non-bitfield members are ignored: | |
3938 | ||
3939 | @smallexample | |
3940 | struct | |
3941 | @{ | |
3942 | char foo; | |
3943 | long : 0; | |
3944 | char bar; | |
3945 | @} t5; | |
3946 | @end smallexample | |
3947 | ||
3948 | Here, @code{t5} will take up 2 bytes. | |
3949 | @end enumerate | |
c1f7febf RK |
3950 | @end table |
3951 | ||
1ccbef77 EC |
3952 | @subsection PowerPC Variable Attributes |
3953 | ||
63d0dca4 DE |
3954 | Three attributes currently are defined for PowerPC configurations: |
3955 | @code{altivec}, @code{ms_struct} and @code{gcc_struct}. | |
1ccbef77 | 3956 | |
63d0dca4 | 3957 | For full documentation of the struct attributes please see the |
38bb2b65 | 3958 | documentation in @ref{i386 Variable Attributes}. |
63d0dca4 DE |
3959 | |
3960 | For documentation of @code{altivec} attribute please see the | |
38bb2b65 | 3961 | documentation in @ref{PowerPC Type Attributes}. |
1ccbef77 | 3962 | |
85d9c13c TS |
3963 | @subsection SPU Variable Attributes |
3964 | ||
3965 | The SPU supports the @code{spu_vector} attribute for variables. For | |
38bb2b65 SL |
3966 | documentation of this attribute please see the documentation in |
3967 | @ref{SPU Type Attributes}. | |
85d9c13c | 3968 | |
54e9a19d DD |
3969 | @subsection Xstormy16 Variable Attributes |
3970 | ||
3971 | One attribute is currently defined for xstormy16 configurations: | |
38bb2b65 | 3972 | @code{below100}. |
54e9a19d DD |
3973 | |
3974 | @table @code | |
3975 | @item below100 | |
3976 | @cindex @code{below100} attribute | |
3977 | ||
3978 | If a variable has the @code{below100} attribute (@code{BELOW100} is | |
3979 | allowed also), GCC will place the variable in the first 0x100 bytes of | |
3980 | memory and use special opcodes to access it. Such variables will be | |
3981 | placed in either the @code{.bss_below100} section or the | |
3982 | @code{.data_below100} section. | |
3983 | ||
3984 | @end table | |
3985 | ||
79532d34 EW |
3986 | @subsection AVR Variable Attributes |
3987 | ||
3988 | @table @code | |
3989 | @item progmem | |
3990 | @cindex @code{progmem} variable attribute | |
3991 | The @code{progmem} attribute is used on the AVR to place data in the Program | |
3992 | Memory address space. The AVR is a Harvard Architecture processor and data | |
3993 | normally resides in the Data Memory address space. | |
3994 | @end table | |
3995 | ||
c1f7febf RK |
3996 | @node Type Attributes |
3997 | @section Specifying Attributes of Types | |
3998 | @cindex attribute of types | |
3999 | @cindex type attributes | |
4000 | ||
4001 | The keyword @code{__attribute__} allows you to specify special | |
b9e75696 JM |
4002 | attributes of @code{struct} and @code{union} types when you define |
4003 | such types. This keyword is followed by an attribute specification | |
4004 | inside double parentheses. Seven attributes are currently defined for | |
4005 | types: @code{aligned}, @code{packed}, @code{transparent_union}, | |
4006 | @code{unused}, @code{deprecated}, @code{visibility}, and | |
4007 | @code{may_alias}. Other attributes are defined for functions | |
4008 | (@pxref{Function Attributes}) and for variables (@pxref{Variable | |
4009 | Attributes}). | |
c1f7febf RK |
4010 | |
4011 | You may also specify any one of these attributes with @samp{__} | |
4012 | preceding and following its keyword. This allows you to use these | |
4013 | attributes in header files without being concerned about a possible | |
4014 | macro of the same name. For example, you may use @code{__aligned__} | |
4015 | instead of @code{aligned}. | |
4016 | ||
4009f2e7 JM |
4017 | You may specify type attributes in an enum, struct or union type |
4018 | declaration or definition, or for other types in a @code{typedef} | |
4019 | declaration. | |
c1f7febf | 4020 | |
b9e75696 JM |
4021 | For an enum, struct or union type, you may specify attributes either |
4022 | between the enum, struct or union tag and the name of the type, or | |
4023 | just past the closing curly brace of the @emph{definition}. The | |
4024 | former syntax is preferred. | |
4051959b | 4025 | |
2c5e91d2 JM |
4026 | @xref{Attribute Syntax}, for details of the exact syntax for using |
4027 | attributes. | |
4028 | ||
c1f7febf RK |
4029 | @table @code |
4030 | @cindex @code{aligned} attribute | |
4031 | @item aligned (@var{alignment}) | |
4032 | This attribute specifies a minimum alignment (in bytes) for variables | |
4033 | of the specified type. For example, the declarations: | |
4034 | ||
4035 | @smallexample | |
f69eecfb JL |
4036 | struct S @{ short f[3]; @} __attribute__ ((aligned (8))); |
4037 | typedef int more_aligned_int __attribute__ ((aligned (8))); | |
c1f7febf RK |
4038 | @end smallexample |
4039 | ||
4040 | @noindent | |
d863830b | 4041 | force the compiler to insure (as far as it can) that each variable whose |
c1f7febf | 4042 | type is @code{struct S} or @code{more_aligned_int} will be allocated and |
981f6289 | 4043 | aligned @emph{at least} on a 8-byte boundary. On a SPARC, having all |
c1f7febf RK |
4044 | variables of type @code{struct S} aligned to 8-byte boundaries allows |
4045 | the compiler to use the @code{ldd} and @code{std} (doubleword load and | |
4046 | store) instructions when copying one variable of type @code{struct S} to | |
4047 | another, thus improving run-time efficiency. | |
4048 | ||
4049 | Note that the alignment of any given @code{struct} or @code{union} type | |
5490d604 | 4050 | is required by the ISO C standard to be at least a perfect multiple of |
c1f7febf RK |
4051 | the lowest common multiple of the alignments of all of the members of |
4052 | the @code{struct} or @code{union} in question. This means that you @emph{can} | |
4053 | effectively adjust the alignment of a @code{struct} or @code{union} | |
4054 | type by attaching an @code{aligned} attribute to any one of the members | |
4055 | of such a type, but the notation illustrated in the example above is a | |
4056 | more obvious, intuitive, and readable way to request the compiler to | |
4057 | adjust the alignment of an entire @code{struct} or @code{union} type. | |
4058 | ||
4059 | As in the preceding example, you can explicitly specify the alignment | |
4060 | (in bytes) that you wish the compiler to use for a given @code{struct} | |
4061 | or @code{union} type. Alternatively, you can leave out the alignment factor | |
4062 | and just ask the compiler to align a type to the maximum | |
4063 | useful alignment for the target machine you are compiling for. For | |
4064 | example, you could write: | |
4065 | ||
4066 | @smallexample | |
4067 | struct S @{ short f[3]; @} __attribute__ ((aligned)); | |
4068 | @end smallexample | |
4069 | ||
4070 | Whenever you leave out the alignment factor in an @code{aligned} | |
4071 | attribute specification, the compiler automatically sets the alignment | |
4072 | for the type to the largest alignment which is ever used for any data | |
4073 | type on the target machine you are compiling for. Doing this can often | |
4074 | make copy operations more efficient, because the compiler can use | |
4075 | whatever instructions copy the biggest chunks of memory when performing | |
4076 | copies to or from the variables which have types that you have aligned | |
4077 | this way. | |
4078 | ||
4079 | In the example above, if the size of each @code{short} is 2 bytes, then | |
4080 | the size of the entire @code{struct S} type is 6 bytes. The smallest | |
4081 | power of two which is greater than or equal to that is 8, so the | |
4082 | compiler sets the alignment for the entire @code{struct S} type to 8 | |
4083 | bytes. | |
4084 | ||
4085 | Note that although you can ask the compiler to select a time-efficient | |
4086 | alignment for a given type and then declare only individual stand-alone | |
4087 | objects of that type, the compiler's ability to select a time-efficient | |
4088 | alignment is primarily useful only when you plan to create arrays of | |
4089 | variables having the relevant (efficiently aligned) type. If you | |
4090 | declare or use arrays of variables of an efficiently-aligned type, then | |
4091 | it is likely that your program will also be doing pointer arithmetic (or | |
4092 | subscripting, which amounts to the same thing) on pointers to the | |
4093 | relevant type, and the code that the compiler generates for these | |
4094 | pointer arithmetic operations will often be more efficient for | |
4095 | efficiently-aligned types than for other types. | |
4096 | ||
4097 | The @code{aligned} attribute can only increase the alignment; but you | |
4098 | can decrease it by specifying @code{packed} as well. See below. | |
4099 | ||
4100 | Note that the effectiveness of @code{aligned} attributes may be limited | |
4101 | by inherent limitations in your linker. On many systems, the linker is | |
4102 | only able to arrange for variables to be aligned up to a certain maximum | |
4103 | alignment. (For some linkers, the maximum supported alignment may | |
4104 | be very very small.) If your linker is only able to align variables | |
4105 | up to a maximum of 8 byte alignment, then specifying @code{aligned(16)} | |
4106 | in an @code{__attribute__} will still only provide you with 8 byte | |
4107 | alignment. See your linker documentation for further information. | |
4108 | ||
4109 | @item packed | |
a5bcc582 | 4110 | This attribute, attached to @code{struct} or @code{union} type |
d1a701eb MM |
4111 | definition, specifies that each member (other than zero-width bitfields) |
4112 | of the structure or union is placed to minimize the memory required. When | |
4113 | attached to an @code{enum} definition, it indicates that the smallest | |
4114 | integral type should be used. | |
c1f7febf | 4115 | |
84330467 | 4116 | @opindex fshort-enums |
c1f7febf RK |
4117 | Specifying this attribute for @code{struct} and @code{union} types is |
4118 | equivalent to specifying the @code{packed} attribute on each of the | |
84330467 | 4119 | structure or union members. Specifying the @option{-fshort-enums} |
c1f7febf RK |
4120 | flag on the line is equivalent to specifying the @code{packed} |
4121 | attribute on all @code{enum} definitions. | |
4122 | ||
a5bcc582 NS |
4123 | In the following example @code{struct my_packed_struct}'s members are |
4124 | packed closely together, but the internal layout of its @code{s} member | |
78466c0e | 4125 | is not packed---to do that, @code{struct my_unpacked_struct} would need to |
a5bcc582 NS |
4126 | be packed too. |
4127 | ||
4128 | @smallexample | |
4129 | struct my_unpacked_struct | |
4130 | @{ | |
4131 | char c; | |
4132 | int i; | |
4133 | @}; | |
4134 | ||
75b66a16 | 4135 | struct __attribute__ ((__packed__)) my_packed_struct |
a5bcc582 NS |
4136 | @{ |
4137 | char c; | |
4138 | int i; | |
4139 | struct my_unpacked_struct s; | |
4140 | @}; | |
4141 | @end smallexample | |
4142 | ||
4143 | You may only specify this attribute on the definition of a @code{enum}, | |
4144 | @code{struct} or @code{union}, not on a @code{typedef} which does not | |
4145 | also define the enumerated type, structure or union. | |
c1f7febf RK |
4146 | |
4147 | @item transparent_union | |
4148 | This attribute, attached to a @code{union} type definition, indicates | |
4149 | that any function parameter having that union type causes calls to that | |
4150 | function to be treated in a special way. | |
4151 | ||
4152 | First, the argument corresponding to a transparent union type can be of | |
4153 | any type in the union; no cast is required. Also, if the union contains | |
4154 | a pointer type, the corresponding argument can be a null pointer | |
4155 | constant or a void pointer expression; and if the union contains a void | |
4156 | pointer type, the corresponding argument can be any pointer expression. | |
4157 | If the union member type is a pointer, qualifiers like @code{const} on | |
4158 | the referenced type must be respected, just as with normal pointer | |
4159 | conversions. | |
4160 | ||
4161 | Second, the argument is passed to the function using the calling | |
64c18e57 | 4162 | conventions of the first member of the transparent union, not the calling |
c1f7febf RK |
4163 | conventions of the union itself. All members of the union must have the |
4164 | same machine representation; this is necessary for this argument passing | |
4165 | to work properly. | |
4166 | ||
4167 | Transparent unions are designed for library functions that have multiple | |
4168 | interfaces for compatibility reasons. For example, suppose the | |
4169 | @code{wait} function must accept either a value of type @code{int *} to | |
4170 | comply with Posix, or a value of type @code{union wait *} to comply with | |
4171 | the 4.1BSD interface. If @code{wait}'s parameter were @code{void *}, | |
4172 | @code{wait} would accept both kinds of arguments, but it would also | |
4173 | accept any other pointer type and this would make argument type checking | |
4174 | less useful. Instead, @code{<sys/wait.h>} might define the interface | |
4175 | as follows: | |
4176 | ||
4177 | @smallexample | |
4009f2e7 | 4178 | typedef union __attribute__ ((__transparent_union__)) |
c1f7febf RK |
4179 | @{ |
4180 | int *__ip; | |
4181 | union wait *__up; | |
4009f2e7 | 4182 | @} wait_status_ptr_t; |
c1f7febf RK |
4183 | |
4184 | pid_t wait (wait_status_ptr_t); | |
4185 | @end smallexample | |
4186 | ||
4187 | This interface allows either @code{int *} or @code{union wait *} | |
4188 | arguments to be passed, using the @code{int *} calling convention. | |
4189 | The program can call @code{wait} with arguments of either type: | |
4190 | ||
3ab51846 | 4191 | @smallexample |
c1f7febf RK |
4192 | int w1 () @{ int w; return wait (&w); @} |
4193 | int w2 () @{ union wait w; return wait (&w); @} | |
3ab51846 | 4194 | @end smallexample |
c1f7febf RK |
4195 | |
4196 | With this interface, @code{wait}'s implementation might look like this: | |
4197 | ||
3ab51846 | 4198 | @smallexample |
c1f7febf RK |
4199 | pid_t wait (wait_status_ptr_t p) |
4200 | @{ | |
4201 | return waitpid (-1, p.__ip, 0); | |
4202 | @} | |
3ab51846 | 4203 | @end smallexample |
d863830b JL |
4204 | |
4205 | @item unused | |
4206 | When attached to a type (including a @code{union} or a @code{struct}), | |
4207 | this attribute means that variables of that type are meant to appear | |
f0523f02 | 4208 | possibly unused. GCC will not produce a warning for any variables of |
d863830b JL |
4209 | that type, even if the variable appears to do nothing. This is often |
4210 | the case with lock or thread classes, which are usually defined and then | |
4211 | not referenced, but contain constructors and destructors that have | |
956d6950 | 4212 | nontrivial bookkeeping functions. |
d863830b | 4213 | |
e23bd218 IR |
4214 | @item deprecated |
4215 | The @code{deprecated} attribute results in a warning if the type | |
4216 | is used anywhere in the source file. This is useful when identifying | |
4217 | types that are expected to be removed in a future version of a program. | |
4218 | If possible, the warning also includes the location of the declaration | |
4219 | of the deprecated type, to enable users to easily find further | |
4220 | information about why the type is deprecated, or what they should do | |
4221 | instead. Note that the warnings only occur for uses and then only | |
adc9fe67 | 4222 | if the type is being applied to an identifier that itself is not being |
e23bd218 IR |
4223 | declared as deprecated. |
4224 | ||
4225 | @smallexample | |
4226 | typedef int T1 __attribute__ ((deprecated)); | |
4227 | T1 x; | |
4228 | typedef T1 T2; | |
4229 | T2 y; | |
4230 | typedef T1 T3 __attribute__ ((deprecated)); | |
4231 | T3 z __attribute__ ((deprecated)); | |
4232 | @end smallexample | |
4233 | ||
4234 | results in a warning on line 2 and 3 but not lines 4, 5, or 6. No | |
4235 | warning is issued for line 4 because T2 is not explicitly | |
4236 | deprecated. Line 5 has no warning because T3 is explicitly | |
4237 | deprecated. Similarly for line 6. | |
4238 | ||
4239 | The @code{deprecated} attribute can also be used for functions and | |
4240 | variables (@pxref{Function Attributes}, @pxref{Variable Attributes}.) | |
4241 | ||
d18b1ed8 OS |
4242 | @item may_alias |
4243 | Accesses to objects with types with this attribute are not subjected to | |
4244 | type-based alias analysis, but are instead assumed to be able to alias | |
4245 | any other type of objects, just like the @code{char} type. See | |
4246 | @option{-fstrict-aliasing} for more information on aliasing issues. | |
4247 | ||
4248 | Example of use: | |
4249 | ||
478c9e72 | 4250 | @smallexample |
d18b1ed8 OS |
4251 | typedef short __attribute__((__may_alias__)) short_a; |
4252 | ||
4253 | int | |
4254 | main (void) | |
4255 | @{ | |
4256 | int a = 0x12345678; | |
4257 | short_a *b = (short_a *) &a; | |
4258 | ||
4259 | b[1] = 0; | |
4260 | ||
4261 | if (a == 0x12345678) | |
4262 | abort(); | |
4263 | ||
4264 | exit(0); | |
4265 | @} | |
478c9e72 | 4266 | @end smallexample |
d18b1ed8 OS |
4267 | |
4268 | If you replaced @code{short_a} with @code{short} in the variable | |
4269 | declaration, the above program would abort when compiled with | |
4270 | @option{-fstrict-aliasing}, which is on by default at @option{-O2} or | |
4271 | above in recent GCC versions. | |
fe77449a | 4272 | |
b9e75696 | 4273 | @item visibility |
b9e75696 JM |
4274 | In C++, attribute visibility (@pxref{Function Attributes}) can also be |
4275 | applied to class, struct, union and enum types. Unlike other type | |
4276 | attributes, the attribute must appear between the initial keyword and | |
4277 | the name of the type; it cannot appear after the body of the type. | |
4278 | ||
b70f0f48 JM |
4279 | Note that the type visibility is applied to vague linkage entities |
4280 | associated with the class (vtable, typeinfo node, etc.). In | |
4281 | particular, if a class is thrown as an exception in one shared object | |
4282 | and caught in another, the class must have default visibility. | |
4283 | Otherwise the two shared objects will be unable to use the same | |
4284 | typeinfo node and exception handling will break. | |
4285 | ||
38bb2b65 SL |
4286 | @end table |
4287 | ||
04fb56d5 MM |
4288 | @subsection ARM Type Attributes |
4289 | ||
4290 | On those ARM targets that support @code{dllimport} (such as Symbian | |
f0eb93a8 | 4291 | OS), you can use the @code{notshared} attribute to indicate that the |
04fb56d5 | 4292 | virtual table and other similar data for a class should not be |
8a36672b | 4293 | exported from a DLL@. For example: |
04fb56d5 MM |
4294 | |
4295 | @smallexample | |
4296 | class __declspec(notshared) C @{ | |
4297 | public: | |
f0eb93a8 | 4298 | __declspec(dllimport) C(); |
04fb56d5 MM |
4299 | virtual void f(); |
4300 | @} | |
4301 | ||
4302 | __declspec(dllexport) | |
4303 | C::C() @{@} | |
4304 | @end smallexample | |
4305 | ||
4306 | In this code, @code{C::C} is exported from the current DLL, but the | |
4307 | virtual table for @code{C} is not exported. (You can use | |
4308 | @code{__attribute__} instead of @code{__declspec} if you prefer, but | |
4309 | most Symbian OS code uses @code{__declspec}.) | |
4310 | ||
63d0dca4 | 4311 | @anchor{i386 Type Attributes} |
fe77449a DR |
4312 | @subsection i386 Type Attributes |
4313 | ||
4314 | Two attributes are currently defined for i386 configurations: | |
38bb2b65 SL |
4315 | @code{ms_struct} and @code{gcc_struct}. |
4316 | ||
4317 | @table @code | |
fe77449a DR |
4318 | |
4319 | @item ms_struct | |
4320 | @itemx gcc_struct | |
4321 | @cindex @code{ms_struct} | |
4322 | @cindex @code{gcc_struct} | |
4323 | ||
4324 | If @code{packed} is used on a structure, or if bit-fields are used | |
4325 | it may be that the Microsoft ABI packs them differently | |
4326 | than GCC would normally pack them. Particularly when moving packed | |
4327 | data between functions compiled with GCC and the native Microsoft compiler | |
4328 | (either via function call or as data in a file), it may be necessary to access | |
4329 | either format. | |
4330 | ||
95fef11f | 4331 | Currently @option{-m[no-]ms-bitfields} is provided for the Microsoft Windows X86 |
fe77449a | 4332 | compilers to match the native Microsoft compiler. |
c1f7febf RK |
4333 | @end table |
4334 | ||
4335 | To specify multiple attributes, separate them by commas within the | |
4336 | double parentheses: for example, @samp{__attribute__ ((aligned (16), | |
4337 | packed))}. | |
4338 | ||
63d0dca4 DE |
4339 | @anchor{PowerPC Type Attributes} |
4340 | @subsection PowerPC Type Attributes | |
4341 | ||
4342 | Three attributes currently are defined for PowerPC configurations: | |
4343 | @code{altivec}, @code{ms_struct} and @code{gcc_struct}. | |
4344 | ||
38bb2b65 SL |
4345 | For full documentation of the @code{ms_struct} and @code{gcc_struct} |
4346 | attributes please see the documentation in @ref{i386 Type Attributes}. | |
63d0dca4 DE |
4347 | |
4348 | The @code{altivec} attribute allows one to declare AltiVec vector data | |
4349 | types supported by the AltiVec Programming Interface Manual. The | |
4350 | attribute requires an argument to specify one of three vector types: | |
4351 | @code{vector__}, @code{pixel__} (always followed by unsigned short), | |
4352 | and @code{bool__} (always followed by unsigned). | |
4353 | ||
4354 | @smallexample | |
4355 | __attribute__((altivec(vector__))) | |
4356 | __attribute__((altivec(pixel__))) unsigned short | |
4357 | __attribute__((altivec(bool__))) unsigned | |
4358 | @end smallexample | |
4359 | ||
4360 | These attributes mainly are intended to support the @code{__vector}, | |
4361 | @code{__pixel}, and @code{__bool} AltiVec keywords. | |
4362 | ||
85d9c13c TS |
4363 | @anchor{SPU Type Attributes} |
4364 | @subsection SPU Type Attributes | |
4365 | ||
4366 | The SPU supports the @code{spu_vector} attribute for types. This attribute | |
4367 | allows one to declare vector data types supported by the Sony/Toshiba/IBM SPU | |
4368 | Language Extensions Specification. It is intended to support the | |
4369 | @code{__vector} keyword. | |
4370 | ||
4371 | ||
c1f7febf RK |
4372 | @node Inline |
4373 | @section An Inline Function is As Fast As a Macro | |
4374 | @cindex inline functions | |
4375 | @cindex integrating function code | |
4376 | @cindex open coding | |
4377 | @cindex macros, inline alternative | |
4378 | ||
0a052b16 GK |
4379 | By declaring a function inline, you can direct GCC to make |
4380 | calls to that function faster. One way GCC can achieve this is to | |
c1f7febf RK |
4381 | integrate that function's code into the code for its callers. This |
4382 | makes execution faster by eliminating the function-call overhead; in | |
0a052b16 GK |
4383 | addition, if any of the actual argument values are constant, their |
4384 | known values may permit simplifications at compile time so that not | |
4385 | all of the inline function's code needs to be included. The effect on | |
4386 | code size is less predictable; object code may be larger or smaller | |
4387 | with function inlining, depending on the particular case. You can | |
4388 | also direct GCC to try to integrate all ``simple enough'' functions | |
4389 | into their callers with the option @option{-finline-functions}. | |
4390 | ||
4391 | GCC implements three different semantics of declaring a function | |
da1c7394 ILT |
4392 | inline. One is available with @option{-std=gnu89} or |
4393 | @option{-fgnu89-inline} or when @code{gnu_inline} attribute is present | |
4394 | on all inline declarations, another when @option{-std=c99} or | |
4395 | @option{-std=gnu99} (without @option{-fgnu89-inline}), and the third | |
4396 | is used when compiling C++. | |
4b404517 | 4397 | |
c1f7febf RK |
4398 | To declare a function inline, use the @code{inline} keyword in its |
4399 | declaration, like this: | |
4400 | ||
3ab51846 | 4401 | @smallexample |
0a052b16 | 4402 | static inline int |
c1f7febf RK |
4403 | inc (int *a) |
4404 | @{ | |
4405 | (*a)++; | |
4406 | @} | |
3ab51846 | 4407 | @end smallexample |
c1f7febf | 4408 | |
0a052b16 GK |
4409 | If you are writing a header file to be included in ISO C89 programs, write |
4410 | @code{__inline__} instead of @code{inline}. @xref{Alternate Keywords}. | |
247b14bd | 4411 | |
0a052b16 GK |
4412 | The three types of inlining behave similarly in two important cases: |
4413 | when the @code{inline} keyword is used on a @code{static} function, | |
4414 | like the example above, and when a function is first declared without | |
4415 | using the @code{inline} keyword and then is defined with | |
4416 | @code{inline}, like this: | |
c1f7febf | 4417 | |
0a052b16 GK |
4418 | @smallexample |
4419 | extern int inc (int *a); | |
4420 | inline int | |
4421 | inc (int *a) | |
4422 | @{ | |
4423 | (*a)++; | |
4424 | @} | |
4425 | @end smallexample | |
c1f7febf | 4426 | |
0a052b16 GK |
4427 | In both of these common cases, the program behaves the same as if you |
4428 | had not used the @code{inline} keyword, except for its speed. | |
c1f7febf RK |
4429 | |
4430 | @cindex inline functions, omission of | |
84330467 | 4431 | @opindex fkeep-inline-functions |
c1f7febf RK |
4432 | When a function is both inline and @code{static}, if all calls to the |
4433 | function are integrated into the caller, and the function's address is | |
4434 | never used, then the function's own assembler code is never referenced. | |
f0523f02 | 4435 | In this case, GCC does not actually output assembler code for the |
84330467 | 4436 | function, unless you specify the option @option{-fkeep-inline-functions}. |
c1f7febf RK |
4437 | Some calls cannot be integrated for various reasons (in particular, |
4438 | calls that precede the function's definition cannot be integrated, and | |
4439 | neither can recursive calls within the definition). If there is a | |
4440 | nonintegrated call, then the function is compiled to assembler code as | |
4441 | usual. The function must also be compiled as usual if the program | |
4442 | refers to its address, because that can't be inlined. | |
4443 | ||
0a052b16 GK |
4444 | @opindex Winline |
4445 | Note that certain usages in a function definition can make it unsuitable | |
4446 | for inline substitution. Among these usages are: use of varargs, use of | |
4447 | alloca, use of variable sized data types (@pxref{Variable Length}), | |
4448 | use of computed goto (@pxref{Labels as Values}), use of nonlocal goto, | |
4449 | and nested functions (@pxref{Nested Functions}). Using @option{-Winline} | |
4450 | will warn when a function marked @code{inline} could not be substituted, | |
4451 | and will give the reason for the failure. | |
4452 | ||
4453 | @cindex automatic @code{inline} for C++ member fns | |
4454 | @cindex @code{inline} automatic for C++ member fns | |
4455 | @cindex member fns, automatically @code{inline} | |
4456 | @cindex C++ member fns, automatically @code{inline} | |
4457 | @opindex fno-default-inline | |
4458 | As required by ISO C++, GCC considers member functions defined within | |
4459 | the body of a class to be marked inline even if they are | |
4460 | not explicitly declared with the @code{inline} keyword. You can | |
4461 | override this with @option{-fno-default-inline}; @pxref{C++ Dialect | |
4462 | Options,,Options Controlling C++ Dialect}. | |
4463 | ||
4464 | GCC does not inline any functions when not optimizing unless you specify | |
4465 | the @samp{always_inline} attribute for the function, like this: | |
4466 | ||
4467 | @smallexample | |
4468 | /* @r{Prototype.} */ | |
4469 | inline void foo (const char) __attribute__((always_inline)); | |
4470 | @end smallexample | |
4471 | ||
4472 | The remainder of this section is specific to GNU C89 inlining. | |
4473 | ||
c1f7febf RK |
4474 | @cindex non-static inline function |
4475 | When an inline function is not @code{static}, then the compiler must assume | |
4476 | that there may be calls from other source files; since a global symbol can | |
4477 | be defined only once in any program, the function must not be defined in | |
4478 | the other source files, so the calls therein cannot be integrated. | |
4479 | Therefore, a non-@code{static} inline function is always compiled on its | |
4480 | own in the usual fashion. | |
4481 | ||
4482 | If you specify both @code{inline} and @code{extern} in the function | |
4483 | definition, then the definition is used only for inlining. In no case | |
4484 | is the function compiled on its own, not even if you refer to its | |
4485 | address explicitly. Such an address becomes an external reference, as | |
4486 | if you had only declared the function, and had not defined it. | |
4487 | ||
4488 | This combination of @code{inline} and @code{extern} has almost the | |
4489 | effect of a macro. The way to use it is to put a function definition in | |
4490 | a header file with these keywords, and put another copy of the | |
4491 | definition (lacking @code{inline} and @code{extern}) in a library file. | |
4492 | The definition in the header file will cause most calls to the function | |
4493 | to be inlined. If any uses of the function remain, they will refer to | |
4494 | the single copy in the library. | |
4495 | ||
c1f7febf RK |
4496 | @node Extended Asm |
4497 | @section Assembler Instructions with C Expression Operands | |
4498 | @cindex extended @code{asm} | |
4499 | @cindex @code{asm} expressions | |
4500 | @cindex assembler instructions | |
4501 | @cindex registers | |
4502 | ||
c85f7c16 JL |
4503 | In an assembler instruction using @code{asm}, you can specify the |
4504 | operands of the instruction using C expressions. This means you need not | |
4505 | guess which registers or memory locations will contain the data you want | |
c1f7febf RK |
4506 | to use. |
4507 | ||
c85f7c16 JL |
4508 | You must specify an assembler instruction template much like what |
4509 | appears in a machine description, plus an operand constraint string for | |
4510 | each operand. | |
c1f7febf RK |
4511 | |
4512 | For example, here is how to use the 68881's @code{fsinx} instruction: | |
4513 | ||
3ab51846 | 4514 | @smallexample |
c1f7febf | 4515 | asm ("fsinx %1,%0" : "=f" (result) : "f" (angle)); |
3ab51846 | 4516 | @end smallexample |
c1f7febf RK |
4517 | |
4518 | @noindent | |
4519 | Here @code{angle} is the C expression for the input operand while | |
4520 | @code{result} is that of the output operand. Each has @samp{"f"} as its | |
c85f7c16 JL |
4521 | operand constraint, saying that a floating point register is required. |
4522 | The @samp{=} in @samp{=f} indicates that the operand is an output; all | |
4523 | output operands' constraints must use @samp{=}. The constraints use the | |
4524 | same language used in the machine description (@pxref{Constraints}). | |
4525 | ||
4526 | Each operand is described by an operand-constraint string followed by | |
4527 | the C expression in parentheses. A colon separates the assembler | |
4528 | template from the first output operand and another separates the last | |
4529 | output operand from the first input, if any. Commas separate the | |
84b72302 RH |
4530 | operands within each group. The total number of operands is currently |
4531 | limited to 30; this limitation may be lifted in some future version of | |
8a36672b | 4532 | GCC@. |
c85f7c16 JL |
4533 | |
4534 | If there are no output operands but there are input operands, you must | |
4535 | place two consecutive colons surrounding the place where the output | |
c1f7febf RK |
4536 | operands would go. |
4537 | ||
84b72302 RH |
4538 | As of GCC version 3.1, it is also possible to specify input and output |
4539 | operands using symbolic names which can be referenced within the | |
4540 | assembler code. These names are specified inside square brackets | |
4541 | preceding the constraint string, and can be referenced inside the | |
4542 | assembler code using @code{%[@var{name}]} instead of a percentage sign | |
4543 | followed by the operand number. Using named operands the above example | |
4544 | could look like: | |
4545 | ||
3ab51846 | 4546 | @smallexample |
84b72302 RH |
4547 | asm ("fsinx %[angle],%[output]" |
4548 | : [output] "=f" (result) | |
4549 | : [angle] "f" (angle)); | |
3ab51846 | 4550 | @end smallexample |
84b72302 RH |
4551 | |
4552 | @noindent | |
4553 | Note that the symbolic operand names have no relation whatsoever to | |
4554 | other C identifiers. You may use any name you like, even those of | |
64c18e57 | 4555 | existing C symbols, but you must ensure that no two operands within the same |
84b72302 RH |
4556 | assembler construct use the same symbolic name. |
4557 | ||
c1f7febf | 4558 | Output operand expressions must be lvalues; the compiler can check this. |
c85f7c16 JL |
4559 | The input operands need not be lvalues. The compiler cannot check |
4560 | whether the operands have data types that are reasonable for the | |
4561 | instruction being executed. It does not parse the assembler instruction | |
4562 | template and does not know what it means or even whether it is valid | |
4563 | assembler input. The extended @code{asm} feature is most often used for | |
4564 | machine instructions the compiler itself does not know exist. If | |
4565 | the output expression cannot be directly addressed (for example, it is a | |
f0523f02 | 4566 | bit-field), your constraint must allow a register. In that case, GCC |
c85f7c16 JL |
4567 | will use the register as the output of the @code{asm}, and then store |
4568 | that register into the output. | |
4569 | ||
f0523f02 | 4570 | The ordinary output operands must be write-only; GCC will assume that |
c85f7c16 JL |
4571 | the values in these operands before the instruction are dead and need |
4572 | not be generated. Extended asm supports input-output or read-write | |
4573 | operands. Use the constraint character @samp{+} to indicate such an | |
373a04f1 JM |
4574 | operand and list it with the output operands. You should only use |
4575 | read-write operands when the constraints for the operand (or the | |
4576 | operand in which only some of the bits are to be changed) allow a | |
4577 | register. | |
4578 | ||
4579 | You may, as an alternative, logically split its function into two | |
4580 | separate operands, one input operand and one write-only output | |
4581 | operand. The connection between them is expressed by constraints | |
4582 | which say they need to be in the same location when the instruction | |
4583 | executes. You can use the same C expression for both operands, or | |
4584 | different expressions. For example, here we write the (fictitious) | |
4585 | @samp{combine} instruction with @code{bar} as its read-only source | |
4586 | operand and @code{foo} as its read-write destination: | |
c1f7febf | 4587 | |
3ab51846 | 4588 | @smallexample |
c1f7febf | 4589 | asm ("combine %2,%0" : "=r" (foo) : "0" (foo), "g" (bar)); |
3ab51846 | 4590 | @end smallexample |
c1f7febf RK |
4591 | |
4592 | @noindent | |
c85f7c16 | 4593 | The constraint @samp{"0"} for operand 1 says that it must occupy the |
84b72302 RH |
4594 | same location as operand 0. A number in constraint is allowed only in |
4595 | an input operand and it must refer to an output operand. | |
c1f7febf | 4596 | |
84b72302 | 4597 | Only a number in the constraint can guarantee that one operand will be in |
c85f7c16 JL |
4598 | the same place as another. The mere fact that @code{foo} is the value |
4599 | of both operands is not enough to guarantee that they will be in the | |
4600 | same place in the generated assembler code. The following would not | |
4601 | work reliably: | |
c1f7febf | 4602 | |
3ab51846 | 4603 | @smallexample |
c1f7febf | 4604 | asm ("combine %2,%0" : "=r" (foo) : "r" (foo), "g" (bar)); |
3ab51846 | 4605 | @end smallexample |
c1f7febf RK |
4606 | |
4607 | Various optimizations or reloading could cause operands 0 and 1 to be in | |
f0523f02 | 4608 | different registers; GCC knows no reason not to do so. For example, the |
c1f7febf RK |
4609 | compiler might find a copy of the value of @code{foo} in one register and |
4610 | use it for operand 1, but generate the output operand 0 in a different | |
4611 | register (copying it afterward to @code{foo}'s own address). Of course, | |
4612 | since the register for operand 1 is not even mentioned in the assembler | |
f0523f02 | 4613 | code, the result will not work, but GCC can't tell that. |
c1f7febf | 4614 | |
84b72302 RH |
4615 | As of GCC version 3.1, one may write @code{[@var{name}]} instead of |
4616 | the operand number for a matching constraint. For example: | |
4617 | ||
3ab51846 | 4618 | @smallexample |
84b72302 RH |
4619 | asm ("cmoveq %1,%2,%[result]" |
4620 | : [result] "=r"(result) | |
4621 | : "r" (test), "r"(new), "[result]"(old)); | |
3ab51846 | 4622 | @end smallexample |
84b72302 | 4623 | |
805c33df HPN |
4624 | Sometimes you need to make an @code{asm} operand be a specific register, |
4625 | but there's no matching constraint letter for that register @emph{by | |
4626 | itself}. To force the operand into that register, use a local variable | |
4627 | for the operand and specify the register in the variable declaration. | |
4628 | @xref{Explicit Reg Vars}. Then for the @code{asm} operand, use any | |
4629 | register constraint letter that matches the register: | |
4630 | ||
4631 | @smallexample | |
4632 | register int *p1 asm ("r0") = @dots{}; | |
4633 | register int *p2 asm ("r1") = @dots{}; | |
4634 | register int *result asm ("r0"); | |
4635 | asm ("sysint" : "=r" (result) : "0" (p1), "r" (p2)); | |
4636 | @end smallexample | |
4637 | ||
b55d5746 HPN |
4638 | @anchor{Example of asm with clobbered asm reg} |
4639 | In the above example, beware that a register that is call-clobbered by | |
4640 | the target ABI will be overwritten by any function call in the | |
4641 | assignment, including library calls for arithmetic operators. | |
4642 | Assuming it is a call-clobbered register, this may happen to @code{r0} | |
4643 | above by the assignment to @code{p2}. If you have to use such a | |
4644 | register, use temporary variables for expressions between the register | |
4645 | assignment and use: | |
4646 | ||
4647 | @smallexample | |
4648 | int t1 = @dots{}; | |
4649 | register int *p1 asm ("r0") = @dots{}; | |
4650 | register int *p2 asm ("r1") = t1; | |
4651 | register int *result asm ("r0"); | |
4652 | asm ("sysint" : "=r" (result) : "0" (p1), "r" (p2)); | |
4653 | @end smallexample | |
4654 | ||
c85f7c16 JL |
4655 | Some instructions clobber specific hard registers. To describe this, |
4656 | write a third colon after the input operands, followed by the names of | |
4657 | the clobbered hard registers (given as strings). Here is a realistic | |
4658 | example for the VAX: | |
c1f7febf | 4659 | |
3ab51846 | 4660 | @smallexample |
c1f7febf | 4661 | asm volatile ("movc3 %0,%1,%2" |
12bcfaa1 | 4662 | : /* @r{no outputs} */ |
c1f7febf RK |
4663 | : "g" (from), "g" (to), "g" (count) |
4664 | : "r0", "r1", "r2", "r3", "r4", "r5"); | |
3ab51846 | 4665 | @end smallexample |
c1f7febf | 4666 | |
c5c76735 JL |
4667 | You may not write a clobber description in a way that overlaps with an |
4668 | input or output operand. For example, you may not have an operand | |
4669 | describing a register class with one member if you mention that register | |
acb5d088 HPN |
4670 | in the clobber list. Variables declared to live in specific registers |
4671 | (@pxref{Explicit Reg Vars}), and used as asm input or output operands must | |
4672 | have no part mentioned in the clobber description. | |
4673 | There is no way for you to specify that an input | |
c5c76735 JL |
4674 | operand is modified without also specifying it as an output |
4675 | operand. Note that if all the output operands you specify are for this | |
4676 | purpose (and hence unused), you will then also need to specify | |
4677 | @code{volatile} for the @code{asm} construct, as described below, to | |
f0523f02 | 4678 | prevent GCC from deleting the @code{asm} statement as unused. |
8fe1938e | 4679 | |
c1f7febf | 4680 | If you refer to a particular hardware register from the assembler code, |
c85f7c16 JL |
4681 | you will probably have to list the register after the third colon to |
4682 | tell the compiler the register's value is modified. In some assemblers, | |
4683 | the register names begin with @samp{%}; to produce one @samp{%} in the | |
4684 | assembler code, you must write @samp{%%} in the input. | |
4685 | ||
4686 | If your assembler instruction can alter the condition code register, add | |
f0523f02 | 4687 | @samp{cc} to the list of clobbered registers. GCC on some machines |
c85f7c16 JL |
4688 | represents the condition codes as a specific hardware register; |
4689 | @samp{cc} serves to name this register. On other machines, the | |
4690 | condition code is handled differently, and specifying @samp{cc} has no | |
4691 | effect. But it is valid no matter what the machine. | |
c1f7febf | 4692 | |
bbf5a54d | 4693 | If your assembler instructions access memory in an unpredictable |
c85f7c16 | 4694 | fashion, add @samp{memory} to the list of clobbered registers. This |
bbf5a54d AJ |
4695 | will cause GCC to not keep memory values cached in registers across the |
4696 | assembler instruction and not optimize stores or loads to that memory. | |
4697 | You will also want to add the @code{volatile} keyword if the memory | |
4698 | affected is not listed in the inputs or outputs of the @code{asm}, as | |
4699 | the @samp{memory} clobber does not count as a side-effect of the | |
4700 | @code{asm}. If you know how large the accessed memory is, you can add | |
4701 | it as input or output but if this is not known, you should add | |
4702 | @samp{memory}. As an example, if you access ten bytes of a string, you | |
4703 | can use a memory input like: | |
4704 | ||
cd1a8088 | 4705 | @smallexample |
bbf5a54d | 4706 | @{"m"( (@{ struct @{ char x[10]; @} *p = (void *)ptr ; *p; @}) )@}. |
cd1a8088 | 4707 | @end smallexample |
bbf5a54d AJ |
4708 | |
4709 | Note that in the following example the memory input is necessary, | |
4710 | otherwise GCC might optimize the store to @code{x} away: | |
cd1a8088 | 4711 | @smallexample |
bbf5a54d AJ |
4712 | int foo () |
4713 | @{ | |
4714 | int x = 42; | |
4715 | int *y = &x; | |
4716 | int result; | |
4717 | asm ("magic stuff accessing an 'int' pointed to by '%1'" | |
4718 | "=&d" (r) : "a" (y), "m" (*y)); | |
f0eb93a8 | 4719 | return result; |
bbf5a54d | 4720 | @} |
cd1a8088 | 4721 | @end smallexample |
c1f7febf | 4722 | |
c85f7c16 | 4723 | You can put multiple assembler instructions together in a single |
8720914b HPN |
4724 | @code{asm} template, separated by the characters normally used in assembly |
4725 | code for the system. A combination that works in most places is a newline | |
4726 | to break the line, plus a tab character to move to the instruction field | |
4727 | (written as @samp{\n\t}). Sometimes semicolons can be used, if the | |
4728 | assembler allows semicolons as a line-breaking character. Note that some | |
4729 | assembler dialects use semicolons to start a comment. | |
4730 | The input operands are guaranteed not to use any of the clobbered | |
c85f7c16 JL |
4731 | registers, and neither will the output operands' addresses, so you can |
4732 | read and write the clobbered registers as many times as you like. Here | |
4733 | is an example of multiple instructions in a template; it assumes the | |
4734 | subroutine @code{_foo} accepts arguments in registers 9 and 10: | |
c1f7febf | 4735 | |
3ab51846 | 4736 | @smallexample |
8720914b | 4737 | asm ("movl %0,r9\n\tmovl %1,r10\n\tcall _foo" |
c1f7febf RK |
4738 | : /* no outputs */ |
4739 | : "g" (from), "g" (to) | |
4740 | : "r9", "r10"); | |
3ab51846 | 4741 | @end smallexample |
c1f7febf | 4742 | |
f0523f02 | 4743 | Unless an output operand has the @samp{&} constraint modifier, GCC |
c85f7c16 JL |
4744 | may allocate it in the same register as an unrelated input operand, on |
4745 | the assumption the inputs are consumed before the outputs are produced. | |
c1f7febf RK |
4746 | This assumption may be false if the assembler code actually consists of |
4747 | more than one instruction. In such a case, use @samp{&} for each output | |
c85f7c16 | 4748 | operand that may not overlap an input. @xref{Modifiers}. |
c1f7febf | 4749 | |
c85f7c16 JL |
4750 | If you want to test the condition code produced by an assembler |
4751 | instruction, you must include a branch and a label in the @code{asm} | |
4752 | construct, as follows: | |
c1f7febf | 4753 | |
3ab51846 | 4754 | @smallexample |
8720914b | 4755 | asm ("clr %0\n\tfrob %1\n\tbeq 0f\n\tmov #1,%0\n0:" |
c1f7febf RK |
4756 | : "g" (result) |
4757 | : "g" (input)); | |
3ab51846 | 4758 | @end smallexample |
c1f7febf RK |
4759 | |
4760 | @noindent | |
4761 | This assumes your assembler supports local labels, as the GNU assembler | |
4762 | and most Unix assemblers do. | |
4763 | ||
4764 | Speaking of labels, jumps from one @code{asm} to another are not | |
c85f7c16 JL |
4765 | supported. The compiler's optimizers do not know about these jumps, and |
4766 | therefore they cannot take account of them when deciding how to | |
c1f7febf RK |
4767 | optimize. |
4768 | ||
4769 | @cindex macros containing @code{asm} | |
4770 | Usually the most convenient way to use these @code{asm} instructions is to | |
4771 | encapsulate them in macros that look like functions. For example, | |
4772 | ||
3ab51846 | 4773 | @smallexample |
c1f7febf RK |
4774 | #define sin(x) \ |
4775 | (@{ double __value, __arg = (x); \ | |
4776 | asm ("fsinx %1,%0": "=f" (__value): "f" (__arg)); \ | |
4777 | __value; @}) | |
3ab51846 | 4778 | @end smallexample |
c1f7febf RK |
4779 | |
4780 | @noindent | |
4781 | Here the variable @code{__arg} is used to make sure that the instruction | |
4782 | operates on a proper @code{double} value, and to accept only those | |
4783 | arguments @code{x} which can convert automatically to a @code{double}. | |
4784 | ||
c85f7c16 JL |
4785 | Another way to make sure the instruction operates on the correct data |
4786 | type is to use a cast in the @code{asm}. This is different from using a | |
c1f7febf RK |
4787 | variable @code{__arg} in that it converts more different types. For |
4788 | example, if the desired type were @code{int}, casting the argument to | |
4789 | @code{int} would accept a pointer with no complaint, while assigning the | |
4790 | argument to an @code{int} variable named @code{__arg} would warn about | |
4791 | using a pointer unless the caller explicitly casts it. | |
4792 | ||
f0523f02 | 4793 | If an @code{asm} has output operands, GCC assumes for optimization |
c85f7c16 JL |
4794 | purposes the instruction has no side effects except to change the output |
4795 | operands. This does not mean instructions with a side effect cannot be | |
4796 | used, but you must be careful, because the compiler may eliminate them | |
4797 | if the output operands aren't used, or move them out of loops, or | |
4798 | replace two with one if they constitute a common subexpression. Also, | |
4799 | if your instruction does have a side effect on a variable that otherwise | |
4800 | appears not to change, the old value of the variable may be reused later | |
4801 | if it happens to be found in a register. | |
c1f7febf | 4802 | |
2f59e40e DJ |
4803 | You can prevent an @code{asm} instruction from being deleted |
4804 | by writing the keyword @code{volatile} after | |
c1f7febf RK |
4805 | the @code{asm}. For example: |
4806 | ||
3ab51846 | 4807 | @smallexample |
310668e8 JM |
4808 | #define get_and_set_priority(new) \ |
4809 | (@{ int __old; \ | |
4810 | asm volatile ("get_and_set_priority %0, %1" \ | |
4811 | : "=g" (__old) : "g" (new)); \ | |
c85f7c16 | 4812 | __old; @}) |
3ab51846 | 4813 | @end smallexample |
c1f7febf RK |
4814 | |
4815 | @noindent | |
e71b34aa MM |
4816 | The @code{volatile} keyword indicates that the instruction has |
4817 | important side-effects. GCC will not delete a volatile @code{asm} if | |
4818 | it is reachable. (The instruction can still be deleted if GCC can | |
4819 | prove that control-flow will never reach the location of the | |
f0eb93a8 | 4820 | instruction.) Note that even a volatile @code{asm} instruction |
2f59e40e | 4821 | can be moved relative to other code, including across jump |
f0eb93a8 JM |
4822 | instructions. For example, on many targets there is a system |
4823 | register which can be set to control the rounding mode of | |
2f59e40e DJ |
4824 | floating point operations. You might try |
4825 | setting it with a volatile @code{asm}, like this PowerPC example: | |
e71b34aa | 4826 | |
3ab51846 | 4827 | @smallexample |
2f59e40e DJ |
4828 | asm volatile("mtfsf 255,%0" : : "f" (fpenv)); |
4829 | sum = x + y; | |
3ab51846 | 4830 | @end smallexample |
e71b34aa | 4831 | |
ebb48a4d | 4832 | @noindent |
2f59e40e DJ |
4833 | This will not work reliably, as the compiler may move the addition back |
4834 | before the volatile @code{asm}. To make it work you need to add an | |
4835 | artificial dependency to the @code{asm} referencing a variable in the code | |
4836 | you don't want moved, for example: | |
4837 | ||
4838 | @smallexample | |
4839 | asm volatile ("mtfsf 255,%1" : "=X"(sum): "f"(fpenv)); | |
4840 | sum = x + y; | |
4841 | @end smallexample | |
4842 | ||
4843 | Similarly, you can't expect a | |
4844 | sequence of volatile @code{asm} instructions to remain perfectly | |
4845 | consecutive. If you want consecutive output, use a single @code{asm}. | |
4846 | Also, GCC will perform some optimizations across a volatile @code{asm} | |
4847 | instruction; GCC does not ``forget everything'' when it encounters | |
4848 | a volatile @code{asm} instruction the way some other compilers do. | |
4849 | ||
4850 | An @code{asm} instruction without any output operands will be treated | |
4851 | identically to a volatile @code{asm} instruction. | |
c1f7febf RK |
4852 | |
4853 | It is a natural idea to look for a way to give access to the condition | |
4854 | code left by the assembler instruction. However, when we attempted to | |
4855 | implement this, we found no way to make it work reliably. The problem | |
4856 | is that output operands might need reloading, which would result in | |
4857 | additional following ``store'' instructions. On most machines, these | |
4858 | instructions would alter the condition code before there was time to | |
4859 | test it. This problem doesn't arise for ordinary ``test'' and | |
4860 | ``compare'' instructions because they don't have any output operands. | |
4861 | ||
eda3fbbe GB |
4862 | For reasons similar to those described above, it is not possible to give |
4863 | an assembler instruction access to the condition code left by previous | |
4864 | instructions. | |
4865 | ||
5490d604 | 4866 | If you are writing a header file that should be includable in ISO C |
c1f7febf RK |
4867 | programs, write @code{__asm__} instead of @code{asm}. @xref{Alternate |
4868 | Keywords}. | |
4869 | ||
ece7fc1c RE |
4870 | @subsection Size of an @code{asm} |
4871 | ||
4872 | Some targets require that GCC track the size of each instruction used in | |
4873 | order to generate correct code. Because the final length of an | |
4874 | @code{asm} is only known by the assembler, GCC must make an estimate as | |
4875 | to how big it will be. The estimate is formed by counting the number of | |
4876 | statements in the pattern of the @code{asm} and multiplying that by the | |
4877 | length of the longest instruction on that processor. Statements in the | |
4878 | @code{asm} are identified by newline characters and whatever statement | |
4879 | separator characters are supported by the assembler; on most processors | |
4880 | this is the `@code{;}' character. | |
4881 | ||
4882 | Normally, GCC's estimate is perfectly adequate to ensure that correct | |
4883 | code is generated, but it is possible to confuse the compiler if you use | |
4884 | pseudo instructions or assembler macros that expand into multiple real | |
4885 | instructions or if you use assembler directives that expand to more | |
4886 | space in the object file than would be needed for a single instruction. | |
4887 | If this happens then the assembler will produce a diagnostic saying that | |
4888 | a label is unreachable. | |
4889 | ||
fe0ce426 JH |
4890 | @subsection i386 floating point asm operands |
4891 | ||
4892 | There are several rules on the usage of stack-like regs in | |
4893 | asm_operands insns. These rules apply only to the operands that are | |
4894 | stack-like regs: | |
4895 | ||
4896 | @enumerate | |
4897 | @item | |
4898 | Given a set of input regs that die in an asm_operands, it is | |
4899 | necessary to know which are implicitly popped by the asm, and | |
4900 | which must be explicitly popped by gcc. | |
4901 | ||
4902 | An input reg that is implicitly popped by the asm must be | |
4903 | explicitly clobbered, unless it is constrained to match an | |
4904 | output operand. | |
4905 | ||
4906 | @item | |
4907 | For any input reg that is implicitly popped by an asm, it is | |
4908 | necessary to know how to adjust the stack to compensate for the pop. | |
4909 | If any non-popped input is closer to the top of the reg-stack than | |
4910 | the implicitly popped reg, it would not be possible to know what the | |
84330467 | 4911 | stack looked like---it's not clear how the rest of the stack ``slides |
fe0ce426 JH |
4912 | up''. |
4913 | ||
4914 | All implicitly popped input regs must be closer to the top of | |
4915 | the reg-stack than any input that is not implicitly popped. | |
4916 | ||
4917 | It is possible that if an input dies in an insn, reload might | |
4918 | use the input reg for an output reload. Consider this example: | |
4919 | ||
3ab51846 | 4920 | @smallexample |
fe0ce426 | 4921 | asm ("foo" : "=t" (a) : "f" (b)); |
3ab51846 | 4922 | @end smallexample |
fe0ce426 JH |
4923 | |
4924 | This asm says that input B is not popped by the asm, and that | |
c771326b | 4925 | the asm pushes a result onto the reg-stack, i.e., the stack is one |
fe0ce426 JH |
4926 | deeper after the asm than it was before. But, it is possible that |
4927 | reload will think that it can use the same reg for both the input and | |
4928 | the output, if input B dies in this insn. | |
4929 | ||
4930 | If any input operand uses the @code{f} constraint, all output reg | |
4931 | constraints must use the @code{&} earlyclobber. | |
4932 | ||
4933 | The asm above would be written as | |
4934 | ||
3ab51846 | 4935 | @smallexample |
fe0ce426 | 4936 | asm ("foo" : "=&t" (a) : "f" (b)); |
3ab51846 | 4937 | @end smallexample |
fe0ce426 JH |
4938 | |
4939 | @item | |
4940 | Some operands need to be in particular places on the stack. All | |
84330467 | 4941 | output operands fall in this category---there is no other way to |
fe0ce426 JH |
4942 | know which regs the outputs appear in unless the user indicates |
4943 | this in the constraints. | |
4944 | ||
4945 | Output operands must specifically indicate which reg an output | |
4946 | appears in after an asm. @code{=f} is not allowed: the operand | |
4947 | constraints must select a class with a single reg. | |
4948 | ||
4949 | @item | |
4950 | Output operands may not be ``inserted'' between existing stack regs. | |
4951 | Since no 387 opcode uses a read/write operand, all output operands | |
4952 | are dead before the asm_operands, and are pushed by the asm_operands. | |
4953 | It makes no sense to push anywhere but the top of the reg-stack. | |
4954 | ||
4955 | Output operands must start at the top of the reg-stack: output | |
4956 | operands may not ``skip'' a reg. | |
4957 | ||
4958 | @item | |
4959 | Some asm statements may need extra stack space for internal | |
4960 | calculations. This can be guaranteed by clobbering stack registers | |
4961 | unrelated to the inputs and outputs. | |
4962 | ||
4963 | @end enumerate | |
4964 | ||
4965 | Here are a couple of reasonable asms to want to write. This asm | |
4966 | takes one input, which is internally popped, and produces two outputs. | |
4967 | ||
3ab51846 | 4968 | @smallexample |
fe0ce426 | 4969 | asm ("fsincos" : "=t" (cos), "=u" (sin) : "0" (inp)); |
3ab51846 | 4970 | @end smallexample |
fe0ce426 JH |
4971 | |
4972 | This asm takes two inputs, which are popped by the @code{fyl2xp1} opcode, | |
4973 | and replaces them with one output. The user must code the @code{st(1)} | |
4974 | clobber for reg-stack.c to know that @code{fyl2xp1} pops both inputs. | |
4975 | ||
3ab51846 | 4976 | @smallexample |
fe0ce426 | 4977 | asm ("fyl2xp1" : "=t" (result) : "0" (x), "u" (y) : "st(1)"); |
3ab51846 | 4978 | @end smallexample |
fe0ce426 | 4979 | |
c1f7febf | 4980 | @include md.texi |
c1f7febf RK |
4981 | |
4982 | @node Asm Labels | |
4983 | @section Controlling Names Used in Assembler Code | |
4984 | @cindex assembler names for identifiers | |
4985 | @cindex names used in assembler code | |
4986 | @cindex identifiers, names in assembler code | |
4987 | ||
4988 | You can specify the name to be used in the assembler code for a C | |
4989 | function or variable by writing the @code{asm} (or @code{__asm__}) | |
4990 | keyword after the declarator as follows: | |
4991 | ||
3ab51846 | 4992 | @smallexample |
c1f7febf | 4993 | int foo asm ("myfoo") = 2; |
3ab51846 | 4994 | @end smallexample |
c1f7febf RK |
4995 | |
4996 | @noindent | |
4997 | This specifies that the name to be used for the variable @code{foo} in | |
4998 | the assembler code should be @samp{myfoo} rather than the usual | |
4999 | @samp{_foo}. | |
5000 | ||
5001 | On systems where an underscore is normally prepended to the name of a C | |
5002 | function or variable, this feature allows you to define names for the | |
5003 | linker that do not start with an underscore. | |
5004 | ||
0adc3c19 MM |
5005 | It does not make sense to use this feature with a non-static local |
5006 | variable since such variables do not have assembler names. If you are | |
5007 | trying to put the variable in a particular register, see @ref{Explicit | |
5008 | Reg Vars}. GCC presently accepts such code with a warning, but will | |
5009 | probably be changed to issue an error, rather than a warning, in the | |
5010 | future. | |
5011 | ||
c1f7febf RK |
5012 | You cannot use @code{asm} in this way in a function @emph{definition}; but |
5013 | you can get the same effect by writing a declaration for the function | |
5014 | before its definition and putting @code{asm} there, like this: | |
5015 | ||
3ab51846 | 5016 | @smallexample |
c1f7febf RK |
5017 | extern func () asm ("FUNC"); |
5018 | ||
5019 | func (x, y) | |
5020 | int x, y; | |
0d893a63 | 5021 | /* @r{@dots{}} */ |
3ab51846 | 5022 | @end smallexample |
c1f7febf RK |
5023 | |
5024 | It is up to you to make sure that the assembler names you choose do not | |
5025 | conflict with any other assembler symbols. Also, you must not use a | |
f0523f02 JM |
5026 | register name; that would produce completely invalid assembler code. GCC |
5027 | does not as yet have the ability to store static variables in registers. | |
c1f7febf RK |
5028 | Perhaps that will be added. |
5029 | ||
5030 | @node Explicit Reg Vars | |
5031 | @section Variables in Specified Registers | |
5032 | @cindex explicit register variables | |
5033 | @cindex variables in specified registers | |
5034 | @cindex specified registers | |
5035 | @cindex registers, global allocation | |
5036 | ||
5037 | GNU C allows you to put a few global variables into specified hardware | |
5038 | registers. You can also specify the register in which an ordinary | |
5039 | register variable should be allocated. | |
5040 | ||
5041 | @itemize @bullet | |
5042 | @item | |
5043 | Global register variables reserve registers throughout the program. | |
5044 | This may be useful in programs such as programming language | |
5045 | interpreters which have a couple of global variables that are accessed | |
5046 | very often. | |
5047 | ||
5048 | @item | |
5049 | Local register variables in specific registers do not reserve the | |
805c33df HPN |
5050 | registers, except at the point where they are used as input or output |
5051 | operands in an @code{asm} statement and the @code{asm} statement itself is | |
5052 | not deleted. The compiler's data flow analysis is capable of determining | |
c1f7febf | 5053 | where the specified registers contain live values, and where they are |
8d344fbc | 5054 | available for other uses. Stores into local register variables may be deleted |
0deaf590 JL |
5055 | when they appear to be dead according to dataflow analysis. References |
5056 | to local register variables may be deleted or moved or simplified. | |
c1f7febf RK |
5057 | |
5058 | These local variables are sometimes convenient for use with the extended | |
5059 | @code{asm} feature (@pxref{Extended Asm}), if you want to write one | |
5060 | output of the assembler instruction directly into a particular register. | |
5061 | (This will work provided the register you specify fits the constraints | |
5062 | specified for that operand in the @code{asm}.) | |
5063 | @end itemize | |
5064 | ||
5065 | @menu | |
5066 | * Global Reg Vars:: | |
5067 | * Local Reg Vars:: | |
5068 | @end menu | |
5069 | ||
5070 | @node Global Reg Vars | |
5071 | @subsection Defining Global Register Variables | |
5072 | @cindex global register variables | |
5073 | @cindex registers, global variables in | |
5074 | ||
5075 | You can define a global register variable in GNU C like this: | |
5076 | ||
3ab51846 | 5077 | @smallexample |
c1f7febf | 5078 | register int *foo asm ("a5"); |
3ab51846 | 5079 | @end smallexample |
c1f7febf RK |
5080 | |
5081 | @noindent | |
5082 | Here @code{a5} is the name of the register which should be used. Choose a | |
5083 | register which is normally saved and restored by function calls on your | |
5084 | machine, so that library routines will not clobber it. | |
5085 | ||
5086 | Naturally the register name is cpu-dependent, so you would need to | |
5087 | conditionalize your program according to cpu type. The register | |
5088 | @code{a5} would be a good choice on a 68000 for a variable of pointer | |
5089 | type. On machines with register windows, be sure to choose a ``global'' | |
5090 | register that is not affected magically by the function call mechanism. | |
5091 | ||
5092 | In addition, operating systems on one type of cpu may differ in how they | |
5093 | name the registers; then you would need additional conditionals. For | |
5094 | example, some 68000 operating systems call this register @code{%a5}. | |
5095 | ||
5096 | Eventually there may be a way of asking the compiler to choose a register | |
5097 | automatically, but first we need to figure out how it should choose and | |
5098 | how to enable you to guide the choice. No solution is evident. | |
5099 | ||
5100 | Defining a global register variable in a certain register reserves that | |
5101 | register entirely for this use, at least within the current compilation. | |
5102 | The register will not be allocated for any other purpose in the functions | |
5103 | in the current compilation. The register will not be saved and restored by | |
5104 | these functions. Stores into this register are never deleted even if they | |
5105 | would appear to be dead, but references may be deleted or moved or | |
5106 | simplified. | |
5107 | ||
5108 | It is not safe to access the global register variables from signal | |
5109 | handlers, or from more than one thread of control, because the system | |
5110 | library routines may temporarily use the register for other things (unless | |
5111 | you recompile them specially for the task at hand). | |
5112 | ||
5113 | @cindex @code{qsort}, and global register variables | |
5114 | It is not safe for one function that uses a global register variable to | |
5115 | call another such function @code{foo} by way of a third function | |
e979f9e8 | 5116 | @code{lose} that was compiled without knowledge of this variable (i.e.@: in a |
c1f7febf RK |
5117 | different source file in which the variable wasn't declared). This is |
5118 | because @code{lose} might save the register and put some other value there. | |
5119 | For example, you can't expect a global register variable to be available in | |
5120 | the comparison-function that you pass to @code{qsort}, since @code{qsort} | |
5121 | might have put something else in that register. (If you are prepared to | |
5122 | recompile @code{qsort} with the same global register variable, you can | |
5123 | solve this problem.) | |
5124 | ||
5125 | If you want to recompile @code{qsort} or other source files which do not | |
5126 | actually use your global register variable, so that they will not use that | |
5127 | register for any other purpose, then it suffices to specify the compiler | |
84330467 | 5128 | option @option{-ffixed-@var{reg}}. You need not actually add a global |
c1f7febf RK |
5129 | register declaration to their source code. |
5130 | ||
5131 | A function which can alter the value of a global register variable cannot | |
5132 | safely be called from a function compiled without this variable, because it | |
5133 | could clobber the value the caller expects to find there on return. | |
5134 | Therefore, the function which is the entry point into the part of the | |
5135 | program that uses the global register variable must explicitly save and | |
5136 | restore the value which belongs to its caller. | |
5137 | ||
5138 | @cindex register variable after @code{longjmp} | |
5139 | @cindex global register after @code{longjmp} | |
5140 | @cindex value after @code{longjmp} | |
5141 | @findex longjmp | |
5142 | @findex setjmp | |
5143 | On most machines, @code{longjmp} will restore to each global register | |
5144 | variable the value it had at the time of the @code{setjmp}. On some | |
5145 | machines, however, @code{longjmp} will not change the value of global | |
5146 | register variables. To be portable, the function that called @code{setjmp} | |
5147 | should make other arrangements to save the values of the global register | |
5148 | variables, and to restore them in a @code{longjmp}. This way, the same | |
5149 | thing will happen regardless of what @code{longjmp} does. | |
5150 | ||
5151 | All global register variable declarations must precede all function | |
5152 | definitions. If such a declaration could appear after function | |
5153 | definitions, the declaration would be too late to prevent the register from | |
5154 | being used for other purposes in the preceding functions. | |
5155 | ||
5156 | Global register variables may not have initial values, because an | |
5157 | executable file has no means to supply initial contents for a register. | |
5158 | ||
981f6289 | 5159 | On the SPARC, there are reports that g3 @dots{} g7 are suitable |
c1f7febf RK |
5160 | registers, but certain library functions, such as @code{getwd}, as well |
5161 | as the subroutines for division and remainder, modify g3 and g4. g1 and | |
5162 | g2 are local temporaries. | |
5163 | ||
5164 | On the 68000, a2 @dots{} a5 should be suitable, as should d2 @dots{} d7. | |
5165 | Of course, it will not do to use more than a few of those. | |
5166 | ||
5167 | @node Local Reg Vars | |
5168 | @subsection Specifying Registers for Local Variables | |
5169 | @cindex local variables, specifying registers | |
5170 | @cindex specifying registers for local variables | |
5171 | @cindex registers for local variables | |
5172 | ||
5173 | You can define a local register variable with a specified register | |
5174 | like this: | |
5175 | ||
3ab51846 | 5176 | @smallexample |
c1f7febf | 5177 | register int *foo asm ("a5"); |
3ab51846 | 5178 | @end smallexample |
c1f7febf RK |
5179 | |
5180 | @noindent | |
5181 | Here @code{a5} is the name of the register which should be used. Note | |
5182 | that this is the same syntax used for defining global register | |
5183 | variables, but for a local variable it would appear within a function. | |
5184 | ||
5185 | Naturally the register name is cpu-dependent, but this is not a | |
5186 | problem, since specific registers are most often useful with explicit | |
5187 | assembler instructions (@pxref{Extended Asm}). Both of these things | |
5188 | generally require that you conditionalize your program according to | |
5189 | cpu type. | |
5190 | ||
5191 | In addition, operating systems on one type of cpu may differ in how they | |
5192 | name the registers; then you would need additional conditionals. For | |
5193 | example, some 68000 operating systems call this register @code{%a5}. | |
5194 | ||
c1f7febf RK |
5195 | Defining such a register variable does not reserve the register; it |
5196 | remains available for other uses in places where flow control determines | |
d754127f | 5197 | the variable's value is not live. |
e5e809f4 | 5198 | |
f0523f02 | 5199 | This option does not guarantee that GCC will generate code that has |
e5e809f4 | 5200 | this variable in the register you specify at all times. You may not |
805c33df HPN |
5201 | code an explicit reference to this register in the @emph{assembler |
5202 | instruction template} part of an @code{asm} statement and assume it will | |
5203 | always refer to this variable. However, using the variable as an | |
5204 | @code{asm} @emph{operand} guarantees that the specified register is used | |
5205 | for the operand. | |
c1f7febf | 5206 | |
8d344fbc | 5207 | Stores into local register variables may be deleted when they appear to be dead |
0deaf590 JL |
5208 | according to dataflow analysis. References to local register variables may |
5209 | be deleted or moved or simplified. | |
5210 | ||
b55d5746 HPN |
5211 | As for global register variables, it's recommended that you choose a |
5212 | register which is normally saved and restored by function calls on | |
5213 | your machine, so that library routines will not clobber it. A common | |
5214 | pitfall is to initialize multiple call-clobbered registers with | |
5215 | arbitrary expressions, where a function call or library call for an | |
5216 | arithmetic operator will overwrite a register value from a previous | |
5217 | assignment, for example @code{r0} below: | |
5218 | @smallexample | |
5219 | register int *p1 asm ("r0") = @dots{}; | |
5220 | register int *p2 asm ("r1") = @dots{}; | |
5221 | @end smallexample | |
5222 | In those cases, a solution is to use a temporary variable for | |
5223 | each arbitrary expression. @xref{Example of asm with clobbered asm reg}. | |
5224 | ||
c1f7febf RK |
5225 | @node Alternate Keywords |
5226 | @section Alternate Keywords | |
5227 | @cindex alternate keywords | |
5228 | @cindex keywords, alternate | |
5229 | ||
5490d604 | 5230 | @option{-ansi} and the various @option{-std} options disable certain |
f458d1d5 ZW |
5231 | keywords. This causes trouble when you want to use GNU C extensions, or |
5232 | a general-purpose header file that should be usable by all programs, | |
5233 | including ISO C programs. The keywords @code{asm}, @code{typeof} and | |
5234 | @code{inline} are not available in programs compiled with | |
5235 | @option{-ansi} or @option{-std} (although @code{inline} can be used in a | |
5236 | program compiled with @option{-std=c99}). The ISO C99 keyword | |
5490d604 JM |
5237 | @code{restrict} is only available when @option{-std=gnu99} (which will |
5238 | eventually be the default) or @option{-std=c99} (or the equivalent | |
bd819a4a | 5239 | @option{-std=iso9899:1999}) is used. |
c1f7febf RK |
5240 | |
5241 | The way to solve these problems is to put @samp{__} at the beginning and | |
5242 | end of each problematical keyword. For example, use @code{__asm__} | |
f458d1d5 | 5243 | instead of @code{asm}, and @code{__inline__} instead of @code{inline}. |
c1f7febf RK |
5244 | |
5245 | Other C compilers won't accept these alternative keywords; if you want to | |
5246 | compile with another compiler, you can define the alternate keywords as | |
5247 | macros to replace them with the customary keywords. It looks like this: | |
5248 | ||
3ab51846 | 5249 | @smallexample |
c1f7febf RK |
5250 | #ifndef __GNUC__ |
5251 | #define __asm__ asm | |
5252 | #endif | |
3ab51846 | 5253 | @end smallexample |
c1f7febf | 5254 | |
6e6b0525 | 5255 | @findex __extension__ |
84330467 JM |
5256 | @opindex pedantic |
5257 | @option{-pedantic} and other options cause warnings for many GNU C extensions. | |
dbe519e0 | 5258 | You can |
c1f7febf RK |
5259 | prevent such warnings within one expression by writing |
5260 | @code{__extension__} before the expression. @code{__extension__} has no | |
5261 | effect aside from this. | |
5262 | ||
5263 | @node Incomplete Enums | |
5264 | @section Incomplete @code{enum} Types | |
5265 | ||
5266 | You can define an @code{enum} tag without specifying its possible values. | |
5267 | This results in an incomplete type, much like what you get if you write | |
5268 | @code{struct foo} without describing the elements. A later declaration | |
5269 | which does specify the possible values completes the type. | |
5270 | ||
5271 | You can't allocate variables or storage using the type while it is | |
5272 | incomplete. However, you can work with pointers to that type. | |
5273 | ||
5274 | This extension may not be very useful, but it makes the handling of | |
5275 | @code{enum} more consistent with the way @code{struct} and @code{union} | |
5276 | are handled. | |
5277 | ||
5278 | This extension is not supported by GNU C++. | |
5279 | ||
5280 | @node Function Names | |
5281 | @section Function Names as Strings | |
e6cc3a24 | 5282 | @cindex @code{__func__} identifier |
4b404517 JM |
5283 | @cindex @code{__FUNCTION__} identifier |
5284 | @cindex @code{__PRETTY_FUNCTION__} identifier | |
c1f7febf | 5285 | |
e6cc3a24 ZW |
5286 | GCC provides three magic variables which hold the name of the current |
5287 | function, as a string. The first of these is @code{__func__}, which | |
5288 | is part of the C99 standard: | |
5289 | ||
e6cc3a24 ZW |
5290 | The identifier @code{__func__} is implicitly declared by the translator |
5291 | as if, immediately following the opening brace of each function | |
5292 | definition, the declaration | |
5293 | ||
5294 | @smallexample | |
5295 | static const char __func__[] = "function-name"; | |
5296 | @end smallexample | |
c1f7febf | 5297 | |
38bb2b65 | 5298 | @noindent |
e6cc3a24 ZW |
5299 | appeared, where function-name is the name of the lexically-enclosing |
5300 | function. This name is the unadorned name of the function. | |
e6cc3a24 ZW |
5301 | |
5302 | @code{__FUNCTION__} is another name for @code{__func__}. Older | |
5303 | versions of GCC recognize only this name. However, it is not | |
5304 | standardized. For maximum portability, we recommend you use | |
5305 | @code{__func__}, but provide a fallback definition with the | |
5306 | preprocessor: | |
5307 | ||
5308 | @smallexample | |
5309 | #if __STDC_VERSION__ < 199901L | |
5310 | # if __GNUC__ >= 2 | |
5311 | # define __func__ __FUNCTION__ | |
5312 | # else | |
5313 | # define __func__ "<unknown>" | |
5314 | # endif | |
5315 | #endif | |
5316 | @end smallexample | |
5317 | ||
5318 | In C, @code{__PRETTY_FUNCTION__} is yet another name for | |
5319 | @code{__func__}. However, in C++, @code{__PRETTY_FUNCTION__} contains | |
5320 | the type signature of the function as well as its bare name. For | |
5321 | example, this program: | |
c1f7febf RK |
5322 | |
5323 | @smallexample | |
5324 | extern "C" @{ | |
5325 | extern int printf (char *, ...); | |
5326 | @} | |
5327 | ||
5328 | class a @{ | |
5329 | public: | |
a721a601 | 5330 | void sub (int i) |
c1f7febf RK |
5331 | @{ |
5332 | printf ("__FUNCTION__ = %s\n", __FUNCTION__); | |
5333 | printf ("__PRETTY_FUNCTION__ = %s\n", __PRETTY_FUNCTION__); | |
5334 | @} | |
5335 | @}; | |
5336 | ||
5337 | int | |
5338 | main (void) | |
5339 | @{ | |
5340 | a ax; | |
5341 | ax.sub (0); | |
5342 | return 0; | |
5343 | @} | |
5344 | @end smallexample | |
5345 | ||
5346 | @noindent | |
5347 | gives this output: | |
5348 | ||
5349 | @smallexample | |
5350 | __FUNCTION__ = sub | |
e6cc3a24 | 5351 | __PRETTY_FUNCTION__ = void a::sub(int) |
22acfb79 NM |
5352 | @end smallexample |
5353 | ||
e6cc3a24 ZW |
5354 | These identifiers are not preprocessor macros. In GCC 3.3 and |
5355 | earlier, in C only, @code{__FUNCTION__} and @code{__PRETTY_FUNCTION__} | |
5356 | were treated as string literals; they could be used to initialize | |
5357 | @code{char} arrays, and they could be concatenated with other string | |
5358 | literals. GCC 3.4 and later treat them as variables, like | |
5359 | @code{__func__}. In C++, @code{__FUNCTION__} and | |
5360 | @code{__PRETTY_FUNCTION__} have always been variables. | |
22acfb79 | 5361 | |
c1f7febf RK |
5362 | @node Return Address |
5363 | @section Getting the Return or Frame Address of a Function | |
5364 | ||
5365 | These functions may be used to get information about the callers of a | |
5366 | function. | |
5367 | ||
84330467 | 5368 | @deftypefn {Built-in Function} {void *} __builtin_return_address (unsigned int @var{level}) |
c1f7febf RK |
5369 | This function returns the return address of the current function, or of |
5370 | one of its callers. The @var{level} argument is number of frames to | |
5371 | scan up the call stack. A value of @code{0} yields the return address | |
5372 | of the current function, a value of @code{1} yields the return address | |
8a36672b | 5373 | of the caller of the current function, and so forth. When inlining |
95b1627e EC |
5374 | the expected behavior is that the function will return the address of |
5375 | the function that will be returned to. To work around this behavior use | |
5376 | the @code{noinline} function attribute. | |
c1f7febf RK |
5377 | |
5378 | The @var{level} argument must be a constant integer. | |
5379 | ||
5380 | On some machines it may be impossible to determine the return address of | |
5381 | any function other than the current one; in such cases, or when the top | |
dd96fbc5 | 5382 | of the stack has been reached, this function will return @code{0} or a |
8a36672b | 5383 | random value. In addition, @code{__builtin_frame_address} may be used |
dd96fbc5 | 5384 | to determine if the top of the stack has been reached. |
c1f7febf | 5385 | |
df2a54e9 | 5386 | This function should only be used with a nonzero argument for debugging |
c1f7febf | 5387 | purposes. |
84330467 | 5388 | @end deftypefn |
c1f7febf | 5389 | |
84330467 | 5390 | @deftypefn {Built-in Function} {void *} __builtin_frame_address (unsigned int @var{level}) |
c1f7febf RK |
5391 | This function is similar to @code{__builtin_return_address}, but it |
5392 | returns the address of the function frame rather than the return address | |
5393 | of the function. Calling @code{__builtin_frame_address} with a value of | |
5394 | @code{0} yields the frame address of the current function, a value of | |
5395 | @code{1} yields the frame address of the caller of the current function, | |
5396 | and so forth. | |
5397 | ||
5398 | The frame is the area on the stack which holds local variables and saved | |
5399 | registers. The frame address is normally the address of the first word | |
5400 | pushed on to the stack by the function. However, the exact definition | |
5401 | depends upon the processor and the calling convention. If the processor | |
5402 | has a dedicated frame pointer register, and the function has a frame, | |
5403 | then @code{__builtin_frame_address} will return the value of the frame | |
5404 | pointer register. | |
5405 | ||
dd96fbc5 L |
5406 | On some machines it may be impossible to determine the frame address of |
5407 | any function other than the current one; in such cases, or when the top | |
5408 | of the stack has been reached, this function will return @code{0} if | |
5409 | the first frame pointer is properly initialized by the startup code. | |
5410 | ||
df2a54e9 | 5411 | This function should only be used with a nonzero argument for debugging |
dd96fbc5 | 5412 | purposes. |
84330467 | 5413 | @end deftypefn |
c1f7febf | 5414 | |
1255c85c BS |
5415 | @node Vector Extensions |
5416 | @section Using vector instructions through built-in functions | |
5417 | ||
5418 | On some targets, the instruction set contains SIMD vector instructions that | |
5419 | operate on multiple values contained in one large register at the same time. | |
5420 | For example, on the i386 the MMX, 3Dnow! and SSE extensions can be used | |
5421 | this way. | |
5422 | ||
5423 | The first step in using these extensions is to provide the necessary data | |
5424 | types. This should be done using an appropriate @code{typedef}: | |
5425 | ||
3ab51846 | 5426 | @smallexample |
4a5eab38 | 5427 | typedef int v4si __attribute__ ((vector_size (16))); |
3ab51846 | 5428 | @end smallexample |
1255c85c | 5429 | |
4a5eab38 PB |
5430 | The @code{int} type specifies the base type, while the attribute specifies |
5431 | the vector size for the variable, measured in bytes. For example, the | |
5432 | declaration above causes the compiler to set the mode for the @code{v4si} | |
5433 | type to be 16 bytes wide and divided into @code{int} sized units. For | |
5434 | a 32-bit @code{int} this means a vector of 4 units of 4 bytes, and the | |
5435 | corresponding mode of @code{foo} will be @acronym{V4SI}. | |
1255c85c | 5436 | |
4a5eab38 PB |
5437 | The @code{vector_size} attribute is only applicable to integral and |
5438 | float scalars, although arrays, pointers, and function return values | |
5439 | are allowed in conjunction with this construct. | |
5440 | ||
5441 | All the basic integer types can be used as base types, both as signed | |
5442 | and as unsigned: @code{char}, @code{short}, @code{int}, @code{long}, | |
5443 | @code{long long}. In addition, @code{float} and @code{double} can be | |
5444 | used to build floating-point vector types. | |
1255c85c | 5445 | |
cb2a532e | 5446 | Specifying a combination that is not valid for the current architecture |
2dd76960 | 5447 | will cause GCC to synthesize the instructions using a narrower mode. |
cb2a532e | 5448 | For example, if you specify a variable of type @code{V4SI} and your |
2dd76960 | 5449 | architecture does not allow for this specific SIMD type, GCC will |
cb2a532e AH |
5450 | produce code that uses 4 @code{SIs}. |
5451 | ||
5452 | The types defined in this manner can be used with a subset of normal C | |
2dd76960 | 5453 | operations. Currently, GCC will allow using the following operators |
3a3e1600 | 5454 | on these types: @code{+, -, *, /, unary minus, ^, |, &, ~}@. |
cb2a532e AH |
5455 | |
5456 | The operations behave like C++ @code{valarrays}. Addition is defined as | |
5457 | the addition of the corresponding elements of the operands. For | |
5458 | example, in the code below, each of the 4 elements in @var{a} will be | |
5459 | added to the corresponding 4 elements in @var{b} and the resulting | |
5460 | vector will be stored in @var{c}. | |
5461 | ||
3ab51846 | 5462 | @smallexample |
4a5eab38 | 5463 | typedef int v4si __attribute__ ((vector_size (16))); |
cb2a532e AH |
5464 | |
5465 | v4si a, b, c; | |
5466 | ||
5467 | c = a + b; | |
3ab51846 | 5468 | @end smallexample |
cb2a532e | 5469 | |
3a3e1600 GK |
5470 | Subtraction, multiplication, division, and the logical operations |
5471 | operate in a similar manner. Likewise, the result of using the unary | |
5472 | minus or complement operators on a vector type is a vector whose | |
5473 | elements are the negative or complemented values of the corresponding | |
cb2a532e AH |
5474 | elements in the operand. |
5475 | ||
5476 | You can declare variables and use them in function calls and returns, as | |
5477 | well as in assignments and some casts. You can specify a vector type as | |
5478 | a return type for a function. Vector types can also be used as function | |
5479 | arguments. It is possible to cast from one vector type to another, | |
5480 | provided they are of the same size (in fact, you can also cast vectors | |
5481 | to and from other datatypes of the same size). | |
5482 | ||
5483 | You cannot operate between vectors of different lengths or different | |
90a21764 | 5484 | signedness without a cast. |
cb2a532e AH |
5485 | |
5486 | A port that supports hardware vector operations, usually provides a set | |
5487 | of built-in functions that can be used to operate on vectors. For | |
5488 | example, a function to add two vectors and multiply the result by a | |
5489 | third could look like this: | |
1255c85c | 5490 | |
3ab51846 | 5491 | @smallexample |
1255c85c BS |
5492 | v4si f (v4si a, v4si b, v4si c) |
5493 | @{ | |
5494 | v4si tmp = __builtin_addv4si (a, b); | |
5495 | return __builtin_mulv4si (tmp, c); | |
5496 | @} | |
5497 | ||
3ab51846 | 5498 | @end smallexample |
1255c85c | 5499 | |
7a3ea201 RH |
5500 | @node Offsetof |
5501 | @section Offsetof | |
5502 | @findex __builtin_offsetof | |
5503 | ||
5504 | GCC implements for both C and C++ a syntactic extension to implement | |
5505 | the @code{offsetof} macro. | |
5506 | ||
5507 | @smallexample | |
5508 | primary: | |
6ccde948 | 5509 | "__builtin_offsetof" "(" @code{typename} "," offsetof_member_designator ")" |
7a3ea201 RH |
5510 | |
5511 | offsetof_member_designator: | |
6ccde948 RW |
5512 | @code{identifier} |
5513 | | offsetof_member_designator "." @code{identifier} | |
5514 | | offsetof_member_designator "[" @code{expr} "]" | |
7a3ea201 RH |
5515 | @end smallexample |
5516 | ||
5517 | This extension is sufficient such that | |
5518 | ||
5519 | @smallexample | |
5520 | #define offsetof(@var{type}, @var{member}) __builtin_offsetof (@var{type}, @var{member}) | |
5521 | @end smallexample | |
5522 | ||
5523 | is a suitable definition of the @code{offsetof} macro. In C++, @var{type} | |
5524 | may be dependent. In either case, @var{member} may consist of a single | |
5525 | identifier, or a sequence of member accesses and array references. | |
5526 | ||
48ae6c13 RH |
5527 | @node Atomic Builtins |
5528 | @section Built-in functions for atomic memory access | |
5529 | ||
5530 | The following builtins are intended to be compatible with those described | |
5531 | in the @cite{Intel Itanium Processor-specific Application Binary Interface}, | |
5532 | section 7.4. As such, they depart from the normal GCC practice of using | |
5533 | the ``__builtin_'' prefix, and further that they are overloaded such that | |
5534 | they work on multiple types. | |
5535 | ||
5536 | The definition given in the Intel documentation allows only for the use of | |
5537 | the types @code{int}, @code{long}, @code{long long} as well as their unsigned | |
5538 | counterparts. GCC will allow any integral scalar or pointer type that is | |
5539 | 1, 2, 4 or 8 bytes in length. | |
5540 | ||
5541 | Not all operations are supported by all target processors. If a particular | |
5542 | operation cannot be implemented on the target processor, a warning will be | |
5543 | generated and a call an external function will be generated. The external | |
5544 | function will carry the same name as the builtin, with an additional suffix | |
5545 | @samp{_@var{n}} where @var{n} is the size of the data type. | |
5546 | ||
5547 | @c ??? Should we have a mechanism to suppress this warning? This is almost | |
5548 | @c useful for implementing the operation under the control of an external | |
5549 | @c mutex. | |
5550 | ||
5551 | In most cases, these builtins are considered a @dfn{full barrier}. That is, | |
5552 | no memory operand will be moved across the operation, either forward or | |
5553 | backward. Further, instructions will be issued as necessary to prevent the | |
5554 | processor from speculating loads across the operation and from queuing stores | |
5555 | after the operation. | |
5556 | ||
5557 | All of the routines are are described in the Intel documentation to take | |
5558 | ``an optional list of variables protected by the memory barrier''. It's | |
5559 | not clear what is meant by that; it could mean that @emph{only} the | |
5560 | following variables are protected, or it could mean that these variables | |
5561 | should in addition be protected. At present GCC ignores this list and | |
5562 | protects all variables which are globally accessible. If in the future | |
5563 | we make some use of this list, an empty list will continue to mean all | |
5564 | globally accessible variables. | |
5565 | ||
5566 | @table @code | |
5567 | @item @var{type} __sync_fetch_and_add (@var{type} *ptr, @var{type} value, ...) | |
5568 | @itemx @var{type} __sync_fetch_and_sub (@var{type} *ptr, @var{type} value, ...) | |
5569 | @itemx @var{type} __sync_fetch_and_or (@var{type} *ptr, @var{type} value, ...) | |
5570 | @itemx @var{type} __sync_fetch_and_and (@var{type} *ptr, @var{type} value, ...) | |
5571 | @itemx @var{type} __sync_fetch_and_xor (@var{type} *ptr, @var{type} value, ...) | |
5572 | @itemx @var{type} __sync_fetch_and_nand (@var{type} *ptr, @var{type} value, ...) | |
5573 | @findex __sync_fetch_and_add | |
5574 | @findex __sync_fetch_and_sub | |
5575 | @findex __sync_fetch_and_or | |
5576 | @findex __sync_fetch_and_and | |
5577 | @findex __sync_fetch_and_xor | |
5578 | @findex __sync_fetch_and_nand | |
5579 | These builtins perform the operation suggested by the name, and | |
5580 | returns the value that had previously been in memory. That is, | |
5581 | ||
5582 | @smallexample | |
5583 | @{ tmp = *ptr; *ptr @var{op}= value; return tmp; @} | |
f12b785d | 5584 | @{ tmp = *ptr; *ptr = ~tmp & value; return tmp; @} // nand |
48ae6c13 RH |
5585 | @end smallexample |
5586 | ||
48ae6c13 RH |
5587 | @item @var{type} __sync_add_and_fetch (@var{type} *ptr, @var{type} value, ...) |
5588 | @itemx @var{type} __sync_sub_and_fetch (@var{type} *ptr, @var{type} value, ...) | |
5589 | @itemx @var{type} __sync_or_and_fetch (@var{type} *ptr, @var{type} value, ...) | |
5590 | @itemx @var{type} __sync_and_and_fetch (@var{type} *ptr, @var{type} value, ...) | |
5591 | @itemx @var{type} __sync_xor_and_fetch (@var{type} *ptr, @var{type} value, ...) | |
5592 | @itemx @var{type} __sync_nand_and_fetch (@var{type} *ptr, @var{type} value, ...) | |
5593 | @findex __sync_add_and_fetch | |
5594 | @findex __sync_sub_and_fetch | |
5595 | @findex __sync_or_and_fetch | |
5596 | @findex __sync_and_and_fetch | |
5597 | @findex __sync_xor_and_fetch | |
5598 | @findex __sync_nand_and_fetch | |
5599 | These builtins perform the operation suggested by the name, and | |
5600 | return the new value. That is, | |
5601 | ||
5602 | @smallexample | |
5603 | @{ *ptr @var{op}= value; return *ptr; @} | |
f12b785d | 5604 | @{ *ptr = ~*ptr & value; return *ptr; @} // nand |
48ae6c13 RH |
5605 | @end smallexample |
5606 | ||
5607 | @item bool __sync_bool_compare_and_swap (@var{type} *ptr, @var{type} oldval @var{type} newval, ...) | |
5608 | @itemx @var{type} __sync_val_compare_and_swap (@var{type} *ptr, @var{type} oldval @var{type} newval, ...) | |
5609 | @findex __sync_bool_compare_and_swap | |
5610 | @findex __sync_val_compare_and_swap | |
5611 | These builtins perform an atomic compare and swap. That is, if the current | |
5612 | value of @code{*@var{ptr}} is @var{oldval}, then write @var{newval} into | |
5613 | @code{*@var{ptr}}. | |
5614 | ||
0ac11108 | 5615 | The ``bool'' version returns true if the comparison is successful and |
48ae6c13 | 5616 | @var{newval} was written. The ``val'' version returns the contents |
f12b785d | 5617 | of @code{*@var{ptr}} before the operation. |
48ae6c13 RH |
5618 | |
5619 | @item __sync_synchronize (...) | |
5620 | @findex __sync_synchronize | |
5621 | This builtin issues a full memory barrier. | |
5622 | ||
5623 | @item @var{type} __sync_lock_test_and_set (@var{type} *ptr, @var{type} value, ...) | |
5624 | @findex __sync_lock_test_and_set | |
5625 | This builtin, as described by Intel, is not a traditional test-and-set | |
5626 | operation, but rather an atomic exchange operation. It writes @var{value} | |
5627 | into @code{*@var{ptr}}, and returns the previous contents of | |
5628 | @code{*@var{ptr}}. | |
5629 | ||
5630 | Many targets have only minimal support for such locks, and do not support | |
5631 | a full exchange operation. In this case, a target may support reduced | |
5632 | functionality here by which the @emph{only} valid value to store is the | |
5633 | immediate constant 1. The exact value actually stored in @code{*@var{ptr}} | |
5634 | is implementation defined. | |
5635 | ||
5636 | This builtin is not a full barrier, but rather an @dfn{acquire barrier}. | |
5637 | This means that references after the builtin cannot move to (or be | |
5638 | speculated to) before the builtin, but previous memory stores may not | |
0ac11108 | 5639 | be globally visible yet, and previous memory loads may not yet be |
48ae6c13 RH |
5640 | satisfied. |
5641 | ||
5642 | @item void __sync_lock_release (@var{type} *ptr, ...) | |
5643 | @findex __sync_lock_release | |
5644 | This builtin releases the lock acquired by @code{__sync_lock_test_and_set}. | |
5645 | Normally this means writing the constant 0 to @code{*@var{ptr}}. | |
5646 | ||
5647 | This builtin is not a full barrier, but rather a @dfn{release barrier}. | |
5648 | This means that all previous memory stores are globally visible, and all | |
5649 | previous memory loads have been satisfied, but following memory reads | |
5650 | are not prevented from being speculated to before the barrier. | |
5651 | @end table | |
5652 | ||
10a0d495 JJ |
5653 | @node Object Size Checking |
5654 | @section Object Size Checking Builtins | |
5655 | @findex __builtin_object_size | |
5656 | @findex __builtin___memcpy_chk | |
5657 | @findex __builtin___mempcpy_chk | |
5658 | @findex __builtin___memmove_chk | |
5659 | @findex __builtin___memset_chk | |
5660 | @findex __builtin___strcpy_chk | |
5661 | @findex __builtin___stpcpy_chk | |
5662 | @findex __builtin___strncpy_chk | |
5663 | @findex __builtin___strcat_chk | |
5664 | @findex __builtin___strncat_chk | |
5665 | @findex __builtin___sprintf_chk | |
5666 | @findex __builtin___snprintf_chk | |
5667 | @findex __builtin___vsprintf_chk | |
5668 | @findex __builtin___vsnprintf_chk | |
5669 | @findex __builtin___printf_chk | |
5670 | @findex __builtin___vprintf_chk | |
5671 | @findex __builtin___fprintf_chk | |
5672 | @findex __builtin___vfprintf_chk | |
5673 | ||
5674 | GCC implements a limited buffer overflow protection mechanism | |
5675 | that can prevent some buffer overflow attacks. | |
5676 | ||
5677 | @deftypefn {Built-in Function} {size_t} __builtin_object_size (void * @var{ptr}, int @var{type}) | |
5678 | is a built-in construct that returns a constant number of bytes from | |
5679 | @var{ptr} to the end of the object @var{ptr} pointer points to | |
5680 | (if known at compile time). @code{__builtin_object_size} never evaluates | |
5681 | its arguments for side-effects. If there are any side-effects in them, it | |
5682 | returns @code{(size_t) -1} for @var{type} 0 or 1 and @code{(size_t) 0} | |
5683 | for @var{type} 2 or 3. If there are multiple objects @var{ptr} can | |
5684 | point to and all of them are known at compile time, the returned number | |
5685 | is the maximum of remaining byte counts in those objects if @var{type} & 2 is | |
a4d05547 | 5686 | 0 and minimum if nonzero. If it is not possible to determine which objects |
10a0d495 JJ |
5687 | @var{ptr} points to at compile time, @code{__builtin_object_size} should |
5688 | return @code{(size_t) -1} for @var{type} 0 or 1 and @code{(size_t) 0} | |
5689 | for @var{type} 2 or 3. | |
5690 | ||
5691 | @var{type} is an integer constant from 0 to 3. If the least significant | |
5692 | bit is clear, objects are whole variables, if it is set, a closest | |
5693 | surrounding subobject is considered the object a pointer points to. | |
5694 | The second bit determines if maximum or minimum of remaining bytes | |
5695 | is computed. | |
5696 | ||
5697 | @smallexample | |
5698 | struct V @{ char buf1[10]; int b; char buf2[10]; @} var; | |
5699 | char *p = &var.buf1[1], *q = &var.b; | |
5700 | ||
5701 | /* Here the object p points to is var. */ | |
5702 | assert (__builtin_object_size (p, 0) == sizeof (var) - 1); | |
5703 | /* The subobject p points to is var.buf1. */ | |
5704 | assert (__builtin_object_size (p, 1) == sizeof (var.buf1) - 1); | |
5705 | /* The object q points to is var. */ | |
5706 | assert (__builtin_object_size (q, 0) | |
6ccde948 | 5707 | == (char *) (&var + 1) - (char *) &var.b); |
10a0d495 JJ |
5708 | /* The subobject q points to is var.b. */ |
5709 | assert (__builtin_object_size (q, 1) == sizeof (var.b)); | |
5710 | @end smallexample | |
5711 | @end deftypefn | |
5712 | ||
5713 | There are built-in functions added for many common string operation | |
021efafc | 5714 | functions, e.g., for @code{memcpy} @code{__builtin___memcpy_chk} |
10a0d495 JJ |
5715 | built-in is provided. This built-in has an additional last argument, |
5716 | which is the number of bytes remaining in object the @var{dest} | |
5717 | argument points to or @code{(size_t) -1} if the size is not known. | |
5718 | ||
5719 | The built-in functions are optimized into the normal string functions | |
5720 | like @code{memcpy} if the last argument is @code{(size_t) -1} or if | |
5721 | it is known at compile time that the destination object will not | |
5722 | be overflown. If the compiler can determine at compile time the | |
5723 | object will be always overflown, it issues a warning. | |
5724 | ||
5725 | The intended use can be e.g. | |
5726 | ||
5727 | @smallexample | |
5728 | #undef memcpy | |
5729 | #define bos0(dest) __builtin_object_size (dest, 0) | |
5730 | #define memcpy(dest, src, n) \ | |
5731 | __builtin___memcpy_chk (dest, src, n, bos0 (dest)) | |
5732 | ||
5733 | char *volatile p; | |
5734 | char buf[10]; | |
5735 | /* It is unknown what object p points to, so this is optimized | |
5736 | into plain memcpy - no checking is possible. */ | |
5737 | memcpy (p, "abcde", n); | |
5738 | /* Destination is known and length too. It is known at compile | |
5739 | time there will be no overflow. */ | |
5740 | memcpy (&buf[5], "abcde", 5); | |
5741 | /* Destination is known, but the length is not known at compile time. | |
5742 | This will result in __memcpy_chk call that can check for overflow | |
5743 | at runtime. */ | |
5744 | memcpy (&buf[5], "abcde", n); | |
5745 | /* Destination is known and it is known at compile time there will | |
5746 | be overflow. There will be a warning and __memcpy_chk call that | |
5747 | will abort the program at runtime. */ | |
5748 | memcpy (&buf[6], "abcde", 5); | |
5749 | @end smallexample | |
5750 | ||
5751 | Such built-in functions are provided for @code{memcpy}, @code{mempcpy}, | |
5752 | @code{memmove}, @code{memset}, @code{strcpy}, @code{stpcpy}, @code{strncpy}, | |
5753 | @code{strcat} and @code{strncat}. | |
5754 | ||
5755 | There are also checking built-in functions for formatted output functions. | |
5756 | @smallexample | |
5757 | int __builtin___sprintf_chk (char *s, int flag, size_t os, const char *fmt, ...); | |
5758 | int __builtin___snprintf_chk (char *s, size_t maxlen, int flag, size_t os, | |
6ccde948 | 5759 | const char *fmt, ...); |
10a0d495 | 5760 | int __builtin___vsprintf_chk (char *s, int flag, size_t os, const char *fmt, |
6ccde948 | 5761 | va_list ap); |
10a0d495 | 5762 | int __builtin___vsnprintf_chk (char *s, size_t maxlen, int flag, size_t os, |
6ccde948 | 5763 | const char *fmt, va_list ap); |
10a0d495 JJ |
5764 | @end smallexample |
5765 | ||
5766 | The added @var{flag} argument is passed unchanged to @code{__sprintf_chk} | |
021efafc | 5767 | etc.@: functions and can contain implementation specific flags on what |
10a0d495 JJ |
5768 | additional security measures the checking function might take, such as |
5769 | handling @code{%n} differently. | |
5770 | ||
5771 | The @var{os} argument is the object size @var{s} points to, like in the | |
a4d05547 | 5772 | other built-in functions. There is a small difference in the behavior |
10a0d495 JJ |
5773 | though, if @var{os} is @code{(size_t) -1}, the built-in functions are |
5774 | optimized into the non-checking functions only if @var{flag} is 0, otherwise | |
5775 | the checking function is called with @var{os} argument set to | |
5776 | @code{(size_t) -1}. | |
5777 | ||
5778 | In addition to this, there are checking built-in functions | |
5779 | @code{__builtin___printf_chk}, @code{__builtin___vprintf_chk}, | |
5780 | @code{__builtin___fprintf_chk} and @code{__builtin___vfprintf_chk}. | |
5781 | These have just one additional argument, @var{flag}, right before | |
5782 | format string @var{fmt}. If the compiler is able to optimize them to | |
021efafc | 5783 | @code{fputc} etc.@: functions, it will, otherwise the checking function |
10a0d495 JJ |
5784 | should be called and the @var{flag} argument passed to it. |
5785 | ||
185ebd6c | 5786 | @node Other Builtins |
f0523f02 | 5787 | @section Other built-in functions provided by GCC |
c771326b | 5788 | @cindex built-in functions |
3bf5906b | 5789 | @findex __builtin_fpclassify |
0c8d3c2b | 5790 | @findex __builtin_isfinite |
8a91c45b | 5791 | @findex __builtin_isnormal |
01702459 JM |
5792 | @findex __builtin_isgreater |
5793 | @findex __builtin_isgreaterequal | |
05f41289 | 5794 | @findex __builtin_isinf_sign |
01702459 JM |
5795 | @findex __builtin_isless |
5796 | @findex __builtin_islessequal | |
5797 | @findex __builtin_islessgreater | |
5798 | @findex __builtin_isunordered | |
17684d46 RG |
5799 | @findex __builtin_powi |
5800 | @findex __builtin_powif | |
5801 | @findex __builtin_powil | |
98ff7c4d KG |
5802 | @findex _Exit |
5803 | @findex _exit | |
01702459 JM |
5804 | @findex abort |
5805 | @findex abs | |
98ff7c4d KG |
5806 | @findex acos |
5807 | @findex acosf | |
5808 | @findex acosh | |
5809 | @findex acoshf | |
5810 | @findex acoshl | |
5811 | @findex acosl | |
01702459 | 5812 | @findex alloca |
98ff7c4d KG |
5813 | @findex asin |
5814 | @findex asinf | |
5815 | @findex asinh | |
5816 | @findex asinhf | |
5817 | @findex asinhl | |
5818 | @findex asinl | |
29f523be | 5819 | @findex atan |
46847aa6 RS |
5820 | @findex atan2 |
5821 | @findex atan2f | |
5822 | @findex atan2l | |
29f523be | 5823 | @findex atanf |
98ff7c4d KG |
5824 | @findex atanh |
5825 | @findex atanhf | |
5826 | @findex atanhl | |
29f523be | 5827 | @findex atanl |
01702459 JM |
5828 | @findex bcmp |
5829 | @findex bzero | |
075ec276 RS |
5830 | @findex cabs |
5831 | @findex cabsf | |
5832 | @findex cabsl | |
11bf0eb0 KG |
5833 | @findex cacos |
5834 | @findex cacosf | |
5835 | @findex cacosh | |
5836 | @findex cacoshf | |
5837 | @findex cacoshl | |
5838 | @findex cacosl | |
1331d16f | 5839 | @findex calloc |
11bf0eb0 KG |
5840 | @findex carg |
5841 | @findex cargf | |
5842 | @findex cargl | |
5843 | @findex casin | |
5844 | @findex casinf | |
5845 | @findex casinh | |
5846 | @findex casinhf | |
5847 | @findex casinhl | |
5848 | @findex casinl | |
5849 | @findex catan | |
5850 | @findex catanf | |
5851 | @findex catanh | |
5852 | @findex catanhf | |
5853 | @findex catanhl | |
5854 | @findex catanl | |
98ff7c4d KG |
5855 | @findex cbrt |
5856 | @findex cbrtf | |
5857 | @findex cbrtl | |
11bf0eb0 KG |
5858 | @findex ccos |
5859 | @findex ccosf | |
5860 | @findex ccosh | |
5861 | @findex ccoshf | |
5862 | @findex ccoshl | |
5863 | @findex ccosl | |
b052d8ee RS |
5864 | @findex ceil |
5865 | @findex ceilf | |
5866 | @findex ceill | |
11bf0eb0 KG |
5867 | @findex cexp |
5868 | @findex cexpf | |
5869 | @findex cexpl | |
341e3d11 JM |
5870 | @findex cimag |
5871 | @findex cimagf | |
5872 | @findex cimagl | |
c3887ef2 PC |
5873 | @findex clog |
5874 | @findex clogf | |
5875 | @findex clogl | |
341e3d11 JM |
5876 | @findex conj |
5877 | @findex conjf | |
5878 | @findex conjl | |
98ff7c4d KG |
5879 | @findex copysign |
5880 | @findex copysignf | |
5881 | @findex copysignl | |
01702459 JM |
5882 | @findex cos |
5883 | @findex cosf | |
98ff7c4d KG |
5884 | @findex cosh |
5885 | @findex coshf | |
5886 | @findex coshl | |
01702459 | 5887 | @findex cosl |
11bf0eb0 KG |
5888 | @findex cpow |
5889 | @findex cpowf | |
5890 | @findex cpowl | |
5891 | @findex cproj | |
5892 | @findex cprojf | |
5893 | @findex cprojl | |
341e3d11 JM |
5894 | @findex creal |
5895 | @findex crealf | |
5896 | @findex creall | |
11bf0eb0 KG |
5897 | @findex csin |
5898 | @findex csinf | |
5899 | @findex csinh | |
5900 | @findex csinhf | |
5901 | @findex csinhl | |
5902 | @findex csinl | |
5903 | @findex csqrt | |
5904 | @findex csqrtf | |
5905 | @findex csqrtl | |
5906 | @findex ctan | |
5907 | @findex ctanf | |
5908 | @findex ctanh | |
5909 | @findex ctanhf | |
5910 | @findex ctanhl | |
5911 | @findex ctanl | |
178b2b9f RS |
5912 | @findex dcgettext |
5913 | @findex dgettext | |
98ff7c4d KG |
5914 | @findex drem |
5915 | @findex dremf | |
5916 | @findex dreml | |
488f17e1 KG |
5917 | @findex erf |
5918 | @findex erfc | |
5919 | @findex erfcf | |
5920 | @findex erfcl | |
5921 | @findex erff | |
5922 | @findex erfl | |
01702459 | 5923 | @findex exit |
e7b489c8 | 5924 | @findex exp |
98ff7c4d KG |
5925 | @findex exp10 |
5926 | @findex exp10f | |
5927 | @findex exp10l | |
5928 | @findex exp2 | |
5929 | @findex exp2f | |
5930 | @findex exp2l | |
e7b489c8 RS |
5931 | @findex expf |
5932 | @findex expl | |
98ff7c4d KG |
5933 | @findex expm1 |
5934 | @findex expm1f | |
5935 | @findex expm1l | |
01702459 JM |
5936 | @findex fabs |
5937 | @findex fabsf | |
5938 | @findex fabsl | |
98ff7c4d KG |
5939 | @findex fdim |
5940 | @findex fdimf | |
5941 | @findex fdiml | |
01702459 | 5942 | @findex ffs |
b052d8ee RS |
5943 | @findex floor |
5944 | @findex floorf | |
5945 | @findex floorl | |
98ff7c4d KG |
5946 | @findex fma |
5947 | @findex fmaf | |
5948 | @findex fmal | |
5949 | @findex fmax | |
5950 | @findex fmaxf | |
5951 | @findex fmaxl | |
5952 | @findex fmin | |
5953 | @findex fminf | |
5954 | @findex fminl | |
b052d8ee RS |
5955 | @findex fmod |
5956 | @findex fmodf | |
5957 | @findex fmodl | |
18f988a0 | 5958 | @findex fprintf |
b4c984fb | 5959 | @findex fprintf_unlocked |
01702459 | 5960 | @findex fputs |
b4c984fb | 5961 | @findex fputs_unlocked |
a2a919aa KG |
5962 | @findex frexp |
5963 | @findex frexpf | |
5964 | @findex frexpl | |
178b2b9f | 5965 | @findex fscanf |
488f17e1 KG |
5966 | @findex gamma |
5967 | @findex gammaf | |
5968 | @findex gammal | |
bf460eec KG |
5969 | @findex gamma_r |
5970 | @findex gammaf_r | |
5971 | @findex gammal_r | |
178b2b9f | 5972 | @findex gettext |
98ff7c4d KG |
5973 | @findex hypot |
5974 | @findex hypotf | |
5975 | @findex hypotl | |
5976 | @findex ilogb | |
5977 | @findex ilogbf | |
5978 | @findex ilogbl | |
e78f4a97 | 5979 | @findex imaxabs |
c7b6c6cd | 5980 | @findex index |
740e5b6f KG |
5981 | @findex isalnum |
5982 | @findex isalpha | |
5983 | @findex isascii | |
5984 | @findex isblank | |
5985 | @findex iscntrl | |
5986 | @findex isdigit | |
5987 | @findex isgraph | |
5988 | @findex islower | |
5989 | @findex isprint | |
5990 | @findex ispunct | |
5991 | @findex isspace | |
5992 | @findex isupper | |
ca4944e1 KG |
5993 | @findex iswalnum |
5994 | @findex iswalpha | |
5995 | @findex iswblank | |
5996 | @findex iswcntrl | |
5997 | @findex iswdigit | |
5998 | @findex iswgraph | |
5999 | @findex iswlower | |
6000 | @findex iswprint | |
6001 | @findex iswpunct | |
6002 | @findex iswspace | |
6003 | @findex iswupper | |
6004 | @findex iswxdigit | |
740e5b6f | 6005 | @findex isxdigit |
488f17e1 KG |
6006 | @findex j0 |
6007 | @findex j0f | |
6008 | @findex j0l | |
6009 | @findex j1 | |
6010 | @findex j1f | |
6011 | @findex j1l | |
6012 | @findex jn | |
6013 | @findex jnf | |
6014 | @findex jnl | |
01702459 | 6015 | @findex labs |
98ff7c4d KG |
6016 | @findex ldexp |
6017 | @findex ldexpf | |
6018 | @findex ldexpl | |
488f17e1 KG |
6019 | @findex lgamma |
6020 | @findex lgammaf | |
6021 | @findex lgammal | |
bf460eec KG |
6022 | @findex lgamma_r |
6023 | @findex lgammaf_r | |
6024 | @findex lgammal_r | |
01702459 | 6025 | @findex llabs |
98ff7c4d KG |
6026 | @findex llrint |
6027 | @findex llrintf | |
6028 | @findex llrintl | |
6029 | @findex llround | |
6030 | @findex llroundf | |
6031 | @findex llroundl | |
e7b489c8 | 6032 | @findex log |
98ff7c4d KG |
6033 | @findex log10 |
6034 | @findex log10f | |
6035 | @findex log10l | |
6036 | @findex log1p | |
6037 | @findex log1pf | |
6038 | @findex log1pl | |
6039 | @findex log2 | |
6040 | @findex log2f | |
6041 | @findex log2l | |
6042 | @findex logb | |
6043 | @findex logbf | |
6044 | @findex logbl | |
e7b489c8 RS |
6045 | @findex logf |
6046 | @findex logl | |
98ff7c4d KG |
6047 | @findex lrint |
6048 | @findex lrintf | |
6049 | @findex lrintl | |
6050 | @findex lround | |
6051 | @findex lroundf | |
6052 | @findex lroundl | |
1331d16f | 6053 | @findex malloc |
2a5fce6d | 6054 | @findex memchr |
01702459 JM |
6055 | @findex memcmp |
6056 | @findex memcpy | |
9cb65f92 | 6057 | @findex mempcpy |
01702459 | 6058 | @findex memset |
a2a919aa KG |
6059 | @findex modf |
6060 | @findex modff | |
6061 | @findex modfl | |
b052d8ee RS |
6062 | @findex nearbyint |
6063 | @findex nearbyintf | |
6064 | @findex nearbyintl | |
98ff7c4d KG |
6065 | @findex nextafter |
6066 | @findex nextafterf | |
6067 | @findex nextafterl | |
6068 | @findex nexttoward | |
6069 | @findex nexttowardf | |
6070 | @findex nexttowardl | |
46847aa6 | 6071 | @findex pow |
98ff7c4d KG |
6072 | @findex pow10 |
6073 | @findex pow10f | |
6074 | @findex pow10l | |
46847aa6 RS |
6075 | @findex powf |
6076 | @findex powl | |
01702459 | 6077 | @findex printf |
b4c984fb | 6078 | @findex printf_unlocked |
08291658 RS |
6079 | @findex putchar |
6080 | @findex puts | |
98ff7c4d KG |
6081 | @findex remainder |
6082 | @findex remainderf | |
6083 | @findex remainderl | |
a2a919aa KG |
6084 | @findex remquo |
6085 | @findex remquof | |
6086 | @findex remquol | |
c7b6c6cd | 6087 | @findex rindex |
98ff7c4d KG |
6088 | @findex rint |
6089 | @findex rintf | |
6090 | @findex rintl | |
b052d8ee RS |
6091 | @findex round |
6092 | @findex roundf | |
6093 | @findex roundl | |
98ff7c4d KG |
6094 | @findex scalb |
6095 | @findex scalbf | |
6096 | @findex scalbl | |
6097 | @findex scalbln | |
6098 | @findex scalblnf | |
6099 | @findex scalblnf | |
6100 | @findex scalbn | |
6101 | @findex scalbnf | |
6102 | @findex scanfnl | |
ef79730c RS |
6103 | @findex signbit |
6104 | @findex signbitf | |
6105 | @findex signbitl | |
44aea9ac JJ |
6106 | @findex signbitd32 |
6107 | @findex signbitd64 | |
6108 | @findex signbitd128 | |
488f17e1 KG |
6109 | @findex significand |
6110 | @findex significandf | |
6111 | @findex significandl | |
01702459 | 6112 | @findex sin |
a2a919aa KG |
6113 | @findex sincos |
6114 | @findex sincosf | |
6115 | @findex sincosl | |
01702459 | 6116 | @findex sinf |
98ff7c4d KG |
6117 | @findex sinh |
6118 | @findex sinhf | |
6119 | @findex sinhl | |
01702459 | 6120 | @findex sinl |
08291658 RS |
6121 | @findex snprintf |
6122 | @findex sprintf | |
01702459 JM |
6123 | @findex sqrt |
6124 | @findex sqrtf | |
6125 | @findex sqrtl | |
08291658 | 6126 | @findex sscanf |
9cb65f92 | 6127 | @findex stpcpy |
e905ac64 KG |
6128 | @findex stpncpy |
6129 | @findex strcasecmp | |
d118937d | 6130 | @findex strcat |
01702459 JM |
6131 | @findex strchr |
6132 | @findex strcmp | |
6133 | @findex strcpy | |
d118937d | 6134 | @findex strcspn |
1331d16f | 6135 | @findex strdup |
178b2b9f RS |
6136 | @findex strfmon |
6137 | @findex strftime | |
01702459 | 6138 | @findex strlen |
e905ac64 | 6139 | @findex strncasecmp |
d118937d | 6140 | @findex strncat |
da9e9f08 KG |
6141 | @findex strncmp |
6142 | @findex strncpy | |
e905ac64 | 6143 | @findex strndup |
01702459 JM |
6144 | @findex strpbrk |
6145 | @findex strrchr | |
d118937d | 6146 | @findex strspn |
01702459 | 6147 | @findex strstr |
29f523be RS |
6148 | @findex tan |
6149 | @findex tanf | |
98ff7c4d KG |
6150 | @findex tanh |
6151 | @findex tanhf | |
6152 | @findex tanhl | |
29f523be | 6153 | @findex tanl |
488f17e1 KG |
6154 | @findex tgamma |
6155 | @findex tgammaf | |
6156 | @findex tgammal | |
740e5b6f KG |
6157 | @findex toascii |
6158 | @findex tolower | |
6159 | @findex toupper | |
ca4944e1 KG |
6160 | @findex towlower |
6161 | @findex towupper | |
4977bab6 ZW |
6162 | @findex trunc |
6163 | @findex truncf | |
6164 | @findex truncl | |
178b2b9f RS |
6165 | @findex vfprintf |
6166 | @findex vfscanf | |
08291658 RS |
6167 | @findex vprintf |
6168 | @findex vscanf | |
6169 | @findex vsnprintf | |
6170 | @findex vsprintf | |
6171 | @findex vsscanf | |
488f17e1 KG |
6172 | @findex y0 |
6173 | @findex y0f | |
6174 | @findex y0l | |
6175 | @findex y1 | |
6176 | @findex y1f | |
6177 | @findex y1l | |
6178 | @findex yn | |
6179 | @findex ynf | |
6180 | @findex ynl | |
185ebd6c | 6181 | |
f0523f02 | 6182 | GCC provides a large number of built-in functions other than the ones |
185ebd6c RH |
6183 | mentioned above. Some of these are for internal use in the processing |
6184 | of exceptions or variable-length argument lists and will not be | |
6185 | documented here because they may change from time to time; we do not | |
6186 | recommend general use of these functions. | |
6187 | ||
6188 | The remaining functions are provided for optimization purposes. | |
6189 | ||
84330467 | 6190 | @opindex fno-builtin |
9c34dbbf ZW |
6191 | GCC includes built-in versions of many of the functions in the standard |
6192 | C library. The versions prefixed with @code{__builtin_} will always be | |
6193 | treated as having the same meaning as the C library function even if you | |
8a36672b | 6194 | specify the @option{-fno-builtin} option. (@pxref{C Dialect Options}) |
9c34dbbf | 6195 | Many of these functions are only optimized in certain cases; if they are |
01702459 JM |
6196 | not optimized in a particular case, a call to the library function will |
6197 | be emitted. | |
6198 | ||
84330467 JM |
6199 | @opindex ansi |
6200 | @opindex std | |
b052d8ee | 6201 | Outside strict ISO C mode (@option{-ansi}, @option{-std=c89} or |
98ff7c4d KG |
6202 | @option{-std=c99}), the functions |
6203 | @code{_exit}, @code{alloca}, @code{bcmp}, @code{bzero}, | |
6204 | @code{dcgettext}, @code{dgettext}, @code{dremf}, @code{dreml}, | |
6205 | @code{drem}, @code{exp10f}, @code{exp10l}, @code{exp10}, @code{ffsll}, | |
bf460eec KG |
6206 | @code{ffsl}, @code{ffs}, @code{fprintf_unlocked}, |
6207 | @code{fputs_unlocked}, @code{gammaf}, @code{gammal}, @code{gamma}, | |
6208 | @code{gammaf_r}, @code{gammal_r}, @code{gamma_r}, @code{gettext}, | |
740e5b6f KG |
6209 | @code{index}, @code{isascii}, @code{j0f}, @code{j0l}, @code{j0}, |
6210 | @code{j1f}, @code{j1l}, @code{j1}, @code{jnf}, @code{jnl}, @code{jn}, | |
bf460eec KG |
6211 | @code{lgammaf_r}, @code{lgammal_r}, @code{lgamma_r}, @code{mempcpy}, |
6212 | @code{pow10f}, @code{pow10l}, @code{pow10}, @code{printf_unlocked}, | |
6213 | @code{rindex}, @code{scalbf}, @code{scalbl}, @code{scalb}, | |
6214 | @code{signbit}, @code{signbitf}, @code{signbitl}, @code{signbitd32}, | |
6215 | @code{signbitd64}, @code{signbitd128}, @code{significandf}, | |
6216 | @code{significandl}, @code{significand}, @code{sincosf}, | |
6217 | @code{sincosl}, @code{sincos}, @code{stpcpy}, @code{stpncpy}, | |
6218 | @code{strcasecmp}, @code{strdup}, @code{strfmon}, @code{strncasecmp}, | |
6219 | @code{strndup}, @code{toascii}, @code{y0f}, @code{y0l}, @code{y0}, | |
6220 | @code{y1f}, @code{y1l}, @code{y1}, @code{ynf}, @code{ynl} and | |
6221 | @code{yn} | |
1331d16f | 6222 | may be handled as built-in functions. |
b052d8ee | 6223 | All these functions have corresponding versions |
9c34dbbf ZW |
6224 | prefixed with @code{__builtin_}, which may be used even in strict C89 |
6225 | mode. | |
01702459 | 6226 | |
075ec276 | 6227 | The ISO C99 functions |
98ff7c4d KG |
6228 | @code{_Exit}, @code{acoshf}, @code{acoshl}, @code{acosh}, @code{asinhf}, |
6229 | @code{asinhl}, @code{asinh}, @code{atanhf}, @code{atanhl}, @code{atanh}, | |
11bf0eb0 KG |
6230 | @code{cabsf}, @code{cabsl}, @code{cabs}, @code{cacosf}, @code{cacoshf}, |
6231 | @code{cacoshl}, @code{cacosh}, @code{cacosl}, @code{cacos}, | |
6232 | @code{cargf}, @code{cargl}, @code{carg}, @code{casinf}, @code{casinhf}, | |
6233 | @code{casinhl}, @code{casinh}, @code{casinl}, @code{casin}, | |
6234 | @code{catanf}, @code{catanhf}, @code{catanhl}, @code{catanh}, | |
6235 | @code{catanl}, @code{catan}, @code{cbrtf}, @code{cbrtl}, @code{cbrt}, | |
6236 | @code{ccosf}, @code{ccoshf}, @code{ccoshl}, @code{ccosh}, @code{ccosl}, | |
6237 | @code{ccos}, @code{cexpf}, @code{cexpl}, @code{cexp}, @code{cimagf}, | |
c3887ef2 PC |
6238 | @code{cimagl}, @code{cimag}, @code{clogf}, @code{clogl}, @code{clog}, |
6239 | @code{conjf}, @code{conjl}, @code{conj}, @code{copysignf}, @code{copysignl}, | |
6240 | @code{copysign}, @code{cpowf}, @code{cpowl}, @code{cpow}, @code{cprojf}, | |
6241 | @code{cprojl}, @code{cproj}, @code{crealf}, @code{creall}, @code{creal}, | |
6242 | @code{csinf}, @code{csinhf}, @code{csinhl}, @code{csinh}, @code{csinl}, | |
6243 | @code{csin}, @code{csqrtf}, @code{csqrtl}, @code{csqrt}, @code{ctanf}, | |
6244 | @code{ctanhf}, @code{ctanhl}, @code{ctanh}, @code{ctanl}, @code{ctan}, | |
6245 | @code{erfcf}, @code{erfcl}, @code{erfc}, @code{erff}, @code{erfl}, | |
6246 | @code{erf}, @code{exp2f}, @code{exp2l}, @code{exp2}, @code{expm1f}, | |
6247 | @code{expm1l}, @code{expm1}, @code{fdimf}, @code{fdiml}, @code{fdim}, | |
6248 | @code{fmaf}, @code{fmal}, @code{fmaxf}, @code{fmaxl}, @code{fmax}, | |
6249 | @code{fma}, @code{fminf}, @code{fminl}, @code{fmin}, @code{hypotf}, | |
6250 | @code{hypotl}, @code{hypot}, @code{ilogbf}, @code{ilogbl}, @code{ilogb}, | |
6251 | @code{imaxabs}, @code{isblank}, @code{iswblank}, @code{lgammaf}, | |
6252 | @code{lgammal}, @code{lgamma}, @code{llabs}, @code{llrintf}, @code{llrintl}, | |
ca4944e1 KG |
6253 | @code{llrint}, @code{llroundf}, @code{llroundl}, @code{llround}, |
6254 | @code{log1pf}, @code{log1pl}, @code{log1p}, @code{log2f}, @code{log2l}, | |
6255 | @code{log2}, @code{logbf}, @code{logbl}, @code{logb}, @code{lrintf}, | |
6256 | @code{lrintl}, @code{lrint}, @code{lroundf}, @code{lroundl}, | |
6257 | @code{lround}, @code{nearbyintf}, @code{nearbyintl}, @code{nearbyint}, | |
740e5b6f KG |
6258 | @code{nextafterf}, @code{nextafterl}, @code{nextafter}, |
6259 | @code{nexttowardf}, @code{nexttowardl}, @code{nexttoward}, | |
6260 | @code{remainderf}, @code{remainderl}, @code{remainder}, @code{remquof}, | |
6261 | @code{remquol}, @code{remquo}, @code{rintf}, @code{rintl}, @code{rint}, | |
6262 | @code{roundf}, @code{roundl}, @code{round}, @code{scalblnf}, | |
6263 | @code{scalblnl}, @code{scalbln}, @code{scalbnf}, @code{scalbnl}, | |
6264 | @code{scalbn}, @code{snprintf}, @code{tgammaf}, @code{tgammal}, | |
6265 | @code{tgamma}, @code{truncf}, @code{truncl}, @code{trunc}, | |
6266 | @code{vfscanf}, @code{vscanf}, @code{vsnprintf} and @code{vsscanf} | |
08291658 | 6267 | are handled as built-in functions |
b052d8ee | 6268 | except in strict ISO C90 mode (@option{-ansi} or @option{-std=c89}). |
46847aa6 | 6269 | |
98ff7c4d KG |
6270 | There are also built-in versions of the ISO C99 functions |
6271 | @code{acosf}, @code{acosl}, @code{asinf}, @code{asinl}, @code{atan2f}, | |
29f523be | 6272 | @code{atan2l}, @code{atanf}, @code{atanl}, @code{ceilf}, @code{ceill}, |
98ff7c4d KG |
6273 | @code{cosf}, @code{coshf}, @code{coshl}, @code{cosl}, @code{expf}, |
6274 | @code{expl}, @code{fabsf}, @code{fabsl}, @code{floorf}, @code{floorl}, | |
a2a919aa KG |
6275 | @code{fmodf}, @code{fmodl}, @code{frexpf}, @code{frexpl}, @code{ldexpf}, |
6276 | @code{ldexpl}, @code{log10f}, @code{log10l}, @code{logf}, @code{logl}, | |
6277 | @code{modfl}, @code{modf}, @code{powf}, @code{powl}, @code{sinf}, | |
6278 | @code{sinhf}, @code{sinhl}, @code{sinl}, @code{sqrtf}, @code{sqrtl}, | |
6279 | @code{tanf}, @code{tanhf}, @code{tanhl} and @code{tanl} | |
46847aa6 RS |
6280 | that are recognized in any mode since ISO C90 reserves these names for |
6281 | the purpose to which ISO C99 puts them. All these functions have | |
6282 | corresponding versions prefixed with @code{__builtin_}. | |
6283 | ||
ca4944e1 KG |
6284 | The ISO C94 functions |
6285 | @code{iswalnum}, @code{iswalpha}, @code{iswcntrl}, @code{iswdigit}, | |
6286 | @code{iswgraph}, @code{iswlower}, @code{iswprint}, @code{iswpunct}, | |
6287 | @code{iswspace}, @code{iswupper}, @code{iswxdigit}, @code{towlower} and | |
6288 | @code{towupper} | |
6289 | are handled as built-in functions | |
6290 | except in strict ISO C90 mode (@option{-ansi} or @option{-std=c89}). | |
6291 | ||
98ff7c4d KG |
6292 | The ISO C90 functions |
6293 | @code{abort}, @code{abs}, @code{acos}, @code{asin}, @code{atan2}, | |
6294 | @code{atan}, @code{calloc}, @code{ceil}, @code{cosh}, @code{cos}, | |
6295 | @code{exit}, @code{exp}, @code{fabs}, @code{floor}, @code{fmod}, | |
740e5b6f KG |
6296 | @code{fprintf}, @code{fputs}, @code{frexp}, @code{fscanf}, |
6297 | @code{isalnum}, @code{isalpha}, @code{iscntrl}, @code{isdigit}, | |
6298 | @code{isgraph}, @code{islower}, @code{isprint}, @code{ispunct}, | |
6299 | @code{isspace}, @code{isupper}, @code{isxdigit}, @code{tolower}, | |
6300 | @code{toupper}, @code{labs}, @code{ldexp}, @code{log10}, @code{log}, | |
2a5fce6d PC |
6301 | @code{malloc}, @code{memchr}, @code{memcmp}, @code{memcpy}, |
6302 | @code{memset}, @code{modf}, @code{pow}, @code{printf}, @code{putchar}, | |
6303 | @code{puts}, @code{scanf}, @code{sinh}, @code{sin}, @code{snprintf}, | |
6304 | @code{sprintf}, @code{sqrt}, @code{sscanf}, @code{strcat}, | |
6305 | @code{strchr}, @code{strcmp}, @code{strcpy}, @code{strcspn}, | |
6306 | @code{strlen}, @code{strncat}, @code{strncmp}, @code{strncpy}, | |
6307 | @code{strpbrk}, @code{strrchr}, @code{strspn}, @code{strstr}, | |
6308 | @code{tanh}, @code{tan}, @code{vfprintf}, @code{vprintf} and @code{vsprintf} | |
08291658 | 6309 | are all recognized as built-in functions unless |
46847aa6 RS |
6310 | @option{-fno-builtin} is specified (or @option{-fno-builtin-@var{function}} |
6311 | is specified for an individual function). All of these functions have | |
4977bab6 | 6312 | corresponding versions prefixed with @code{__builtin_}. |
9c34dbbf ZW |
6313 | |
6314 | GCC provides built-in versions of the ISO C99 floating point comparison | |
6315 | macros that avoid raising exceptions for unordered operands. They have | |
6316 | the same names as the standard macros ( @code{isgreater}, | |
6317 | @code{isgreaterequal}, @code{isless}, @code{islessequal}, | |
6318 | @code{islessgreater}, and @code{isunordered}) , with @code{__builtin_} | |
6319 | prefixed. We intend for a library implementor to be able to simply | |
6320 | @code{#define} each standard macro to its built-in equivalent. | |
3bf5906b KG |
6321 | In the same fashion, GCC provides @code{fpclassify}, @code{isfinite}, |
6322 | @code{isinf_sign} and @code{isnormal} built-ins used with | |
6323 | @code{__builtin_} prefixed. The @code{isinf} and @code{isnan} | |
6324 | builtins appear both with and without the @code{__builtin_} prefix. | |
185ebd6c | 6325 | |
ecbcf7b3 AH |
6326 | @deftypefn {Built-in Function} int __builtin_types_compatible_p (@var{type1}, @var{type2}) |
6327 | ||
6328 | You can use the built-in function @code{__builtin_types_compatible_p} to | |
6329 | determine whether two types are the same. | |
6330 | ||
6331 | This built-in function returns 1 if the unqualified versions of the | |
6332 | types @var{type1} and @var{type2} (which are types, not expressions) are | |
6333 | compatible, 0 otherwise. The result of this built-in function can be | |
6334 | used in integer constant expressions. | |
6335 | ||
6336 | This built-in function ignores top level qualifiers (e.g., @code{const}, | |
6337 | @code{volatile}). For example, @code{int} is equivalent to @code{const | |
6338 | int}. | |
6339 | ||
6340 | The type @code{int[]} and @code{int[5]} are compatible. On the other | |
6341 | hand, @code{int} and @code{char *} are not compatible, even if the size | |
6342 | of their types, on the particular architecture are the same. Also, the | |
6343 | amount of pointer indirection is taken into account when determining | |
6344 | similarity. Consequently, @code{short *} is not similar to | |
6345 | @code{short **}. Furthermore, two types that are typedefed are | |
6346 | considered compatible if their underlying types are compatible. | |
6347 | ||
bca63328 JM |
6348 | An @code{enum} type is not considered to be compatible with another |
6349 | @code{enum} type even if both are compatible with the same integer | |
6350 | type; this is what the C standard specifies. | |
6351 | For example, @code{enum @{foo, bar@}} is not similar to | |
ecbcf7b3 AH |
6352 | @code{enum @{hot, dog@}}. |
6353 | ||
6354 | You would typically use this function in code whose execution varies | |
6355 | depending on the arguments' types. For example: | |
6356 | ||
6357 | @smallexample | |
6e5bb5ad JM |
6358 | #define foo(x) \ |
6359 | (@{ \ | |
b7886f14 | 6360 | typeof (x) tmp = (x); \ |
6e5bb5ad JM |
6361 | if (__builtin_types_compatible_p (typeof (x), long double)) \ |
6362 | tmp = foo_long_double (tmp); \ | |
6363 | else if (__builtin_types_compatible_p (typeof (x), double)) \ | |
6364 | tmp = foo_double (tmp); \ | |
6365 | else if (__builtin_types_compatible_p (typeof (x), float)) \ | |
6366 | tmp = foo_float (tmp); \ | |
6367 | else \ | |
6368 | abort (); \ | |
6369 | tmp; \ | |
ecbcf7b3 AH |
6370 | @}) |
6371 | @end smallexample | |
6372 | ||
8a36672b | 6373 | @emph{Note:} This construct is only available for C@. |
ecbcf7b3 AH |
6374 | |
6375 | @end deftypefn | |
6376 | ||
6377 | @deftypefn {Built-in Function} @var{type} __builtin_choose_expr (@var{const_exp}, @var{exp1}, @var{exp2}) | |
6378 | ||
6379 | You can use the built-in function @code{__builtin_choose_expr} to | |
6380 | evaluate code depending on the value of a constant expression. This | |
6381 | built-in function returns @var{exp1} if @var{const_exp}, which is a | |
6382 | constant expression that must be able to be determined at compile time, | |
6383 | is nonzero. Otherwise it returns 0. | |
6384 | ||
6385 | This built-in function is analogous to the @samp{? :} operator in C, | |
6386 | except that the expression returned has its type unaltered by promotion | |
6387 | rules. Also, the built-in function does not evaluate the expression | |
6388 | that was not chosen. For example, if @var{const_exp} evaluates to true, | |
6389 | @var{exp2} is not evaluated even if it has side-effects. | |
6390 | ||
6391 | This built-in function can return an lvalue if the chosen argument is an | |
6392 | lvalue. | |
6393 | ||
6394 | If @var{exp1} is returned, the return type is the same as @var{exp1}'s | |
6395 | type. Similarly, if @var{exp2} is returned, its return type is the same | |
6396 | as @var{exp2}. | |
6397 | ||
6398 | Example: | |
6399 | ||
6400 | @smallexample | |
478c9e72 JJ |
6401 | #define foo(x) \ |
6402 | __builtin_choose_expr ( \ | |
6403 | __builtin_types_compatible_p (typeof (x), double), \ | |
6404 | foo_double (x), \ | |
6405 | __builtin_choose_expr ( \ | |
6406 | __builtin_types_compatible_p (typeof (x), float), \ | |
6407 | foo_float (x), \ | |
6408 | /* @r{The void expression results in a compile-time error} \ | |
6409 | @r{when assigning the result to something.} */ \ | |
ecbcf7b3 AH |
6410 | (void)0)) |
6411 | @end smallexample | |
6412 | ||
8a36672b | 6413 | @emph{Note:} This construct is only available for C@. Furthermore, the |
ecbcf7b3 AH |
6414 | unused expression (@var{exp1} or @var{exp2} depending on the value of |
6415 | @var{const_exp}) may still generate syntax errors. This may change in | |
6416 | future revisions. | |
6417 | ||
6418 | @end deftypefn | |
6419 | ||
84330467 JM |
6420 | @deftypefn {Built-in Function} int __builtin_constant_p (@var{exp}) |
6421 | You can use the built-in function @code{__builtin_constant_p} to | |
185ebd6c | 6422 | determine if a value is known to be constant at compile-time and hence |
f0523f02 | 6423 | that GCC can perform constant-folding on expressions involving that |
185ebd6c RH |
6424 | value. The argument of the function is the value to test. The function |
6425 | returns the integer 1 if the argument is known to be a compile-time | |
6426 | constant and 0 if it is not known to be a compile-time constant. A | |
6427 | return of 0 does not indicate that the value is @emph{not} a constant, | |
f0523f02 | 6428 | but merely that GCC cannot prove it is a constant with the specified |
84330467 | 6429 | value of the @option{-O} option. |
185ebd6c RH |
6430 | |
6431 | You would typically use this function in an embedded application where | |
6432 | memory was a critical resource. If you have some complex calculation, | |
6433 | you may want it to be folded if it involves constants, but need to call | |
6434 | a function if it does not. For example: | |
6435 | ||
4d390518 | 6436 | @smallexample |
310668e8 JM |
6437 | #define Scale_Value(X) \ |
6438 | (__builtin_constant_p (X) \ | |
6439 | ? ((X) * SCALE + OFFSET) : Scale (X)) | |
185ebd6c RH |
6440 | @end smallexample |
6441 | ||
84330467 | 6442 | You may use this built-in function in either a macro or an inline |
185ebd6c | 6443 | function. However, if you use it in an inlined function and pass an |
f0523f02 | 6444 | argument of the function as the argument to the built-in, GCC will |
185ebd6c | 6445 | never return 1 when you call the inline function with a string constant |
4b404517 | 6446 | or compound literal (@pxref{Compound Literals}) and will not return 1 |
185ebd6c | 6447 | when you pass a constant numeric value to the inline function unless you |
84330467 | 6448 | specify the @option{-O} option. |
13104975 ZW |
6449 | |
6450 | You may also use @code{__builtin_constant_p} in initializers for static | |
6451 | data. For instance, you can write | |
6452 | ||
6453 | @smallexample | |
79323c50 | 6454 | static const int table[] = @{ |
13104975 | 6455 | __builtin_constant_p (EXPRESSION) ? (EXPRESSION) : -1, |
0d893a63 | 6456 | /* @r{@dots{}} */ |
79323c50 | 6457 | @}; |
13104975 ZW |
6458 | @end smallexample |
6459 | ||
6460 | @noindent | |
6461 | This is an acceptable initializer even if @var{EXPRESSION} is not a | |
6462 | constant expression. GCC must be more conservative about evaluating the | |
6463 | built-in in this case, because it has no opportunity to perform | |
6464 | optimization. | |
6465 | ||
6466 | Previous versions of GCC did not accept this built-in in data | |
6467 | initializers. The earliest version where it is completely safe is | |
6468 | 3.0.1. | |
84330467 | 6469 | @end deftypefn |
185ebd6c | 6470 | |
84330467 JM |
6471 | @deftypefn {Built-in Function} long __builtin_expect (long @var{exp}, long @var{c}) |
6472 | @opindex fprofile-arcs | |
02f52e19 | 6473 | You may use @code{__builtin_expect} to provide the compiler with |
994a57cd | 6474 | branch prediction information. In general, you should prefer to |
84330467 | 6475 | use actual profile feedback for this (@option{-fprofile-arcs}), as |
994a57cd | 6476 | programmers are notoriously bad at predicting how their programs |
60b6e1f5 | 6477 | actually perform. However, there are applications in which this |
994a57cd RH |
6478 | data is hard to collect. |
6479 | ||
ef950eba JH |
6480 | The return value is the value of @var{exp}, which should be an integral |
6481 | expression. The semantics of the built-in are that it is expected that | |
6482 | @var{exp} == @var{c}. For example: | |
994a57cd RH |
6483 | |
6484 | @smallexample | |
6485 | if (__builtin_expect (x, 0)) | |
6486 | foo (); | |
6487 | @end smallexample | |
6488 | ||
6489 | @noindent | |
6490 | would indicate that we do not expect to call @code{foo}, since | |
6491 | we expect @code{x} to be zero. Since you are limited to integral | |
6492 | expressions for @var{exp}, you should use constructions such as | |
6493 | ||
6494 | @smallexample | |
6495 | if (__builtin_expect (ptr != NULL, 1)) | |
6496 | error (); | |
6497 | @end smallexample | |
6498 | ||
6499 | @noindent | |
6500 | when testing pointer or floating-point values. | |
84330467 | 6501 | @end deftypefn |
994a57cd | 6502 | |
a18c20ec AP |
6503 | @deftypefn {Built-in Function} void __builtin_trap (void) |
6504 | This function causes the program to exit abnormally. GCC implements | |
6505 | this function by using a target-dependent mechanism (such as | |
6506 | intentionally executing an illegal instruction) or by calling | |
6507 | @code{abort}. The mechanism used may vary from release to release so | |
6508 | you should not rely on any particular implementation. | |
6509 | @end deftypefn | |
6510 | ||
677feb77 DD |
6511 | @deftypefn {Built-in Function} void __builtin___clear_cache (char *@var{begin}, char *@var{end}) |
6512 | This function is used to flush the processor's instruction cache for | |
6513 | the region of memory between @var{begin} inclusive and @var{end} | |
6514 | exclusive. Some targets require that the instruction cache be | |
6515 | flushed, after modifying memory containing code, in order to obtain | |
6516 | deterministic behavior. | |
6517 | ||
6518 | If the target does not require instruction cache flushes, | |
6519 | @code{__builtin___clear_cache} has no effect. Otherwise either | |
6520 | instructions are emitted in-line to clear the instruction cache or a | |
6521 | call to the @code{__clear_cache} function in libgcc is made. | |
6522 | @end deftypefn | |
6523 | ||
3bca17dd | 6524 | @deftypefn {Built-in Function} void __builtin_prefetch (const void *@var{addr}, ...) |
a9ccbb60 JJ |
6525 | This function is used to minimize cache-miss latency by moving data into |
6526 | a cache before it is accessed. | |
6527 | You can insert calls to @code{__builtin_prefetch} into code for which | |
6528 | you know addresses of data in memory that is likely to be accessed soon. | |
6529 | If the target supports them, data prefetch instructions will be generated. | |
6530 | If the prefetch is done early enough before the access then the data will | |
6531 | be in the cache by the time it is accessed. | |
6532 | ||
6533 | The value of @var{addr} is the address of the memory to prefetch. | |
e83d297b | 6534 | There are two optional arguments, @var{rw} and @var{locality}. |
a9ccbb60 | 6535 | The value of @var{rw} is a compile-time constant one or zero; one |
e83d297b JJ |
6536 | means that the prefetch is preparing for a write to the memory address |
6537 | and zero, the default, means that the prefetch is preparing for a read. | |
a9ccbb60 JJ |
6538 | The value @var{locality} must be a compile-time constant integer between |
6539 | zero and three. A value of zero means that the data has no temporal | |
6540 | locality, so it need not be left in the cache after the access. A value | |
6541 | of three means that the data has a high degree of temporal locality and | |
6542 | should be left in all levels of cache possible. Values of one and two | |
e83d297b JJ |
6543 | mean, respectively, a low or moderate degree of temporal locality. The |
6544 | default is three. | |
a9ccbb60 JJ |
6545 | |
6546 | @smallexample | |
6547 | for (i = 0; i < n; i++) | |
6548 | @{ | |
6549 | a[i] = a[i] + b[i]; | |
6550 | __builtin_prefetch (&a[i+j], 1, 1); | |
6551 | __builtin_prefetch (&b[i+j], 0, 1); | |
0d893a63 | 6552 | /* @r{@dots{}} */ |
a9ccbb60 JJ |
6553 | @} |
6554 | @end smallexample | |
6555 | ||
f282ffb3 | 6556 | Data prefetch does not generate faults if @var{addr} is invalid, but |
a9ccbb60 JJ |
6557 | the address expression itself must be valid. For example, a prefetch |
6558 | of @code{p->next} will not fault if @code{p->next} is not a valid | |
6559 | address, but evaluation will fault if @code{p} is not a valid address. | |
6560 | ||
6561 | If the target does not support data prefetch, the address expression | |
6562 | is evaluated if it includes side effects but no other code is generated | |
6563 | and GCC does not issue a warning. | |
6564 | @end deftypefn | |
6565 | ||
ab5e2615 RH |
6566 | @deftypefn {Built-in Function} double __builtin_huge_val (void) |
6567 | Returns a positive infinity, if supported by the floating-point format, | |
6568 | else @code{DBL_MAX}. This function is suitable for implementing the | |
6569 | ISO C macro @code{HUGE_VAL}. | |
6570 | @end deftypefn | |
6571 | ||
6572 | @deftypefn {Built-in Function} float __builtin_huge_valf (void) | |
6573 | Similar to @code{__builtin_huge_val}, except the return type is @code{float}. | |
6574 | @end deftypefn | |
6575 | ||
dad78426 | 6576 | @deftypefn {Built-in Function} {long double} __builtin_huge_vall (void) |
ab5e2615 RH |
6577 | Similar to @code{__builtin_huge_val}, except the return |
6578 | type is @code{long double}. | |
6579 | @end deftypefn | |
6580 | ||
3bf5906b KG |
6581 | @deftypefn {Built-in Function} int __builtin_fpclassify (int, int, int, int, int, ...) |
6582 | This built-in implements the C99 fpclassify functionality. The first | |
6583 | five int arguments should be the target library's notion of the | |
6584 | possible FP classes and are used for return values. They must be | |
6585 | constant values and they must appear in this order: @code{FP_NAN}, | |
32101f99 | 6586 | @code{FP_INFINITE}, @code{FP_NORMAL}, @code{FP_SUBNORMAL} and |
3bf5906b KG |
6587 | @code{FP_ZERO}. The ellipsis is for exactly one floating point value |
6588 | to classify. GCC treats the last argument as type-generic, which | |
6589 | means it does not do default promotion from float to double. | |
6590 | @end deftypefn | |
6591 | ||
ab5e2615 RH |
6592 | @deftypefn {Built-in Function} double __builtin_inf (void) |
6593 | Similar to @code{__builtin_huge_val}, except a warning is generated | |
6594 | if the target floating-point format does not support infinities. | |
ab5e2615 RH |
6595 | @end deftypefn |
6596 | ||
9a8ce21f JG |
6597 | @deftypefn {Built-in Function} _Decimal32 __builtin_infd32 (void) |
6598 | Similar to @code{__builtin_inf}, except the return type is @code{_Decimal32}. | |
6599 | @end deftypefn | |
6600 | ||
6601 | @deftypefn {Built-in Function} _Decimal64 __builtin_infd64 (void) | |
6602 | Similar to @code{__builtin_inf}, except the return type is @code{_Decimal64}. | |
6603 | @end deftypefn | |
6604 | ||
6605 | @deftypefn {Built-in Function} _Decimal128 __builtin_infd128 (void) | |
6606 | Similar to @code{__builtin_inf}, except the return type is @code{_Decimal128}. | |
6607 | @end deftypefn | |
6608 | ||
ab5e2615 RH |
6609 | @deftypefn {Built-in Function} float __builtin_inff (void) |
6610 | Similar to @code{__builtin_inf}, except the return type is @code{float}. | |
9c86fc0b | 6611 | This function is suitable for implementing the ISO C99 macro @code{INFINITY}. |
ab5e2615 RH |
6612 | @end deftypefn |
6613 | ||
dad78426 | 6614 | @deftypefn {Built-in Function} {long double} __builtin_infl (void) |
ab5e2615 RH |
6615 | Similar to @code{__builtin_inf}, except the return |
6616 | type is @code{long double}. | |
6617 | @end deftypefn | |
6618 | ||
05f41289 KG |
6619 | @deftypefn {Built-in Function} int __builtin_isinf_sign (...) |
6620 | Similar to @code{isinf}, except the return value will be negative for | |
6621 | an argument of @code{-Inf}. Note while the parameter list is an | |
6622 | ellipsis, this function only accepts exactly one floating point | |
6623 | argument. GCC treats this parameter as type-generic, which means it | |
6624 | does not do default promotion from float to double. | |
6625 | @end deftypefn | |
6626 | ||
1472e41c RH |
6627 | @deftypefn {Built-in Function} double __builtin_nan (const char *str) |
6628 | This is an implementation of the ISO C99 function @code{nan}. | |
6629 | ||
6630 | Since ISO C99 defines this function in terms of @code{strtod}, which we | |
c0478a66 | 6631 | do not implement, a description of the parsing is in order. The string |
1472e41c RH |
6632 | is parsed as by @code{strtol}; that is, the base is recognized by |
6633 | leading @samp{0} or @samp{0x} prefixes. The number parsed is placed | |
6634 | in the significand such that the least significant bit of the number | |
daf2f129 | 6635 | is at the least significant bit of the significand. The number is |
1472e41c | 6636 | truncated to fit the significand field provided. The significand is |
8a36672b | 6637 | forced to be a quiet NaN@. |
1472e41c | 6638 | |
a7d37464 GK |
6639 | This function, if given a string literal all of which would have been |
6640 | consumed by strtol, is evaluated early enough that it is considered a | |
6641 | compile-time constant. | |
1472e41c RH |
6642 | @end deftypefn |
6643 | ||
9a8ce21f JG |
6644 | @deftypefn {Built-in Function} _Decimal32 __builtin_nand32 (const char *str) |
6645 | Similar to @code{__builtin_nan}, except the return type is @code{_Decimal32}. | |
6646 | @end deftypefn | |
6647 | ||
6648 | @deftypefn {Built-in Function} _Decimal64 __builtin_nand64 (const char *str) | |
6649 | Similar to @code{__builtin_nan}, except the return type is @code{_Decimal64}. | |
6650 | @end deftypefn | |
6651 | ||
6652 | @deftypefn {Built-in Function} _Decimal128 __builtin_nand128 (const char *str) | |
6653 | Similar to @code{__builtin_nan}, except the return type is @code{_Decimal128}. | |
6654 | @end deftypefn | |
6655 | ||
1472e41c RH |
6656 | @deftypefn {Built-in Function} float __builtin_nanf (const char *str) |
6657 | Similar to @code{__builtin_nan}, except the return type is @code{float}. | |
6658 | @end deftypefn | |
6659 | ||
dad78426 | 6660 | @deftypefn {Built-in Function} {long double} __builtin_nanl (const char *str) |
1472e41c RH |
6661 | Similar to @code{__builtin_nan}, except the return type is @code{long double}. |
6662 | @end deftypefn | |
6663 | ||
6664 | @deftypefn {Built-in Function} double __builtin_nans (const char *str) | |
daf2f129 | 6665 | Similar to @code{__builtin_nan}, except the significand is forced |
8a36672b | 6666 | to be a signaling NaN@. The @code{nans} function is proposed by |
aaa67502 | 6667 | @uref{http://www.open-std.org/jtc1/sc22/wg14/www/docs/n965.htm,,WG14 N965}. |
1472e41c RH |
6668 | @end deftypefn |
6669 | ||
6670 | @deftypefn {Built-in Function} float __builtin_nansf (const char *str) | |
6671 | Similar to @code{__builtin_nans}, except the return type is @code{float}. | |
6672 | @end deftypefn | |
6673 | ||
dad78426 | 6674 | @deftypefn {Built-in Function} {long double} __builtin_nansl (const char *str) |
1472e41c RH |
6675 | Similar to @code{__builtin_nans}, except the return type is @code{long double}. |
6676 | @end deftypefn | |
6677 | ||
2928cd7a RH |
6678 | @deftypefn {Built-in Function} int __builtin_ffs (unsigned int x) |
6679 | Returns one plus the index of the least significant 1-bit of @var{x}, or | |
6680 | if @var{x} is zero, returns zero. | |
6681 | @end deftypefn | |
6682 | ||
6683 | @deftypefn {Built-in Function} int __builtin_clz (unsigned int x) | |
6684 | Returns the number of leading 0-bits in @var{x}, starting at the most | |
6685 | significant bit position. If @var{x} is 0, the result is undefined. | |
6686 | @end deftypefn | |
6687 | ||
6688 | @deftypefn {Built-in Function} int __builtin_ctz (unsigned int x) | |
6689 | Returns the number of trailing 0-bits in @var{x}, starting at the least | |
6690 | significant bit position. If @var{x} is 0, the result is undefined. | |
6691 | @end deftypefn | |
6692 | ||
6693 | @deftypefn {Built-in Function} int __builtin_popcount (unsigned int x) | |
6694 | Returns the number of 1-bits in @var{x}. | |
6695 | @end deftypefn | |
6696 | ||
6697 | @deftypefn {Built-in Function} int __builtin_parity (unsigned int x) | |
8a36672b | 6698 | Returns the parity of @var{x}, i.e.@: the number of 1-bits in @var{x} |
2928cd7a RH |
6699 | modulo 2. |
6700 | @end deftypefn | |
6701 | ||
6702 | @deftypefn {Built-in Function} int __builtin_ffsl (unsigned long) | |
6703 | Similar to @code{__builtin_ffs}, except the argument type is | |
6704 | @code{unsigned long}. | |
6705 | @end deftypefn | |
6706 | ||
6707 | @deftypefn {Built-in Function} int __builtin_clzl (unsigned long) | |
6708 | Similar to @code{__builtin_clz}, except the argument type is | |
6709 | @code{unsigned long}. | |
6710 | @end deftypefn | |
6711 | ||
6712 | @deftypefn {Built-in Function} int __builtin_ctzl (unsigned long) | |
6713 | Similar to @code{__builtin_ctz}, except the argument type is | |
6714 | @code{unsigned long}. | |
6715 | @end deftypefn | |
6716 | ||
6717 | @deftypefn {Built-in Function} int __builtin_popcountl (unsigned long) | |
6718 | Similar to @code{__builtin_popcount}, except the argument type is | |
6719 | @code{unsigned long}. | |
6720 | @end deftypefn | |
6721 | ||
6722 | @deftypefn {Built-in Function} int __builtin_parityl (unsigned long) | |
6723 | Similar to @code{__builtin_parity}, except the argument type is | |
6724 | @code{unsigned long}. | |
6725 | @end deftypefn | |
6726 | ||
6727 | @deftypefn {Built-in Function} int __builtin_ffsll (unsigned long long) | |
6728 | Similar to @code{__builtin_ffs}, except the argument type is | |
6729 | @code{unsigned long long}. | |
6730 | @end deftypefn | |
6731 | ||
6732 | @deftypefn {Built-in Function} int __builtin_clzll (unsigned long long) | |
6733 | Similar to @code{__builtin_clz}, except the argument type is | |
6734 | @code{unsigned long long}. | |
6735 | @end deftypefn | |
6736 | ||
6737 | @deftypefn {Built-in Function} int __builtin_ctzll (unsigned long long) | |
6738 | Similar to @code{__builtin_ctz}, except the argument type is | |
6739 | @code{unsigned long long}. | |
6740 | @end deftypefn | |
6741 | ||
6742 | @deftypefn {Built-in Function} int __builtin_popcountll (unsigned long long) | |
6743 | Similar to @code{__builtin_popcount}, except the argument type is | |
6744 | @code{unsigned long long}. | |
6745 | @end deftypefn | |
6746 | ||
6747 | @deftypefn {Built-in Function} int __builtin_parityll (unsigned long long) | |
6748 | Similar to @code{__builtin_parity}, except the argument type is | |
6749 | @code{unsigned long long}. | |
6750 | @end deftypefn | |
6751 | ||
17684d46 RG |
6752 | @deftypefn {Built-in Function} double __builtin_powi (double, int) |
6753 | Returns the first argument raised to the power of the second. Unlike the | |
6754 | @code{pow} function no guarantees about precision and rounding are made. | |
6755 | @end deftypefn | |
6756 | ||
6757 | @deftypefn {Built-in Function} float __builtin_powif (float, int) | |
6758 | Similar to @code{__builtin_powi}, except the argument and return types | |
6759 | are @code{float}. | |
6760 | @end deftypefn | |
6761 | ||
6762 | @deftypefn {Built-in Function} {long double} __builtin_powil (long double, int) | |
6763 | Similar to @code{__builtin_powi}, except the argument and return types | |
6764 | are @code{long double}. | |
6765 | @end deftypefn | |
6766 | ||
167fa32c EC |
6767 | @deftypefn {Built-in Function} int32_t __builtin_bswap32 (int32_t x) |
6768 | Returns @var{x} with the order of the bytes reversed; for example, | |
6769 | @code{0xaabbccdd} becomes @code{0xddccbbaa}. Byte here always means | |
6770 | exactly 8 bits. | |
6771 | @end deftypefn | |
6772 | ||
6773 | @deftypefn {Built-in Function} int64_t __builtin_bswap64 (int64_t x) | |
6774 | Similar to @code{__builtin_bswap32}, except the argument and return types | |
6775 | are 64-bit. | |
6776 | @end deftypefn | |
2928cd7a | 6777 | |
0975678f JM |
6778 | @node Target Builtins |
6779 | @section Built-in Functions Specific to Particular Target Machines | |
6780 | ||
6781 | On some target machines, GCC supports many built-in functions specific | |
6782 | to those machines. Generally these generate calls to specific machine | |
6783 | instructions, but allow the compiler to schedule those calls. | |
6784 | ||
6785 | @menu | |
6d8fd7bb | 6786 | * Alpha Built-in Functions:: |
88f77cba JB |
6787 | * ARM iWMMXt Built-in Functions:: |
6788 | * ARM NEON Intrinsics:: | |
161c21b6 | 6789 | * Blackfin Built-in Functions:: |
c3ee0579 | 6790 | * FR-V Built-in Functions:: |
0975678f | 6791 | * X86 Built-in Functions:: |
118ea793 | 6792 | * MIPS DSP Built-in Functions:: |
d840bfd3 | 6793 | * MIPS Paired-Single Support:: |
93581857 | 6794 | * MIPS Loongson Built-in Functions:: |
333c8841 | 6795 | * PowerPC AltiVec Built-in Functions:: |
c5145ceb | 6796 | * SPARC VIS Built-in Functions:: |
85d9c13c | 6797 | * SPU Built-in Functions:: |
0975678f JM |
6798 | @end menu |
6799 | ||
6d8fd7bb RH |
6800 | @node Alpha Built-in Functions |
6801 | @subsection Alpha Built-in Functions | |
6802 | ||
6803 | These built-in functions are available for the Alpha family of | |
6804 | processors, depending on the command-line switches used. | |
6805 | ||
95b1627e | 6806 | The following built-in functions are always available. They |
6d8fd7bb RH |
6807 | all generate the machine instruction that is part of the name. |
6808 | ||
3ab51846 | 6809 | @smallexample |
6d8fd7bb RH |
6810 | long __builtin_alpha_implver (void) |
6811 | long __builtin_alpha_rpcc (void) | |
6812 | long __builtin_alpha_amask (long) | |
6813 | long __builtin_alpha_cmpbge (long, long) | |
c4b50f1a RH |
6814 | long __builtin_alpha_extbl (long, long) |
6815 | long __builtin_alpha_extwl (long, long) | |
6816 | long __builtin_alpha_extll (long, long) | |
6d8fd7bb | 6817 | long __builtin_alpha_extql (long, long) |
c4b50f1a RH |
6818 | long __builtin_alpha_extwh (long, long) |
6819 | long __builtin_alpha_extlh (long, long) | |
6d8fd7bb | 6820 | long __builtin_alpha_extqh (long, long) |
c4b50f1a RH |
6821 | long __builtin_alpha_insbl (long, long) |
6822 | long __builtin_alpha_inswl (long, long) | |
6823 | long __builtin_alpha_insll (long, long) | |
6824 | long __builtin_alpha_insql (long, long) | |
6825 | long __builtin_alpha_inswh (long, long) | |
6826 | long __builtin_alpha_inslh (long, long) | |
6827 | long __builtin_alpha_insqh (long, long) | |
6828 | long __builtin_alpha_mskbl (long, long) | |
6829 | long __builtin_alpha_mskwl (long, long) | |
6830 | long __builtin_alpha_mskll (long, long) | |
6831 | long __builtin_alpha_mskql (long, long) | |
6832 | long __builtin_alpha_mskwh (long, long) | |
6833 | long __builtin_alpha_msklh (long, long) | |
6834 | long __builtin_alpha_mskqh (long, long) | |
6835 | long __builtin_alpha_umulh (long, long) | |
6d8fd7bb RH |
6836 | long __builtin_alpha_zap (long, long) |
6837 | long __builtin_alpha_zapnot (long, long) | |
3ab51846 | 6838 | @end smallexample |
6d8fd7bb RH |
6839 | |
6840 | The following built-in functions are always with @option{-mmax} | |
6841 | or @option{-mcpu=@var{cpu}} where @var{cpu} is @code{pca56} or | |
6842 | later. They all generate the machine instruction that is part | |
6843 | of the name. | |
6844 | ||
3ab51846 | 6845 | @smallexample |
6d8fd7bb RH |
6846 | long __builtin_alpha_pklb (long) |
6847 | long __builtin_alpha_pkwb (long) | |
6848 | long __builtin_alpha_unpkbl (long) | |
6849 | long __builtin_alpha_unpkbw (long) | |
6850 | long __builtin_alpha_minub8 (long, long) | |
6851 | long __builtin_alpha_minsb8 (long, long) | |
6852 | long __builtin_alpha_minuw4 (long, long) | |
6853 | long __builtin_alpha_minsw4 (long, long) | |
6854 | long __builtin_alpha_maxub8 (long, long) | |
6855 | long __builtin_alpha_maxsb8 (long, long) | |
6856 | long __builtin_alpha_maxuw4 (long, long) | |
6857 | long __builtin_alpha_maxsw4 (long, long) | |
6858 | long __builtin_alpha_perr (long, long) | |
3ab51846 | 6859 | @end smallexample |
6d8fd7bb | 6860 | |
c4b50f1a RH |
6861 | The following built-in functions are always with @option{-mcix} |
6862 | or @option{-mcpu=@var{cpu}} where @var{cpu} is @code{ev67} or | |
6863 | later. They all generate the machine instruction that is part | |
6864 | of the name. | |
6865 | ||
3ab51846 | 6866 | @smallexample |
c4b50f1a RH |
6867 | long __builtin_alpha_cttz (long) |
6868 | long __builtin_alpha_ctlz (long) | |
6869 | long __builtin_alpha_ctpop (long) | |
3ab51846 | 6870 | @end smallexample |
c4b50f1a | 6871 | |
116b7a5e RH |
6872 | The following builtins are available on systems that use the OSF/1 |
6873 | PALcode. Normally they invoke the @code{rduniq} and @code{wruniq} | |
6874 | PAL calls, but when invoked with @option{-mtls-kernel}, they invoke | |
6875 | @code{rdval} and @code{wrval}. | |
6876 | ||
3ab51846 | 6877 | @smallexample |
116b7a5e RH |
6878 | void *__builtin_thread_pointer (void) |
6879 | void __builtin_set_thread_pointer (void *) | |
3ab51846 | 6880 | @end smallexample |
116b7a5e | 6881 | |
88f77cba JB |
6882 | @node ARM iWMMXt Built-in Functions |
6883 | @subsection ARM iWMMXt Built-in Functions | |
4bc73018 NC |
6884 | |
6885 | These built-in functions are available for the ARM family of | |
88f77cba | 6886 | processors when the @option{-mcpu=iwmmxt} switch is used: |
4bc73018 | 6887 | |
3ab51846 | 6888 | @smallexample |
d63851eb ILT |
6889 | typedef int v2si __attribute__ ((vector_size (8))); |
6890 | typedef short v4hi __attribute__ ((vector_size (8))); | |
6891 | typedef char v8qi __attribute__ ((vector_size (8))); | |
6892 | ||
6893 | int __builtin_arm_getwcx (int) | |
6894 | void __builtin_arm_setwcx (int, int) | |
6895 | int __builtin_arm_textrmsb (v8qi, int) | |
6896 | int __builtin_arm_textrmsh (v4hi, int) | |
6897 | int __builtin_arm_textrmsw (v2si, int) | |
6898 | int __builtin_arm_textrmub (v8qi, int) | |
6899 | int __builtin_arm_textrmuh (v4hi, int) | |
6900 | int __builtin_arm_textrmuw (v2si, int) | |
6901 | v8qi __builtin_arm_tinsrb (v8qi, int) | |
6902 | v4hi __builtin_arm_tinsrh (v4hi, int) | |
6903 | v2si __builtin_arm_tinsrw (v2si, int) | |
6904 | long long __builtin_arm_tmia (long long, int, int) | |
6905 | long long __builtin_arm_tmiabb (long long, int, int) | |
6906 | long long __builtin_arm_tmiabt (long long, int, int) | |
6907 | long long __builtin_arm_tmiaph (long long, int, int) | |
6908 | long long __builtin_arm_tmiatb (long long, int, int) | |
6909 | long long __builtin_arm_tmiatt (long long, int, int) | |
6910 | int __builtin_arm_tmovmskb (v8qi) | |
6911 | int __builtin_arm_tmovmskh (v4hi) | |
6912 | int __builtin_arm_tmovmskw (v2si) | |
6913 | long long __builtin_arm_waccb (v8qi) | |
6914 | long long __builtin_arm_wacch (v4hi) | |
6915 | long long __builtin_arm_waccw (v2si) | |
6916 | v8qi __builtin_arm_waddb (v8qi, v8qi) | |
6917 | v8qi __builtin_arm_waddbss (v8qi, v8qi) | |
6918 | v8qi __builtin_arm_waddbus (v8qi, v8qi) | |
6919 | v4hi __builtin_arm_waddh (v4hi, v4hi) | |
6920 | v4hi __builtin_arm_waddhss (v4hi, v4hi) | |
6921 | v4hi __builtin_arm_waddhus (v4hi, v4hi) | |
4bc73018 | 6922 | v2si __builtin_arm_waddw (v2si, v2si) |
4bc73018 | 6923 | v2si __builtin_arm_waddwss (v2si, v2si) |
4bc73018 | 6924 | v2si __builtin_arm_waddwus (v2si, v2si) |
d63851eb ILT |
6925 | v8qi __builtin_arm_walign (v8qi, v8qi, int) |
6926 | long long __builtin_arm_wand(long long, long long) | |
6927 | long long __builtin_arm_wandn (long long, long long) | |
6928 | v8qi __builtin_arm_wavg2b (v8qi, v8qi) | |
6929 | v8qi __builtin_arm_wavg2br (v8qi, v8qi) | |
6930 | v4hi __builtin_arm_wavg2h (v4hi, v4hi) | |
6931 | v4hi __builtin_arm_wavg2hr (v4hi, v4hi) | |
6932 | v8qi __builtin_arm_wcmpeqb (v8qi, v8qi) | |
6933 | v4hi __builtin_arm_wcmpeqh (v4hi, v4hi) | |
4bc73018 | 6934 | v2si __builtin_arm_wcmpeqw (v2si, v2si) |
d63851eb ILT |
6935 | v8qi __builtin_arm_wcmpgtsb (v8qi, v8qi) |
6936 | v4hi __builtin_arm_wcmpgtsh (v4hi, v4hi) | |
4bc73018 | 6937 | v2si __builtin_arm_wcmpgtsw (v2si, v2si) |
d63851eb ILT |
6938 | v8qi __builtin_arm_wcmpgtub (v8qi, v8qi) |
6939 | v4hi __builtin_arm_wcmpgtuh (v4hi, v4hi) | |
6940 | v2si __builtin_arm_wcmpgtuw (v2si, v2si) | |
6941 | long long __builtin_arm_wmacs (long long, v4hi, v4hi) | |
6942 | long long __builtin_arm_wmacsz (v4hi, v4hi) | |
6943 | long long __builtin_arm_wmacu (long long, v4hi, v4hi) | |
6944 | long long __builtin_arm_wmacuz (v4hi, v4hi) | |
6945 | v4hi __builtin_arm_wmadds (v4hi, v4hi) | |
6946 | v4hi __builtin_arm_wmaddu (v4hi, v4hi) | |
6947 | v8qi __builtin_arm_wmaxsb (v8qi, v8qi) | |
6948 | v4hi __builtin_arm_wmaxsh (v4hi, v4hi) | |
4bc73018 | 6949 | v2si __builtin_arm_wmaxsw (v2si, v2si) |
d63851eb ILT |
6950 | v8qi __builtin_arm_wmaxub (v8qi, v8qi) |
6951 | v4hi __builtin_arm_wmaxuh (v4hi, v4hi) | |
4bc73018 | 6952 | v2si __builtin_arm_wmaxuw (v2si, v2si) |
d63851eb ILT |
6953 | v8qi __builtin_arm_wminsb (v8qi, v8qi) |
6954 | v4hi __builtin_arm_wminsh (v4hi, v4hi) | |
4bc73018 | 6955 | v2si __builtin_arm_wminsw (v2si, v2si) |
d63851eb ILT |
6956 | v8qi __builtin_arm_wminub (v8qi, v8qi) |
6957 | v4hi __builtin_arm_wminuh (v4hi, v4hi) | |
4bc73018 | 6958 | v2si __builtin_arm_wminuw (v2si, v2si) |
d63851eb ILT |
6959 | v4hi __builtin_arm_wmulsm (v4hi, v4hi) |
6960 | v4hi __builtin_arm_wmulul (v4hi, v4hi) | |
6961 | v4hi __builtin_arm_wmulum (v4hi, v4hi) | |
6962 | long long __builtin_arm_wor (long long, long long) | |
6963 | v2si __builtin_arm_wpackdss (long long, long long) | |
6964 | v2si __builtin_arm_wpackdus (long long, long long) | |
6965 | v8qi __builtin_arm_wpackhss (v4hi, v4hi) | |
6966 | v8qi __builtin_arm_wpackhus (v4hi, v4hi) | |
6967 | v4hi __builtin_arm_wpackwss (v2si, v2si) | |
6968 | v4hi __builtin_arm_wpackwus (v2si, v2si) | |
6969 | long long __builtin_arm_wrord (long long, long long) | |
6970 | long long __builtin_arm_wrordi (long long, int) | |
6971 | v4hi __builtin_arm_wrorh (v4hi, long long) | |
6972 | v4hi __builtin_arm_wrorhi (v4hi, int) | |
6973 | v2si __builtin_arm_wrorw (v2si, long long) | |
6974 | v2si __builtin_arm_wrorwi (v2si, int) | |
6975 | v2si __builtin_arm_wsadb (v8qi, v8qi) | |
6976 | v2si __builtin_arm_wsadbz (v8qi, v8qi) | |
6977 | v2si __builtin_arm_wsadh (v4hi, v4hi) | |
6978 | v2si __builtin_arm_wsadhz (v4hi, v4hi) | |
6979 | v4hi __builtin_arm_wshufh (v4hi, int) | |
6980 | long long __builtin_arm_wslld (long long, long long) | |
6981 | long long __builtin_arm_wslldi (long long, int) | |
6982 | v4hi __builtin_arm_wsllh (v4hi, long long) | |
6983 | v4hi __builtin_arm_wsllhi (v4hi, int) | |
6984 | v2si __builtin_arm_wsllw (v2si, long long) | |
4bc73018 | 6985 | v2si __builtin_arm_wsllwi (v2si, int) |
d63851eb ILT |
6986 | long long __builtin_arm_wsrad (long long, long long) |
6987 | long long __builtin_arm_wsradi (long long, int) | |
6988 | v4hi __builtin_arm_wsrah (v4hi, long long) | |
6989 | v4hi __builtin_arm_wsrahi (v4hi, int) | |
6990 | v2si __builtin_arm_wsraw (v2si, long long) | |
4bc73018 | 6991 | v2si __builtin_arm_wsrawi (v2si, int) |
d63851eb ILT |
6992 | long long __builtin_arm_wsrld (long long, long long) |
6993 | long long __builtin_arm_wsrldi (long long, int) | |
6994 | v4hi __builtin_arm_wsrlh (v4hi, long long) | |
6995 | v4hi __builtin_arm_wsrlhi (v4hi, int) | |
6996 | v2si __builtin_arm_wsrlw (v2si, long long) | |
4bc73018 | 6997 | v2si __builtin_arm_wsrlwi (v2si, int) |
d63851eb ILT |
6998 | v8qi __builtin_arm_wsubb (v8qi, v8qi) |
6999 | v8qi __builtin_arm_wsubbss (v8qi, v8qi) | |
7000 | v8qi __builtin_arm_wsubbus (v8qi, v8qi) | |
7001 | v4hi __builtin_arm_wsubh (v4hi, v4hi) | |
7002 | v4hi __builtin_arm_wsubhss (v4hi, v4hi) | |
7003 | v4hi __builtin_arm_wsubhus (v4hi, v4hi) | |
7004 | v2si __builtin_arm_wsubw (v2si, v2si) | |
7005 | v2si __builtin_arm_wsubwss (v2si, v2si) | |
7006 | v2si __builtin_arm_wsubwus (v2si, v2si) | |
7007 | v4hi __builtin_arm_wunpckehsb (v8qi) | |
7008 | v2si __builtin_arm_wunpckehsh (v4hi) | |
7009 | long long __builtin_arm_wunpckehsw (v2si) | |
7010 | v4hi __builtin_arm_wunpckehub (v8qi) | |
7011 | v2si __builtin_arm_wunpckehuh (v4hi) | |
7012 | long long __builtin_arm_wunpckehuw (v2si) | |
7013 | v4hi __builtin_arm_wunpckelsb (v8qi) | |
7014 | v2si __builtin_arm_wunpckelsh (v4hi) | |
7015 | long long __builtin_arm_wunpckelsw (v2si) | |
7016 | v4hi __builtin_arm_wunpckelub (v8qi) | |
7017 | v2si __builtin_arm_wunpckeluh (v4hi) | |
7018 | long long __builtin_arm_wunpckeluw (v2si) | |
7019 | v8qi __builtin_arm_wunpckihb (v8qi, v8qi) | |
7020 | v4hi __builtin_arm_wunpckihh (v4hi, v4hi) | |
4bc73018 | 7021 | v2si __builtin_arm_wunpckihw (v2si, v2si) |
d63851eb ILT |
7022 | v8qi __builtin_arm_wunpckilb (v8qi, v8qi) |
7023 | v4hi __builtin_arm_wunpckilh (v4hi, v4hi) | |
4bc73018 | 7024 | v2si __builtin_arm_wunpckilw (v2si, v2si) |
d63851eb ILT |
7025 | long long __builtin_arm_wxor (long long, long long) |
7026 | long long __builtin_arm_wzero () | |
3ab51846 | 7027 | @end smallexample |
4bc73018 | 7028 | |
88f77cba JB |
7029 | @node ARM NEON Intrinsics |
7030 | @subsection ARM NEON Intrinsics | |
7031 | ||
7032 | These built-in intrinsics for the ARM Advanced SIMD extension are available | |
7033 | when the @option{-mfpu=neon} switch is used: | |
7034 | ||
7035 | @include arm-neon-intrinsics.texi | |
7036 | ||
161c21b6 BS |
7037 | @node Blackfin Built-in Functions |
7038 | @subsection Blackfin Built-in Functions | |
7039 | ||
7040 | Currently, there are two Blackfin-specific built-in functions. These are | |
7041 | used for generating @code{CSYNC} and @code{SSYNC} machine insns without | |
7042 | using inline assembly; by using these built-in functions the compiler can | |
7043 | automatically add workarounds for hardware errata involving these | |
7044 | instructions. These functions are named as follows: | |
7045 | ||
7046 | @smallexample | |
7047 | void __builtin_bfin_csync (void) | |
7048 | void __builtin_bfin_ssync (void) | |
7049 | @end smallexample | |
7050 | ||
c3ee0579 RS |
7051 | @node FR-V Built-in Functions |
7052 | @subsection FR-V Built-in Functions | |
7053 | ||
7054 | GCC provides many FR-V-specific built-in functions. In general, | |
7055 | these functions are intended to be compatible with those described | |
7056 | by @cite{FR-V Family, Softune C/C++ Compiler Manual (V6), Fujitsu | |
7057 | Semiconductor}. The two exceptions are @code{__MDUNPACKH} and | |
7058 | @code{__MBTOHE}, the gcc forms of which pass 128-bit values by | |
7059 | pointer rather than by value. | |
7060 | ||
7061 | Most of the functions are named after specific FR-V instructions. | |
27ef2cdd | 7062 | Such functions are said to be ``directly mapped'' and are summarized |
c3ee0579 RS |
7063 | here in tabular form. |
7064 | ||
7065 | @menu | |
7066 | * Argument Types:: | |
7067 | * Directly-mapped Integer Functions:: | |
7068 | * Directly-mapped Media Functions:: | |
c14ff86e | 7069 | * Raw read/write Functions:: |
c3ee0579 RS |
7070 | * Other Built-in Functions:: |
7071 | @end menu | |
7072 | ||
7073 | @node Argument Types | |
7074 | @subsubsection Argument Types | |
7075 | ||
7076 | The arguments to the built-in functions can be divided into three groups: | |
7077 | register numbers, compile-time constants and run-time values. In order | |
7078 | to make this classification clear at a glance, the arguments and return | |
7079 | values are given the following pseudo types: | |
7080 | ||
7081 | @multitable @columnfractions .20 .30 .15 .35 | |
7082 | @item Pseudo type @tab Real C type @tab Constant? @tab Description | |
7083 | @item @code{uh} @tab @code{unsigned short} @tab No @tab an unsigned halfword | |
7084 | @item @code{uw1} @tab @code{unsigned int} @tab No @tab an unsigned word | |
7085 | @item @code{sw1} @tab @code{int} @tab No @tab a signed word | |
7086 | @item @code{uw2} @tab @code{unsigned long long} @tab No | |
7087 | @tab an unsigned doubleword | |
7088 | @item @code{sw2} @tab @code{long long} @tab No @tab a signed doubleword | |
7089 | @item @code{const} @tab @code{int} @tab Yes @tab an integer constant | |
7090 | @item @code{acc} @tab @code{int} @tab Yes @tab an ACC register number | |
7091 | @item @code{iacc} @tab @code{int} @tab Yes @tab an IACC register number | |
7092 | @end multitable | |
7093 | ||
7094 | These pseudo types are not defined by GCC, they are simply a notational | |
7095 | convenience used in this manual. | |
7096 | ||
7097 | Arguments of type @code{uh}, @code{uw1}, @code{sw1}, @code{uw2} | |
7098 | and @code{sw2} are evaluated at run time. They correspond to | |
7099 | register operands in the underlying FR-V instructions. | |
7100 | ||
7101 | @code{const} arguments represent immediate operands in the underlying | |
7102 | FR-V instructions. They must be compile-time constants. | |
7103 | ||
7104 | @code{acc} arguments are evaluated at compile time and specify the number | |
7105 | of an accumulator register. For example, an @code{acc} argument of 2 | |
7106 | will select the ACC2 register. | |
7107 | ||
7108 | @code{iacc} arguments are similar to @code{acc} arguments but specify the | |
7109 | number of an IACC register. See @pxref{Other Built-in Functions} | |
7110 | for more details. | |
7111 | ||
7112 | @node Directly-mapped Integer Functions | |
7113 | @subsubsection Directly-mapped Integer Functions | |
7114 | ||
7115 | The functions listed below map directly to FR-V I-type instructions. | |
7116 | ||
7117 | @multitable @columnfractions .45 .32 .23 | |
7118 | @item Function prototype @tab Example usage @tab Assembly output | |
7119 | @item @code{sw1 __ADDSS (sw1, sw1)} | |
7120 | @tab @code{@var{c} = __ADDSS (@var{a}, @var{b})} | |
7121 | @tab @code{ADDSS @var{a},@var{b},@var{c}} | |
7122 | @item @code{sw1 __SCAN (sw1, sw1)} | |
7123 | @tab @code{@var{c} = __SCAN (@var{a}, @var{b})} | |
7124 | @tab @code{SCAN @var{a},@var{b},@var{c}} | |
7125 | @item @code{sw1 __SCUTSS (sw1)} | |
7126 | @tab @code{@var{b} = __SCUTSS (@var{a})} | |
7127 | @tab @code{SCUTSS @var{a},@var{b}} | |
7128 | @item @code{sw1 __SLASS (sw1, sw1)} | |
7129 | @tab @code{@var{c} = __SLASS (@var{a}, @var{b})} | |
7130 | @tab @code{SLASS @var{a},@var{b},@var{c}} | |
7131 | @item @code{void __SMASS (sw1, sw1)} | |
7132 | @tab @code{__SMASS (@var{a}, @var{b})} | |
7133 | @tab @code{SMASS @var{a},@var{b}} | |
7134 | @item @code{void __SMSSS (sw1, sw1)} | |
7135 | @tab @code{__SMSSS (@var{a}, @var{b})} | |
7136 | @tab @code{SMSSS @var{a},@var{b}} | |
7137 | @item @code{void __SMU (sw1, sw1)} | |
7138 | @tab @code{__SMU (@var{a}, @var{b})} | |
7139 | @tab @code{SMU @var{a},@var{b}} | |
7140 | @item @code{sw2 __SMUL (sw1, sw1)} | |
7141 | @tab @code{@var{c} = __SMUL (@var{a}, @var{b})} | |
7142 | @tab @code{SMUL @var{a},@var{b},@var{c}} | |
7143 | @item @code{sw1 __SUBSS (sw1, sw1)} | |
7144 | @tab @code{@var{c} = __SUBSS (@var{a}, @var{b})} | |
7145 | @tab @code{SUBSS @var{a},@var{b},@var{c}} | |
7146 | @item @code{uw2 __UMUL (uw1, uw1)} | |
7147 | @tab @code{@var{c} = __UMUL (@var{a}, @var{b})} | |
7148 | @tab @code{UMUL @var{a},@var{b},@var{c}} | |
7149 | @end multitable | |
7150 | ||
7151 | @node Directly-mapped Media Functions | |
7152 | @subsubsection Directly-mapped Media Functions | |
7153 | ||
7154 | The functions listed below map directly to FR-V M-type instructions. | |
7155 | ||
7156 | @multitable @columnfractions .45 .32 .23 | |
7157 | @item Function prototype @tab Example usage @tab Assembly output | |
7158 | @item @code{uw1 __MABSHS (sw1)} | |
7159 | @tab @code{@var{b} = __MABSHS (@var{a})} | |
7160 | @tab @code{MABSHS @var{a},@var{b}} | |
7161 | @item @code{void __MADDACCS (acc, acc)} | |
7162 | @tab @code{__MADDACCS (@var{b}, @var{a})} | |
7163 | @tab @code{MADDACCS @var{a},@var{b}} | |
7164 | @item @code{sw1 __MADDHSS (sw1, sw1)} | |
7165 | @tab @code{@var{c} = __MADDHSS (@var{a}, @var{b})} | |
7166 | @tab @code{MADDHSS @var{a},@var{b},@var{c}} | |
7167 | @item @code{uw1 __MADDHUS (uw1, uw1)} | |
7168 | @tab @code{@var{c} = __MADDHUS (@var{a}, @var{b})} | |
7169 | @tab @code{MADDHUS @var{a},@var{b},@var{c}} | |
7170 | @item @code{uw1 __MAND (uw1, uw1)} | |
7171 | @tab @code{@var{c} = __MAND (@var{a}, @var{b})} | |
7172 | @tab @code{MAND @var{a},@var{b},@var{c}} | |
7173 | @item @code{void __MASACCS (acc, acc)} | |
7174 | @tab @code{__MASACCS (@var{b}, @var{a})} | |
7175 | @tab @code{MASACCS @var{a},@var{b}} | |
7176 | @item @code{uw1 __MAVEH (uw1, uw1)} | |
7177 | @tab @code{@var{c} = __MAVEH (@var{a}, @var{b})} | |
7178 | @tab @code{MAVEH @var{a},@var{b},@var{c}} | |
7179 | @item @code{uw2 __MBTOH (uw1)} | |
7180 | @tab @code{@var{b} = __MBTOH (@var{a})} | |
7181 | @tab @code{MBTOH @var{a},@var{b}} | |
7182 | @item @code{void __MBTOHE (uw1 *, uw1)} | |
7183 | @tab @code{__MBTOHE (&@var{b}, @var{a})} | |
7184 | @tab @code{MBTOHE @var{a},@var{b}} | |
7185 | @item @code{void __MCLRACC (acc)} | |
7186 | @tab @code{__MCLRACC (@var{a})} | |
7187 | @tab @code{MCLRACC @var{a}} | |
7188 | @item @code{void __MCLRACCA (void)} | |
7189 | @tab @code{__MCLRACCA ()} | |
7190 | @tab @code{MCLRACCA} | |
7191 | @item @code{uw1 __Mcop1 (uw1, uw1)} | |
7192 | @tab @code{@var{c} = __Mcop1 (@var{a}, @var{b})} | |
7193 | @tab @code{Mcop1 @var{a},@var{b},@var{c}} | |
7194 | @item @code{uw1 __Mcop2 (uw1, uw1)} | |
7195 | @tab @code{@var{c} = __Mcop2 (@var{a}, @var{b})} | |
7196 | @tab @code{Mcop2 @var{a},@var{b},@var{c}} | |
7197 | @item @code{uw1 __MCPLHI (uw2, const)} | |
7198 | @tab @code{@var{c} = __MCPLHI (@var{a}, @var{b})} | |
7199 | @tab @code{MCPLHI @var{a},#@var{b},@var{c}} | |
7200 | @item @code{uw1 __MCPLI (uw2, const)} | |
7201 | @tab @code{@var{c} = __MCPLI (@var{a}, @var{b})} | |
7202 | @tab @code{MCPLI @var{a},#@var{b},@var{c}} | |
7203 | @item @code{void __MCPXIS (acc, sw1, sw1)} | |
7204 | @tab @code{__MCPXIS (@var{c}, @var{a}, @var{b})} | |
7205 | @tab @code{MCPXIS @var{a},@var{b},@var{c}} | |
7206 | @item @code{void __MCPXIU (acc, uw1, uw1)} | |
7207 | @tab @code{__MCPXIU (@var{c}, @var{a}, @var{b})} | |
7208 | @tab @code{MCPXIU @var{a},@var{b},@var{c}} | |
7209 | @item @code{void __MCPXRS (acc, sw1, sw1)} | |
7210 | @tab @code{__MCPXRS (@var{c}, @var{a}, @var{b})} | |
7211 | @tab @code{MCPXRS @var{a},@var{b},@var{c}} | |
7212 | @item @code{void __MCPXRU (acc, uw1, uw1)} | |
7213 | @tab @code{__MCPXRU (@var{c}, @var{a}, @var{b})} | |
7214 | @tab @code{MCPXRU @var{a},@var{b},@var{c}} | |
7215 | @item @code{uw1 __MCUT (acc, uw1)} | |
7216 | @tab @code{@var{c} = __MCUT (@var{a}, @var{b})} | |
7217 | @tab @code{MCUT @var{a},@var{b},@var{c}} | |
7218 | @item @code{uw1 __MCUTSS (acc, sw1)} | |
7219 | @tab @code{@var{c} = __MCUTSS (@var{a}, @var{b})} | |
7220 | @tab @code{MCUTSS @var{a},@var{b},@var{c}} | |
7221 | @item @code{void __MDADDACCS (acc, acc)} | |
7222 | @tab @code{__MDADDACCS (@var{b}, @var{a})} | |
7223 | @tab @code{MDADDACCS @var{a},@var{b}} | |
7224 | @item @code{void __MDASACCS (acc, acc)} | |
7225 | @tab @code{__MDASACCS (@var{b}, @var{a})} | |
7226 | @tab @code{MDASACCS @var{a},@var{b}} | |
7227 | @item @code{uw2 __MDCUTSSI (acc, const)} | |
7228 | @tab @code{@var{c} = __MDCUTSSI (@var{a}, @var{b})} | |
7229 | @tab @code{MDCUTSSI @var{a},#@var{b},@var{c}} | |
7230 | @item @code{uw2 __MDPACKH (uw2, uw2)} | |
7231 | @tab @code{@var{c} = __MDPACKH (@var{a}, @var{b})} | |
7232 | @tab @code{MDPACKH @var{a},@var{b},@var{c}} | |
7233 | @item @code{uw2 __MDROTLI (uw2, const)} | |
7234 | @tab @code{@var{c} = __MDROTLI (@var{a}, @var{b})} | |
7235 | @tab @code{MDROTLI @var{a},#@var{b},@var{c}} | |
7236 | @item @code{void __MDSUBACCS (acc, acc)} | |
7237 | @tab @code{__MDSUBACCS (@var{b}, @var{a})} | |
7238 | @tab @code{MDSUBACCS @var{a},@var{b}} | |
7239 | @item @code{void __MDUNPACKH (uw1 *, uw2)} | |
7240 | @tab @code{__MDUNPACKH (&@var{b}, @var{a})} | |
7241 | @tab @code{MDUNPACKH @var{a},@var{b}} | |
7242 | @item @code{uw2 __MEXPDHD (uw1, const)} | |
7243 | @tab @code{@var{c} = __MEXPDHD (@var{a}, @var{b})} | |
7244 | @tab @code{MEXPDHD @var{a},#@var{b},@var{c}} | |
7245 | @item @code{uw1 __MEXPDHW (uw1, const)} | |
7246 | @tab @code{@var{c} = __MEXPDHW (@var{a}, @var{b})} | |
7247 | @tab @code{MEXPDHW @var{a},#@var{b},@var{c}} | |
7248 | @item @code{uw1 __MHDSETH (uw1, const)} | |
7249 | @tab @code{@var{c} = __MHDSETH (@var{a}, @var{b})} | |
7250 | @tab @code{MHDSETH @var{a},#@var{b},@var{c}} | |
7251 | @item @code{sw1 __MHDSETS (const)} | |
7252 | @tab @code{@var{b} = __MHDSETS (@var{a})} | |
7253 | @tab @code{MHDSETS #@var{a},@var{b}} | |
7254 | @item @code{uw1 __MHSETHIH (uw1, const)} | |
7255 | @tab @code{@var{b} = __MHSETHIH (@var{b}, @var{a})} | |
7256 | @tab @code{MHSETHIH #@var{a},@var{b}} | |
7257 | @item @code{sw1 __MHSETHIS (sw1, const)} | |
7258 | @tab @code{@var{b} = __MHSETHIS (@var{b}, @var{a})} | |
7259 | @tab @code{MHSETHIS #@var{a},@var{b}} | |
7260 | @item @code{uw1 __MHSETLOH (uw1, const)} | |
7261 | @tab @code{@var{b} = __MHSETLOH (@var{b}, @var{a})} | |
7262 | @tab @code{MHSETLOH #@var{a},@var{b}} | |
7263 | @item @code{sw1 __MHSETLOS (sw1, const)} | |
7264 | @tab @code{@var{b} = __MHSETLOS (@var{b}, @var{a})} | |
7265 | @tab @code{MHSETLOS #@var{a},@var{b}} | |
7266 | @item @code{uw1 __MHTOB (uw2)} | |
7267 | @tab @code{@var{b} = __MHTOB (@var{a})} | |
7268 | @tab @code{MHTOB @var{a},@var{b}} | |
7269 | @item @code{void __MMACHS (acc, sw1, sw1)} | |
7270 | @tab @code{__MMACHS (@var{c}, @var{a}, @var{b})} | |
7271 | @tab @code{MMACHS @var{a},@var{b},@var{c}} | |
7272 | @item @code{void __MMACHU (acc, uw1, uw1)} | |
7273 | @tab @code{__MMACHU (@var{c}, @var{a}, @var{b})} | |
7274 | @tab @code{MMACHU @var{a},@var{b},@var{c}} | |
7275 | @item @code{void __MMRDHS (acc, sw1, sw1)} | |
7276 | @tab @code{__MMRDHS (@var{c}, @var{a}, @var{b})} | |
7277 | @tab @code{MMRDHS @var{a},@var{b},@var{c}} | |
7278 | @item @code{void __MMRDHU (acc, uw1, uw1)} | |
7279 | @tab @code{__MMRDHU (@var{c}, @var{a}, @var{b})} | |
7280 | @tab @code{MMRDHU @var{a},@var{b},@var{c}} | |
7281 | @item @code{void __MMULHS (acc, sw1, sw1)} | |
7282 | @tab @code{__MMULHS (@var{c}, @var{a}, @var{b})} | |
7283 | @tab @code{MMULHS @var{a},@var{b},@var{c}} | |
7284 | @item @code{void __MMULHU (acc, uw1, uw1)} | |
7285 | @tab @code{__MMULHU (@var{c}, @var{a}, @var{b})} | |
7286 | @tab @code{MMULHU @var{a},@var{b},@var{c}} | |
7287 | @item @code{void __MMULXHS (acc, sw1, sw1)} | |
7288 | @tab @code{__MMULXHS (@var{c}, @var{a}, @var{b})} | |
7289 | @tab @code{MMULXHS @var{a},@var{b},@var{c}} | |
7290 | @item @code{void __MMULXHU (acc, uw1, uw1)} | |
7291 | @tab @code{__MMULXHU (@var{c}, @var{a}, @var{b})} | |
7292 | @tab @code{MMULXHU @var{a},@var{b},@var{c}} | |
7293 | @item @code{uw1 __MNOT (uw1)} | |
7294 | @tab @code{@var{b} = __MNOT (@var{a})} | |
7295 | @tab @code{MNOT @var{a},@var{b}} | |
7296 | @item @code{uw1 __MOR (uw1, uw1)} | |
7297 | @tab @code{@var{c} = __MOR (@var{a}, @var{b})} | |
7298 | @tab @code{MOR @var{a},@var{b},@var{c}} | |
7299 | @item @code{uw1 __MPACKH (uh, uh)} | |
7300 | @tab @code{@var{c} = __MPACKH (@var{a}, @var{b})} | |
7301 | @tab @code{MPACKH @var{a},@var{b},@var{c}} | |
7302 | @item @code{sw2 __MQADDHSS (sw2, sw2)} | |
7303 | @tab @code{@var{c} = __MQADDHSS (@var{a}, @var{b})} | |
7304 | @tab @code{MQADDHSS @var{a},@var{b},@var{c}} | |
7305 | @item @code{uw2 __MQADDHUS (uw2, uw2)} | |
7306 | @tab @code{@var{c} = __MQADDHUS (@var{a}, @var{b})} | |
7307 | @tab @code{MQADDHUS @var{a},@var{b},@var{c}} | |
7308 | @item @code{void __MQCPXIS (acc, sw2, sw2)} | |
7309 | @tab @code{__MQCPXIS (@var{c}, @var{a}, @var{b})} | |
7310 | @tab @code{MQCPXIS @var{a},@var{b},@var{c}} | |
7311 | @item @code{void __MQCPXIU (acc, uw2, uw2)} | |
7312 | @tab @code{__MQCPXIU (@var{c}, @var{a}, @var{b})} | |
7313 | @tab @code{MQCPXIU @var{a},@var{b},@var{c}} | |
7314 | @item @code{void __MQCPXRS (acc, sw2, sw2)} | |
7315 | @tab @code{__MQCPXRS (@var{c}, @var{a}, @var{b})} | |
7316 | @tab @code{MQCPXRS @var{a},@var{b},@var{c}} | |
7317 | @item @code{void __MQCPXRU (acc, uw2, uw2)} | |
7318 | @tab @code{__MQCPXRU (@var{c}, @var{a}, @var{b})} | |
7319 | @tab @code{MQCPXRU @var{a},@var{b},@var{c}} | |
7320 | @item @code{sw2 __MQLCLRHS (sw2, sw2)} | |
7321 | @tab @code{@var{c} = __MQLCLRHS (@var{a}, @var{b})} | |
7322 | @tab @code{MQLCLRHS @var{a},@var{b},@var{c}} | |
7323 | @item @code{sw2 __MQLMTHS (sw2, sw2)} | |
7324 | @tab @code{@var{c} = __MQLMTHS (@var{a}, @var{b})} | |
7325 | @tab @code{MQLMTHS @var{a},@var{b},@var{c}} | |
7326 | @item @code{void __MQMACHS (acc, sw2, sw2)} | |
7327 | @tab @code{__MQMACHS (@var{c}, @var{a}, @var{b})} | |
7328 | @tab @code{MQMACHS @var{a},@var{b},@var{c}} | |
7329 | @item @code{void __MQMACHU (acc, uw2, uw2)} | |
7330 | @tab @code{__MQMACHU (@var{c}, @var{a}, @var{b})} | |
7331 | @tab @code{MQMACHU @var{a},@var{b},@var{c}} | |
7332 | @item @code{void __MQMACXHS (acc, sw2, sw2)} | |
7333 | @tab @code{__MQMACXHS (@var{c}, @var{a}, @var{b})} | |
7334 | @tab @code{MQMACXHS @var{a},@var{b},@var{c}} | |
7335 | @item @code{void __MQMULHS (acc, sw2, sw2)} | |
7336 | @tab @code{__MQMULHS (@var{c}, @var{a}, @var{b})} | |
7337 | @tab @code{MQMULHS @var{a},@var{b},@var{c}} | |
7338 | @item @code{void __MQMULHU (acc, uw2, uw2)} | |
7339 | @tab @code{__MQMULHU (@var{c}, @var{a}, @var{b})} | |
7340 | @tab @code{MQMULHU @var{a},@var{b},@var{c}} | |
7341 | @item @code{void __MQMULXHS (acc, sw2, sw2)} | |
7342 | @tab @code{__MQMULXHS (@var{c}, @var{a}, @var{b})} | |
7343 | @tab @code{MQMULXHS @var{a},@var{b},@var{c}} | |
7344 | @item @code{void __MQMULXHU (acc, uw2, uw2)} | |
7345 | @tab @code{__MQMULXHU (@var{c}, @var{a}, @var{b})} | |
7346 | @tab @code{MQMULXHU @var{a},@var{b},@var{c}} | |
7347 | @item @code{sw2 __MQSATHS (sw2, sw2)} | |
7348 | @tab @code{@var{c} = __MQSATHS (@var{a}, @var{b})} | |
7349 | @tab @code{MQSATHS @var{a},@var{b},@var{c}} | |
7350 | @item @code{uw2 __MQSLLHI (uw2, int)} | |
7351 | @tab @code{@var{c} = __MQSLLHI (@var{a}, @var{b})} | |
7352 | @tab @code{MQSLLHI @var{a},@var{b},@var{c}} | |
7353 | @item @code{sw2 __MQSRAHI (sw2, int)} | |
7354 | @tab @code{@var{c} = __MQSRAHI (@var{a}, @var{b})} | |
7355 | @tab @code{MQSRAHI @var{a},@var{b},@var{c}} | |
7356 | @item @code{sw2 __MQSUBHSS (sw2, sw2)} | |
7357 | @tab @code{@var{c} = __MQSUBHSS (@var{a}, @var{b})} | |
7358 | @tab @code{MQSUBHSS @var{a},@var{b},@var{c}} | |
7359 | @item @code{uw2 __MQSUBHUS (uw2, uw2)} | |
7360 | @tab @code{@var{c} = __MQSUBHUS (@var{a}, @var{b})} | |
7361 | @tab @code{MQSUBHUS @var{a},@var{b},@var{c}} | |
7362 | @item @code{void __MQXMACHS (acc, sw2, sw2)} | |
7363 | @tab @code{__MQXMACHS (@var{c}, @var{a}, @var{b})} | |
7364 | @tab @code{MQXMACHS @var{a},@var{b},@var{c}} | |
7365 | @item @code{void __MQXMACXHS (acc, sw2, sw2)} | |
7366 | @tab @code{__MQXMACXHS (@var{c}, @var{a}, @var{b})} | |
7367 | @tab @code{MQXMACXHS @var{a},@var{b},@var{c}} | |
7368 | @item @code{uw1 __MRDACC (acc)} | |
7369 | @tab @code{@var{b} = __MRDACC (@var{a})} | |
7370 | @tab @code{MRDACC @var{a},@var{b}} | |
7371 | @item @code{uw1 __MRDACCG (acc)} | |
7372 | @tab @code{@var{b} = __MRDACCG (@var{a})} | |
7373 | @tab @code{MRDACCG @var{a},@var{b}} | |
7374 | @item @code{uw1 __MROTLI (uw1, const)} | |
7375 | @tab @code{@var{c} = __MROTLI (@var{a}, @var{b})} | |
7376 | @tab @code{MROTLI @var{a},#@var{b},@var{c}} | |
7377 | @item @code{uw1 __MROTRI (uw1, const)} | |
7378 | @tab @code{@var{c} = __MROTRI (@var{a}, @var{b})} | |
7379 | @tab @code{MROTRI @var{a},#@var{b},@var{c}} | |
7380 | @item @code{sw1 __MSATHS (sw1, sw1)} | |
7381 | @tab @code{@var{c} = __MSATHS (@var{a}, @var{b})} | |
7382 | @tab @code{MSATHS @var{a},@var{b},@var{c}} | |
7383 | @item @code{uw1 __MSATHU (uw1, uw1)} | |
7384 | @tab @code{@var{c} = __MSATHU (@var{a}, @var{b})} | |
7385 | @tab @code{MSATHU @var{a},@var{b},@var{c}} | |
7386 | @item @code{uw1 __MSLLHI (uw1, const)} | |
7387 | @tab @code{@var{c} = __MSLLHI (@var{a}, @var{b})} | |
7388 | @tab @code{MSLLHI @var{a},#@var{b},@var{c}} | |
7389 | @item @code{sw1 __MSRAHI (sw1, const)} | |
7390 | @tab @code{@var{c} = __MSRAHI (@var{a}, @var{b})} | |
7391 | @tab @code{MSRAHI @var{a},#@var{b},@var{c}} | |
7392 | @item @code{uw1 __MSRLHI (uw1, const)} | |
7393 | @tab @code{@var{c} = __MSRLHI (@var{a}, @var{b})} | |
7394 | @tab @code{MSRLHI @var{a},#@var{b},@var{c}} | |
7395 | @item @code{void __MSUBACCS (acc, acc)} | |
7396 | @tab @code{__MSUBACCS (@var{b}, @var{a})} | |
7397 | @tab @code{MSUBACCS @var{a},@var{b}} | |
7398 | @item @code{sw1 __MSUBHSS (sw1, sw1)} | |
7399 | @tab @code{@var{c} = __MSUBHSS (@var{a}, @var{b})} | |
7400 | @tab @code{MSUBHSS @var{a},@var{b},@var{c}} | |
7401 | @item @code{uw1 __MSUBHUS (uw1, uw1)} | |
7402 | @tab @code{@var{c} = __MSUBHUS (@var{a}, @var{b})} | |
7403 | @tab @code{MSUBHUS @var{a},@var{b},@var{c}} | |
7404 | @item @code{void __MTRAP (void)} | |
7405 | @tab @code{__MTRAP ()} | |
7406 | @tab @code{MTRAP} | |
7407 | @item @code{uw2 __MUNPACKH (uw1)} | |
7408 | @tab @code{@var{b} = __MUNPACKH (@var{a})} | |
7409 | @tab @code{MUNPACKH @var{a},@var{b}} | |
7410 | @item @code{uw1 __MWCUT (uw2, uw1)} | |
7411 | @tab @code{@var{c} = __MWCUT (@var{a}, @var{b})} | |
7412 | @tab @code{MWCUT @var{a},@var{b},@var{c}} | |
7413 | @item @code{void __MWTACC (acc, uw1)} | |
7414 | @tab @code{__MWTACC (@var{b}, @var{a})} | |
7415 | @tab @code{MWTACC @var{a},@var{b}} | |
7416 | @item @code{void __MWTACCG (acc, uw1)} | |
7417 | @tab @code{__MWTACCG (@var{b}, @var{a})} | |
7418 | @tab @code{MWTACCG @var{a},@var{b}} | |
7419 | @item @code{uw1 __MXOR (uw1, uw1)} | |
7420 | @tab @code{@var{c} = __MXOR (@var{a}, @var{b})} | |
7421 | @tab @code{MXOR @var{a},@var{b},@var{c}} | |
7422 | @end multitable | |
7423 | ||
c14ff86e AH |
7424 | @node Raw read/write Functions |
7425 | @subsubsection Raw read/write Functions | |
7426 | ||
7427 | This sections describes built-in functions related to read and write | |
7428 | instructions to access memory. These functions generate | |
7429 | @code{membar} instructions to flush the I/O load and stores where | |
7430 | appropriate, as described in Fujitsu's manual described above. | |
7431 | ||
7432 | @table @code | |
7433 | ||
7434 | @item unsigned char __builtin_read8 (void *@var{data}) | |
7435 | @item unsigned short __builtin_read16 (void *@var{data}) | |
7436 | @item unsigned long __builtin_read32 (void *@var{data}) | |
7437 | @item unsigned long long __builtin_read64 (void *@var{data}) | |
7438 | ||
7439 | @item void __builtin_write8 (void *@var{data}, unsigned char @var{datum}) | |
7440 | @item void __builtin_write16 (void *@var{data}, unsigned short @var{datum}) | |
7441 | @item void __builtin_write32 (void *@var{data}, unsigned long @var{datum}) | |
7442 | @item void __builtin_write64 (void *@var{data}, unsigned long long @var{datum}) | |
7443 | @end table | |
7444 | ||
c3ee0579 RS |
7445 | @node Other Built-in Functions |
7446 | @subsubsection Other Built-in Functions | |
7447 | ||
7448 | This section describes built-in functions that are not named after | |
7449 | a specific FR-V instruction. | |
7450 | ||
7451 | @table @code | |
7452 | @item sw2 __IACCreadll (iacc @var{reg}) | |
7453 | Return the full 64-bit value of IACC0@. The @var{reg} argument is reserved | |
7454 | for future expansion and must be 0. | |
7455 | ||
7456 | @item sw1 __IACCreadl (iacc @var{reg}) | |
7457 | Return the value of IACC0H if @var{reg} is 0 and IACC0L if @var{reg} is 1. | |
7458 | Other values of @var{reg} are rejected as invalid. | |
7459 | ||
7460 | @item void __IACCsetll (iacc @var{reg}, sw2 @var{x}) | |
7461 | Set the full 64-bit value of IACC0 to @var{x}. The @var{reg} argument | |
7462 | is reserved for future expansion and must be 0. | |
7463 | ||
7464 | @item void __IACCsetl (iacc @var{reg}, sw1 @var{x}) | |
7465 | Set IACC0H to @var{x} if @var{reg} is 0 and IACC0L to @var{x} if @var{reg} | |
7466 | is 1. Other values of @var{reg} are rejected as invalid. | |
7467 | ||
7468 | @item void __data_prefetch0 (const void *@var{x}) | |
7469 | Use the @code{dcpl} instruction to load the contents of address @var{x} | |
7470 | into the data cache. | |
7471 | ||
7472 | @item void __data_prefetch (const void *@var{x}) | |
7473 | Use the @code{nldub} instruction to load the contents of address @var{x} | |
7474 | into the data cache. The instruction will be issued in slot I1@. | |
7475 | @end table | |
7476 | ||
0975678f JM |
7477 | @node X86 Built-in Functions |
7478 | @subsection X86 Built-in Functions | |
7479 | ||
7480 | These built-in functions are available for the i386 and x86-64 family | |
7481 | of computers, depending on the command-line switches used. | |
7482 | ||
75576871 BB |
7483 | Note that, if you specify command-line switches such as @option{-msse}, |
7484 | the compiler could use the extended instruction sets even if the built-ins | |
7485 | are not used explicitly in the program. For this reason, applications | |
7486 | which perform runtime CPU detection must compile separate files for each | |
7487 | supported architecture, using the appropriate flags. In particular, | |
7488 | the file containing the CPU detection code should be compiled without | |
7489 | these options. | |
7490 | ||
0975678f | 7491 | The following machine modes are available for use with MMX built-in functions |
333c8841 AH |
7492 | (@pxref{Vector Extensions}): @code{V2SI} for a vector of two 32-bit integers, |
7493 | @code{V4HI} for a vector of four 16-bit integers, and @code{V8QI} for a | |
7494 | vector of eight 8-bit integers. Some of the built-in functions operate on | |
75d8b30e | 7495 | MMX registers as a whole 64-bit entity, these use @code{V1DI} as their mode. |
0975678f JM |
7496 | |
7497 | If 3Dnow extensions are enabled, @code{V2SF} is used as a mode for a vector | |
333c8841 | 7498 | of two 32-bit floating point values. |
0975678f | 7499 | |
333c8841 AH |
7500 | If SSE extensions are enabled, @code{V4SF} is used for a vector of four 32-bit |
7501 | floating point values. Some instructions use a vector of four 32-bit | |
0975678f | 7502 | integers, these use @code{V4SI}. Finally, some instructions operate on an |
333c8841 | 7503 | entire vector register, interpreting it as a 128-bit integer, these use mode |
0975678f JM |
7504 | @code{TI}. |
7505 | ||
27f56cb1 | 7506 | In 64-bit mode, the x86-64 family of processors uses additional built-in |
5513e239 UB |
7507 | functions for efficient use of @code{TF} (@code{__float128}) 128-bit |
7508 | floating point and @code{TC} 128-bit complex floating point values. | |
7509 | ||
27f56cb1 GP |
7510 | The following floating point built-in functions are available in 64-bit |
7511 | mode. All of them implement the function that is part of the name. | |
5513e239 UB |
7512 | |
7513 | @smallexample | |
7514 | __float128 __builtin_fabsq (__float128) | |
7515 | __float128 __builtin_copysignq (__float128, __float128) | |
7516 | @end smallexample | |
7517 | ||
7518 | The following floating point built-in functions are made available in the | |
7519 | 64-bit mode. | |
7520 | ||
7521 | @table @code | |
7522 | @item __float128 __builtin_infq (void) | |
7523 | Similar to @code{__builtin_inf}, except the return type is @code{__float128}. | |
7524 | @end table | |
7525 | ||
0975678f JM |
7526 | The following built-in functions are made available by @option{-mmmx}. |
7527 | All of them generate the machine instruction that is part of the name. | |
7528 | ||
3ab51846 | 7529 | @smallexample |
0975678f JM |
7530 | v8qi __builtin_ia32_paddb (v8qi, v8qi) |
7531 | v4hi __builtin_ia32_paddw (v4hi, v4hi) | |
7532 | v2si __builtin_ia32_paddd (v2si, v2si) | |
7533 | v8qi __builtin_ia32_psubb (v8qi, v8qi) | |
7534 | v4hi __builtin_ia32_psubw (v4hi, v4hi) | |
7535 | v2si __builtin_ia32_psubd (v2si, v2si) | |
7536 | v8qi __builtin_ia32_paddsb (v8qi, v8qi) | |
7537 | v4hi __builtin_ia32_paddsw (v4hi, v4hi) | |
7538 | v8qi __builtin_ia32_psubsb (v8qi, v8qi) | |
7539 | v4hi __builtin_ia32_psubsw (v4hi, v4hi) | |
7540 | v8qi __builtin_ia32_paddusb (v8qi, v8qi) | |
7541 | v4hi __builtin_ia32_paddusw (v4hi, v4hi) | |
7542 | v8qi __builtin_ia32_psubusb (v8qi, v8qi) | |
7543 | v4hi __builtin_ia32_psubusw (v4hi, v4hi) | |
7544 | v4hi __builtin_ia32_pmullw (v4hi, v4hi) | |
7545 | v4hi __builtin_ia32_pmulhw (v4hi, v4hi) | |
7546 | di __builtin_ia32_pand (di, di) | |
7547 | di __builtin_ia32_pandn (di,di) | |
7548 | di __builtin_ia32_por (di, di) | |
7549 | di __builtin_ia32_pxor (di, di) | |
7550 | v8qi __builtin_ia32_pcmpeqb (v8qi, v8qi) | |
7551 | v4hi __builtin_ia32_pcmpeqw (v4hi, v4hi) | |
7552 | v2si __builtin_ia32_pcmpeqd (v2si, v2si) | |
7553 | v8qi __builtin_ia32_pcmpgtb (v8qi, v8qi) | |
7554 | v4hi __builtin_ia32_pcmpgtw (v4hi, v4hi) | |
7555 | v2si __builtin_ia32_pcmpgtd (v2si, v2si) | |
7556 | v8qi __builtin_ia32_punpckhbw (v8qi, v8qi) | |
7557 | v4hi __builtin_ia32_punpckhwd (v4hi, v4hi) | |
7558 | v2si __builtin_ia32_punpckhdq (v2si, v2si) | |
7559 | v8qi __builtin_ia32_punpcklbw (v8qi, v8qi) | |
7560 | v4hi __builtin_ia32_punpcklwd (v4hi, v4hi) | |
7561 | v2si __builtin_ia32_punpckldq (v2si, v2si) | |
7562 | v8qi __builtin_ia32_packsswb (v4hi, v4hi) | |
7563 | v4hi __builtin_ia32_packssdw (v2si, v2si) | |
7564 | v8qi __builtin_ia32_packuswb (v4hi, v4hi) | |
10a97ae6 | 7565 | |
52eaae97 UB |
7566 | v4hi __builtin_ia32_psllw (v4hi, v4hi) |
7567 | v2si __builtin_ia32_pslld (v2si, v2si) | |
7568 | v1di __builtin_ia32_psllq (v1di, v1di) | |
7569 | v4hi __builtin_ia32_psrlw (v4hi, v4hi) | |
7570 | v2si __builtin_ia32_psrld (v2si, v2si) | |
7571 | v1di __builtin_ia32_psrlq (v1di, v1di) | |
7572 | v4hi __builtin_ia32_psraw (v4hi, v4hi) | |
7573 | v2si __builtin_ia32_psrad (v2si, v2si) | |
7574 | v4hi __builtin_ia32_psllwi (v4hi, int) | |
7575 | v2si __builtin_ia32_pslldi (v2si, int) | |
7576 | v1di __builtin_ia32_psllqi (v1di, int) | |
7577 | v4hi __builtin_ia32_psrlwi (v4hi, int) | |
7578 | v2si __builtin_ia32_psrldi (v2si, int) | |
7579 | v1di __builtin_ia32_psrlqi (v1di, int) | |
7580 | v4hi __builtin_ia32_psrawi (v4hi, int) | |
7581 | v2si __builtin_ia32_psradi (v2si, int) | |
10a97ae6 | 7582 | |
3ab51846 | 7583 | @end smallexample |
0975678f JM |
7584 | |
7585 | The following built-in functions are made available either with | |
7586 | @option{-msse}, or with a combination of @option{-m3dnow} and | |
7587 | @option{-march=athlon}. All of them generate the machine | |
7588 | instruction that is part of the name. | |
7589 | ||
3ab51846 | 7590 | @smallexample |
0975678f JM |
7591 | v4hi __builtin_ia32_pmulhuw (v4hi, v4hi) |
7592 | v8qi __builtin_ia32_pavgb (v8qi, v8qi) | |
7593 | v4hi __builtin_ia32_pavgw (v4hi, v4hi) | |
ab555a5b | 7594 | v1di __builtin_ia32_psadbw (v8qi, v8qi) |
0975678f JM |
7595 | v8qi __builtin_ia32_pmaxub (v8qi, v8qi) |
7596 | v4hi __builtin_ia32_pmaxsw (v4hi, v4hi) | |
7597 | v8qi __builtin_ia32_pminub (v8qi, v8qi) | |
7598 | v4hi __builtin_ia32_pminsw (v4hi, v4hi) | |
7599 | int __builtin_ia32_pextrw (v4hi, int) | |
7600 | v4hi __builtin_ia32_pinsrw (v4hi, int, int) | |
7601 | int __builtin_ia32_pmovmskb (v8qi) | |
7602 | void __builtin_ia32_maskmovq (v8qi, v8qi, char *) | |
7603 | void __builtin_ia32_movntq (di *, di) | |
7604 | void __builtin_ia32_sfence (void) | |
3ab51846 | 7605 | @end smallexample |
0975678f JM |
7606 | |
7607 | The following built-in functions are available when @option{-msse} is used. | |
7608 | All of them generate the machine instruction that is part of the name. | |
7609 | ||
3ab51846 | 7610 | @smallexample |
0975678f JM |
7611 | int __builtin_ia32_comieq (v4sf, v4sf) |
7612 | int __builtin_ia32_comineq (v4sf, v4sf) | |
7613 | int __builtin_ia32_comilt (v4sf, v4sf) | |
7614 | int __builtin_ia32_comile (v4sf, v4sf) | |
7615 | int __builtin_ia32_comigt (v4sf, v4sf) | |
7616 | int __builtin_ia32_comige (v4sf, v4sf) | |
7617 | int __builtin_ia32_ucomieq (v4sf, v4sf) | |
7618 | int __builtin_ia32_ucomineq (v4sf, v4sf) | |
7619 | int __builtin_ia32_ucomilt (v4sf, v4sf) | |
7620 | int __builtin_ia32_ucomile (v4sf, v4sf) | |
7621 | int __builtin_ia32_ucomigt (v4sf, v4sf) | |
7622 | int __builtin_ia32_ucomige (v4sf, v4sf) | |
7623 | v4sf __builtin_ia32_addps (v4sf, v4sf) | |
7624 | v4sf __builtin_ia32_subps (v4sf, v4sf) | |
7625 | v4sf __builtin_ia32_mulps (v4sf, v4sf) | |
7626 | v4sf __builtin_ia32_divps (v4sf, v4sf) | |
7627 | v4sf __builtin_ia32_addss (v4sf, v4sf) | |
7628 | v4sf __builtin_ia32_subss (v4sf, v4sf) | |
7629 | v4sf __builtin_ia32_mulss (v4sf, v4sf) | |
7630 | v4sf __builtin_ia32_divss (v4sf, v4sf) | |
7631 | v4si __builtin_ia32_cmpeqps (v4sf, v4sf) | |
7632 | v4si __builtin_ia32_cmpltps (v4sf, v4sf) | |
7633 | v4si __builtin_ia32_cmpleps (v4sf, v4sf) | |
7634 | v4si __builtin_ia32_cmpgtps (v4sf, v4sf) | |
7635 | v4si __builtin_ia32_cmpgeps (v4sf, v4sf) | |
7636 | v4si __builtin_ia32_cmpunordps (v4sf, v4sf) | |
7637 | v4si __builtin_ia32_cmpneqps (v4sf, v4sf) | |
7638 | v4si __builtin_ia32_cmpnltps (v4sf, v4sf) | |
7639 | v4si __builtin_ia32_cmpnleps (v4sf, v4sf) | |
7640 | v4si __builtin_ia32_cmpngtps (v4sf, v4sf) | |
7641 | v4si __builtin_ia32_cmpngeps (v4sf, v4sf) | |
7642 | v4si __builtin_ia32_cmpordps (v4sf, v4sf) | |
7643 | v4si __builtin_ia32_cmpeqss (v4sf, v4sf) | |
7644 | v4si __builtin_ia32_cmpltss (v4sf, v4sf) | |
7645 | v4si __builtin_ia32_cmpless (v4sf, v4sf) | |
0975678f JM |
7646 | v4si __builtin_ia32_cmpunordss (v4sf, v4sf) |
7647 | v4si __builtin_ia32_cmpneqss (v4sf, v4sf) | |
7648 | v4si __builtin_ia32_cmpnlts (v4sf, v4sf) | |
7649 | v4si __builtin_ia32_cmpnless (v4sf, v4sf) | |
0975678f JM |
7650 | v4si __builtin_ia32_cmpordss (v4sf, v4sf) |
7651 | v4sf __builtin_ia32_maxps (v4sf, v4sf) | |
7652 | v4sf __builtin_ia32_maxss (v4sf, v4sf) | |
7653 | v4sf __builtin_ia32_minps (v4sf, v4sf) | |
7654 | v4sf __builtin_ia32_minss (v4sf, v4sf) | |
7655 | v4sf __builtin_ia32_andps (v4sf, v4sf) | |
7656 | v4sf __builtin_ia32_andnps (v4sf, v4sf) | |
7657 | v4sf __builtin_ia32_orps (v4sf, v4sf) | |
7658 | v4sf __builtin_ia32_xorps (v4sf, v4sf) | |
7659 | v4sf __builtin_ia32_movss (v4sf, v4sf) | |
7660 | v4sf __builtin_ia32_movhlps (v4sf, v4sf) | |
7661 | v4sf __builtin_ia32_movlhps (v4sf, v4sf) | |
7662 | v4sf __builtin_ia32_unpckhps (v4sf, v4sf) | |
7663 | v4sf __builtin_ia32_unpcklps (v4sf, v4sf) | |
7664 | v4sf __builtin_ia32_cvtpi2ps (v4sf, v2si) | |
7665 | v4sf __builtin_ia32_cvtsi2ss (v4sf, int) | |
7666 | v2si __builtin_ia32_cvtps2pi (v4sf) | |
7667 | int __builtin_ia32_cvtss2si (v4sf) | |
7668 | v2si __builtin_ia32_cvttps2pi (v4sf) | |
7669 | int __builtin_ia32_cvttss2si (v4sf) | |
7670 | v4sf __builtin_ia32_rcpps (v4sf) | |
7671 | v4sf __builtin_ia32_rsqrtps (v4sf) | |
7672 | v4sf __builtin_ia32_sqrtps (v4sf) | |
7673 | v4sf __builtin_ia32_rcpss (v4sf) | |
7674 | v4sf __builtin_ia32_rsqrtss (v4sf) | |
7675 | v4sf __builtin_ia32_sqrtss (v4sf) | |
7676 | v4sf __builtin_ia32_shufps (v4sf, v4sf, int) | |
7677 | void __builtin_ia32_movntps (float *, v4sf) | |
7678 | int __builtin_ia32_movmskps (v4sf) | |
3ab51846 | 7679 | @end smallexample |
0975678f JM |
7680 | |
7681 | The following built-in functions are available when @option{-msse} is used. | |
7682 | ||
7683 | @table @code | |
7684 | @item v4sf __builtin_ia32_loadaps (float *) | |
7685 | Generates the @code{movaps} machine instruction as a load from memory. | |
7686 | @item void __builtin_ia32_storeaps (float *, v4sf) | |
7687 | Generates the @code{movaps} machine instruction as a store to memory. | |
7688 | @item v4sf __builtin_ia32_loadups (float *) | |
7689 | Generates the @code{movups} machine instruction as a load from memory. | |
7690 | @item void __builtin_ia32_storeups (float *, v4sf) | |
7691 | Generates the @code{movups} machine instruction as a store to memory. | |
7692 | @item v4sf __builtin_ia32_loadsss (float *) | |
7693 | Generates the @code{movss} machine instruction as a load from memory. | |
7694 | @item void __builtin_ia32_storess (float *, v4sf) | |
7695 | Generates the @code{movss} machine instruction as a store to memory. | |
bb1418c1 | 7696 | @item v4sf __builtin_ia32_loadhps (v4sf, const v2sf *) |
0975678f | 7697 | Generates the @code{movhps} machine instruction as a load from memory. |
bb1418c1 | 7698 | @item v4sf __builtin_ia32_loadlps (v4sf, const v2sf *) |
0975678f | 7699 | Generates the @code{movlps} machine instruction as a load from memory |
bb1418c1 | 7700 | @item void __builtin_ia32_storehps (v2sf *, v4sf) |
0975678f | 7701 | Generates the @code{movhps} machine instruction as a store to memory. |
bb1418c1 | 7702 | @item void __builtin_ia32_storelps (v2sf *, v4sf) |
0975678f JM |
7703 | Generates the @code{movlps} machine instruction as a store to memory. |
7704 | @end table | |
7705 | ||
d7aa4788 RG |
7706 | The following built-in functions are available when @option{-msse2} is used. |
7707 | All of them generate the machine instruction that is part of the name. | |
7708 | ||
7709 | @smallexample | |
7710 | int __builtin_ia32_comisdeq (v2df, v2df) | |
7711 | int __builtin_ia32_comisdlt (v2df, v2df) | |
7712 | int __builtin_ia32_comisdle (v2df, v2df) | |
7713 | int __builtin_ia32_comisdgt (v2df, v2df) | |
7714 | int __builtin_ia32_comisdge (v2df, v2df) | |
7715 | int __builtin_ia32_comisdneq (v2df, v2df) | |
7716 | int __builtin_ia32_ucomisdeq (v2df, v2df) | |
7717 | int __builtin_ia32_ucomisdlt (v2df, v2df) | |
7718 | int __builtin_ia32_ucomisdle (v2df, v2df) | |
7719 | int __builtin_ia32_ucomisdgt (v2df, v2df) | |
7720 | int __builtin_ia32_ucomisdge (v2df, v2df) | |
7721 | int __builtin_ia32_ucomisdneq (v2df, v2df) | |
7722 | v2df __builtin_ia32_cmpeqpd (v2df, v2df) | |
7723 | v2df __builtin_ia32_cmpltpd (v2df, v2df) | |
7724 | v2df __builtin_ia32_cmplepd (v2df, v2df) | |
7725 | v2df __builtin_ia32_cmpgtpd (v2df, v2df) | |
7726 | v2df __builtin_ia32_cmpgepd (v2df, v2df) | |
7727 | v2df __builtin_ia32_cmpunordpd (v2df, v2df) | |
7728 | v2df __builtin_ia32_cmpneqpd (v2df, v2df) | |
7729 | v2df __builtin_ia32_cmpnltpd (v2df, v2df) | |
7730 | v2df __builtin_ia32_cmpnlepd (v2df, v2df) | |
7731 | v2df __builtin_ia32_cmpngtpd (v2df, v2df) | |
7732 | v2df __builtin_ia32_cmpngepd (v2df, v2df) | |
7733 | v2df __builtin_ia32_cmpordpd (v2df, v2df) | |
7734 | v2df __builtin_ia32_cmpeqsd (v2df, v2df) | |
7735 | v2df __builtin_ia32_cmpltsd (v2df, v2df) | |
7736 | v2df __builtin_ia32_cmplesd (v2df, v2df) | |
7737 | v2df __builtin_ia32_cmpunordsd (v2df, v2df) | |
7738 | v2df __builtin_ia32_cmpneqsd (v2df, v2df) | |
7739 | v2df __builtin_ia32_cmpnltsd (v2df, v2df) | |
7740 | v2df __builtin_ia32_cmpnlesd (v2df, v2df) | |
7741 | v2df __builtin_ia32_cmpordsd (v2df, v2df) | |
7742 | v2di __builtin_ia32_paddq (v2di, v2di) | |
7743 | v2di __builtin_ia32_psubq (v2di, v2di) | |
7744 | v2df __builtin_ia32_addpd (v2df, v2df) | |
7745 | v2df __builtin_ia32_subpd (v2df, v2df) | |
7746 | v2df __builtin_ia32_mulpd (v2df, v2df) | |
7747 | v2df __builtin_ia32_divpd (v2df, v2df) | |
7748 | v2df __builtin_ia32_addsd (v2df, v2df) | |
7749 | v2df __builtin_ia32_subsd (v2df, v2df) | |
7750 | v2df __builtin_ia32_mulsd (v2df, v2df) | |
7751 | v2df __builtin_ia32_divsd (v2df, v2df) | |
7752 | v2df __builtin_ia32_minpd (v2df, v2df) | |
7753 | v2df __builtin_ia32_maxpd (v2df, v2df) | |
7754 | v2df __builtin_ia32_minsd (v2df, v2df) | |
7755 | v2df __builtin_ia32_maxsd (v2df, v2df) | |
7756 | v2df __builtin_ia32_andpd (v2df, v2df) | |
7757 | v2df __builtin_ia32_andnpd (v2df, v2df) | |
7758 | v2df __builtin_ia32_orpd (v2df, v2df) | |
7759 | v2df __builtin_ia32_xorpd (v2df, v2df) | |
7760 | v2df __builtin_ia32_movsd (v2df, v2df) | |
7761 | v2df __builtin_ia32_unpckhpd (v2df, v2df) | |
7762 | v2df __builtin_ia32_unpcklpd (v2df, v2df) | |
7763 | v16qi __builtin_ia32_paddb128 (v16qi, v16qi) | |
7764 | v8hi __builtin_ia32_paddw128 (v8hi, v8hi) | |
7765 | v4si __builtin_ia32_paddd128 (v4si, v4si) | |
7766 | v2di __builtin_ia32_paddq128 (v2di, v2di) | |
7767 | v16qi __builtin_ia32_psubb128 (v16qi, v16qi) | |
7768 | v8hi __builtin_ia32_psubw128 (v8hi, v8hi) | |
7769 | v4si __builtin_ia32_psubd128 (v4si, v4si) | |
7770 | v2di __builtin_ia32_psubq128 (v2di, v2di) | |
7771 | v8hi __builtin_ia32_pmullw128 (v8hi, v8hi) | |
7772 | v8hi __builtin_ia32_pmulhw128 (v8hi, v8hi) | |
7773 | v2di __builtin_ia32_pand128 (v2di, v2di) | |
7774 | v2di __builtin_ia32_pandn128 (v2di, v2di) | |
7775 | v2di __builtin_ia32_por128 (v2di, v2di) | |
7776 | v2di __builtin_ia32_pxor128 (v2di, v2di) | |
7777 | v16qi __builtin_ia32_pavgb128 (v16qi, v16qi) | |
7778 | v8hi __builtin_ia32_pavgw128 (v8hi, v8hi) | |
7779 | v16qi __builtin_ia32_pcmpeqb128 (v16qi, v16qi) | |
7780 | v8hi __builtin_ia32_pcmpeqw128 (v8hi, v8hi) | |
7781 | v4si __builtin_ia32_pcmpeqd128 (v4si, v4si) | |
7782 | v16qi __builtin_ia32_pcmpgtb128 (v16qi, v16qi) | |
7783 | v8hi __builtin_ia32_pcmpgtw128 (v8hi, v8hi) | |
7784 | v4si __builtin_ia32_pcmpgtd128 (v4si, v4si) | |
7785 | v16qi __builtin_ia32_pmaxub128 (v16qi, v16qi) | |
7786 | v8hi __builtin_ia32_pmaxsw128 (v8hi, v8hi) | |
7787 | v16qi __builtin_ia32_pminub128 (v16qi, v16qi) | |
7788 | v8hi __builtin_ia32_pminsw128 (v8hi, v8hi) | |
7789 | v16qi __builtin_ia32_punpckhbw128 (v16qi, v16qi) | |
7790 | v8hi __builtin_ia32_punpckhwd128 (v8hi, v8hi) | |
7791 | v4si __builtin_ia32_punpckhdq128 (v4si, v4si) | |
7792 | v2di __builtin_ia32_punpckhqdq128 (v2di, v2di) | |
7793 | v16qi __builtin_ia32_punpcklbw128 (v16qi, v16qi) | |
7794 | v8hi __builtin_ia32_punpcklwd128 (v8hi, v8hi) | |
7795 | v4si __builtin_ia32_punpckldq128 (v4si, v4si) | |
7796 | v2di __builtin_ia32_punpcklqdq128 (v2di, v2di) | |
1b667c82 L |
7797 | v16qi __builtin_ia32_packsswb128 (v8hi, v8hi) |
7798 | v8hi __builtin_ia32_packssdw128 (v4si, v4si) | |
7799 | v16qi __builtin_ia32_packuswb128 (v8hi, v8hi) | |
d7aa4788 RG |
7800 | v8hi __builtin_ia32_pmulhuw128 (v8hi, v8hi) |
7801 | void __builtin_ia32_maskmovdqu (v16qi, v16qi) | |
7802 | v2df __builtin_ia32_loadupd (double *) | |
7803 | void __builtin_ia32_storeupd (double *, v2df) | |
bb1418c1 L |
7804 | v2df __builtin_ia32_loadhpd (v2df, double const *) |
7805 | v2df __builtin_ia32_loadlpd (v2df, double const *) | |
d7aa4788 RG |
7806 | int __builtin_ia32_movmskpd (v2df) |
7807 | int __builtin_ia32_pmovmskb128 (v16qi) | |
7808 | void __builtin_ia32_movnti (int *, int) | |
7809 | void __builtin_ia32_movntpd (double *, v2df) | |
7810 | void __builtin_ia32_movntdq (v2df *, v2df) | |
7811 | v4si __builtin_ia32_pshufd (v4si, int) | |
7812 | v8hi __builtin_ia32_pshuflw (v8hi, int) | |
7813 | v8hi __builtin_ia32_pshufhw (v8hi, int) | |
7814 | v2di __builtin_ia32_psadbw128 (v16qi, v16qi) | |
7815 | v2df __builtin_ia32_sqrtpd (v2df) | |
7816 | v2df __builtin_ia32_sqrtsd (v2df) | |
7817 | v2df __builtin_ia32_shufpd (v2df, v2df, int) | |
7818 | v2df __builtin_ia32_cvtdq2pd (v4si) | |
7819 | v4sf __builtin_ia32_cvtdq2ps (v4si) | |
7820 | v4si __builtin_ia32_cvtpd2dq (v2df) | |
7821 | v2si __builtin_ia32_cvtpd2pi (v2df) | |
7822 | v4sf __builtin_ia32_cvtpd2ps (v2df) | |
7823 | v4si __builtin_ia32_cvttpd2dq (v2df) | |
7824 | v2si __builtin_ia32_cvttpd2pi (v2df) | |
7825 | v2df __builtin_ia32_cvtpi2pd (v2si) | |
7826 | int __builtin_ia32_cvtsd2si (v2df) | |
7827 | int __builtin_ia32_cvttsd2si (v2df) | |
7828 | long long __builtin_ia32_cvtsd2si64 (v2df) | |
7829 | long long __builtin_ia32_cvttsd2si64 (v2df) | |
7830 | v4si __builtin_ia32_cvtps2dq (v4sf) | |
7831 | v2df __builtin_ia32_cvtps2pd (v4sf) | |
7832 | v4si __builtin_ia32_cvttps2dq (v4sf) | |
7833 | v2df __builtin_ia32_cvtsi2sd (v2df, int) | |
7834 | v2df __builtin_ia32_cvtsi642sd (v2df, long long) | |
7835 | v4sf __builtin_ia32_cvtsd2ss (v4sf, v2df) | |
7836 | v2df __builtin_ia32_cvtss2sd (v2df, v4sf) | |
7837 | void __builtin_ia32_clflush (const void *) | |
7838 | void __builtin_ia32_lfence (void) | |
7839 | void __builtin_ia32_mfence (void) | |
7840 | v16qi __builtin_ia32_loaddqu (const char *) | |
7841 | void __builtin_ia32_storedqu (char *, v16qi) | |
ab555a5b | 7842 | v1di __builtin_ia32_pmuludq (v2si, v2si) |
d7aa4788 | 7843 | v2di __builtin_ia32_pmuludq128 (v4si, v4si) |
52eaae97 UB |
7844 | v8hi __builtin_ia32_psllw128 (v8hi, v8hi) |
7845 | v4si __builtin_ia32_pslld128 (v4si, v4si) | |
7846 | v2di __builtin_ia32_psllq128 (v2di, v2di) | |
7847 | v8hi __builtin_ia32_psrlw128 (v8hi, v8hi) | |
7848 | v4si __builtin_ia32_psrld128 (v4si, v4si) | |
d7aa4788 | 7849 | v2di __builtin_ia32_psrlq128 (v2di, v2di) |
52eaae97 UB |
7850 | v8hi __builtin_ia32_psraw128 (v8hi, v8hi) |
7851 | v4si __builtin_ia32_psrad128 (v4si, v4si) | |
d7aa4788 RG |
7852 | v2di __builtin_ia32_pslldqi128 (v2di, int) |
7853 | v8hi __builtin_ia32_psllwi128 (v8hi, int) | |
7854 | v4si __builtin_ia32_pslldi128 (v4si, int) | |
7855 | v2di __builtin_ia32_psllqi128 (v2di, int) | |
7856 | v2di __builtin_ia32_psrldqi128 (v2di, int) | |
7857 | v8hi __builtin_ia32_psrlwi128 (v8hi, int) | |
7858 | v4si __builtin_ia32_psrldi128 (v4si, int) | |
7859 | v2di __builtin_ia32_psrlqi128 (v2di, int) | |
7860 | v8hi __builtin_ia32_psrawi128 (v8hi, int) | |
7861 | v4si __builtin_ia32_psradi128 (v4si, int) | |
7862 | v4si __builtin_ia32_pmaddwd128 (v8hi, v8hi) | |
7863 | @end smallexample | |
7864 | ||
9e200aaf | 7865 | The following built-in functions are available when @option{-msse3} is used. |
22c7c85e L |
7866 | All of them generate the machine instruction that is part of the name. |
7867 | ||
3ab51846 | 7868 | @smallexample |
22c7c85e | 7869 | v2df __builtin_ia32_addsubpd (v2df, v2df) |
d7aa4788 | 7870 | v4sf __builtin_ia32_addsubps (v4sf, v4sf) |
22c7c85e | 7871 | v2df __builtin_ia32_haddpd (v2df, v2df) |
d7aa4788 | 7872 | v4sf __builtin_ia32_haddps (v4sf, v4sf) |
22c7c85e | 7873 | v2df __builtin_ia32_hsubpd (v2df, v2df) |
d7aa4788 | 7874 | v4sf __builtin_ia32_hsubps (v4sf, v4sf) |
22c7c85e L |
7875 | v16qi __builtin_ia32_lddqu (char const *) |
7876 | void __builtin_ia32_monitor (void *, unsigned int, unsigned int) | |
7877 | v2df __builtin_ia32_movddup (v2df) | |
7878 | v4sf __builtin_ia32_movshdup (v4sf) | |
7879 | v4sf __builtin_ia32_movsldup (v4sf) | |
7880 | void __builtin_ia32_mwait (unsigned int, unsigned int) | |
3ab51846 | 7881 | @end smallexample |
22c7c85e | 7882 | |
9e200aaf | 7883 | The following built-in functions are available when @option{-msse3} is used. |
22c7c85e L |
7884 | |
7885 | @table @code | |
7886 | @item v2df __builtin_ia32_loadddup (double const *) | |
7887 | Generates the @code{movddup} machine instruction as a load from memory. | |
7888 | @end table | |
7889 | ||
b1875f52 L |
7890 | The following built-in functions are available when @option{-mssse3} is used. |
7891 | All of them generate the machine instruction that is part of the name | |
7892 | with MMX registers. | |
7893 | ||
7894 | @smallexample | |
7895 | v2si __builtin_ia32_phaddd (v2si, v2si) | |
7896 | v4hi __builtin_ia32_phaddw (v4hi, v4hi) | |
7897 | v4hi __builtin_ia32_phaddsw (v4hi, v4hi) | |
7898 | v2si __builtin_ia32_phsubd (v2si, v2si) | |
7899 | v4hi __builtin_ia32_phsubw (v4hi, v4hi) | |
7900 | v4hi __builtin_ia32_phsubsw (v4hi, v4hi) | |
1b667c82 | 7901 | v4hi __builtin_ia32_pmaddubsw (v8qi, v8qi) |
b1875f52 L |
7902 | v4hi __builtin_ia32_pmulhrsw (v4hi, v4hi) |
7903 | v8qi __builtin_ia32_pshufb (v8qi, v8qi) | |
7904 | v8qi __builtin_ia32_psignb (v8qi, v8qi) | |
7905 | v2si __builtin_ia32_psignd (v2si, v2si) | |
7906 | v4hi __builtin_ia32_psignw (v4hi, v4hi) | |
99c25ac1 | 7907 | v1di __builtin_ia32_palignr (v1di, v1di, int) |
b1875f52 L |
7908 | v8qi __builtin_ia32_pabsb (v8qi) |
7909 | v2si __builtin_ia32_pabsd (v2si) | |
7910 | v4hi __builtin_ia32_pabsw (v4hi) | |
7911 | @end smallexample | |
7912 | ||
7913 | The following built-in functions are available when @option{-mssse3} is used. | |
7914 | All of them generate the machine instruction that is part of the name | |
7915 | with SSE registers. | |
7916 | ||
7917 | @smallexample | |
7918 | v4si __builtin_ia32_phaddd128 (v4si, v4si) | |
7919 | v8hi __builtin_ia32_phaddw128 (v8hi, v8hi) | |
7920 | v8hi __builtin_ia32_phaddsw128 (v8hi, v8hi) | |
7921 | v4si __builtin_ia32_phsubd128 (v4si, v4si) | |
7922 | v8hi __builtin_ia32_phsubw128 (v8hi, v8hi) | |
7923 | v8hi __builtin_ia32_phsubsw128 (v8hi, v8hi) | |
1b667c82 | 7924 | v8hi __builtin_ia32_pmaddubsw128 (v16qi, v16qi) |
b1875f52 L |
7925 | v8hi __builtin_ia32_pmulhrsw128 (v8hi, v8hi) |
7926 | v16qi __builtin_ia32_pshufb128 (v16qi, v16qi) | |
7927 | v16qi __builtin_ia32_psignb128 (v16qi, v16qi) | |
7928 | v4si __builtin_ia32_psignd128 (v4si, v4si) | |
7929 | v8hi __builtin_ia32_psignw128 (v8hi, v8hi) | |
858e5e79 | 7930 | v2di __builtin_ia32_palignr128 (v2di, v2di, int) |
b1875f52 L |
7931 | v16qi __builtin_ia32_pabsb128 (v16qi) |
7932 | v4si __builtin_ia32_pabsd128 (v4si) | |
7933 | v8hi __builtin_ia32_pabsw128 (v8hi) | |
7934 | @end smallexample | |
7935 | ||
9a5cee02 L |
7936 | The following built-in functions are available when @option{-msse4.1} is |
7937 | used. All of them generate the machine instruction that is part of the | |
7938 | name. | |
7939 | ||
7940 | @smallexample | |
7941 | v2df __builtin_ia32_blendpd (v2df, v2df, const int) | |
7942 | v4sf __builtin_ia32_blendps (v4sf, v4sf, const int) | |
7943 | v2df __builtin_ia32_blendvpd (v2df, v2df, v2df) | |
7944 | v4sf __builtin_ia32_blendvps (v4sf, v4sf, v4sf) | |
291d9a2d | 7945 | v2df __builtin_ia32_dppd (v2df, v2df, const int) |
9a5cee02 L |
7946 | v4sf __builtin_ia32_dpps (v4sf, v4sf, const int) |
7947 | v4sf __builtin_ia32_insertps128 (v4sf, v4sf, const int) | |
7948 | v2di __builtin_ia32_movntdqa (v2di *); | |
7949 | v16qi __builtin_ia32_mpsadbw128 (v16qi, v16qi, const int) | |
7950 | v8hi __builtin_ia32_packusdw128 (v4si, v4si) | |
7951 | v16qi __builtin_ia32_pblendvb128 (v16qi, v16qi, v16qi) | |
7952 | v8hi __builtin_ia32_pblendw128 (v8hi, v8hi, const int) | |
7953 | v2di __builtin_ia32_pcmpeqq (v2di, v2di) | |
7954 | v8hi __builtin_ia32_phminposuw128 (v8hi) | |
7955 | v16qi __builtin_ia32_pmaxsb128 (v16qi, v16qi) | |
7956 | v4si __builtin_ia32_pmaxsd128 (v4si, v4si) | |
7957 | v4si __builtin_ia32_pmaxud128 (v4si, v4si) | |
7958 | v8hi __builtin_ia32_pmaxuw128 (v8hi, v8hi) | |
7959 | v16qi __builtin_ia32_pminsb128 (v16qi, v16qi) | |
7960 | v4si __builtin_ia32_pminsd128 (v4si, v4si) | |
7961 | v4si __builtin_ia32_pminud128 (v4si, v4si) | |
7962 | v8hi __builtin_ia32_pminuw128 (v8hi, v8hi) | |
7963 | v4si __builtin_ia32_pmovsxbd128 (v16qi) | |
7964 | v2di __builtin_ia32_pmovsxbq128 (v16qi) | |
7965 | v8hi __builtin_ia32_pmovsxbw128 (v16qi) | |
7966 | v2di __builtin_ia32_pmovsxdq128 (v4si) | |
7967 | v4si __builtin_ia32_pmovsxwd128 (v8hi) | |
7968 | v2di __builtin_ia32_pmovsxwq128 (v8hi) | |
7969 | v4si __builtin_ia32_pmovzxbd128 (v16qi) | |
7970 | v2di __builtin_ia32_pmovzxbq128 (v16qi) | |
7971 | v8hi __builtin_ia32_pmovzxbw128 (v16qi) | |
7972 | v2di __builtin_ia32_pmovzxdq128 (v4si) | |
7973 | v4si __builtin_ia32_pmovzxwd128 (v8hi) | |
7974 | v2di __builtin_ia32_pmovzxwq128 (v8hi) | |
7975 | v2di __builtin_ia32_pmuldq128 (v4si, v4si) | |
7976 | v4si __builtin_ia32_pmulld128 (v4si, v4si) | |
7977 | int __builtin_ia32_ptestc128 (v2di, v2di) | |
7978 | int __builtin_ia32_ptestnzc128 (v2di, v2di) | |
7979 | int __builtin_ia32_ptestz128 (v2di, v2di) | |
7980 | v2df __builtin_ia32_roundpd (v2df, const int) | |
7981 | v4sf __builtin_ia32_roundps (v4sf, const int) | |
7982 | v2df __builtin_ia32_roundsd (v2df, v2df, const int) | |
7983 | v4sf __builtin_ia32_roundss (v4sf, v4sf, const int) | |
7984 | @end smallexample | |
7985 | ||
7986 | The following built-in functions are available when @option{-msse4.1} is | |
7987 | used. | |
7988 | ||
7989 | @table @code | |
7990 | @item v4sf __builtin_ia32_vec_set_v4sf (v4sf, float, const int) | |
7991 | Generates the @code{insertps} machine instruction. | |
7992 | @item int __builtin_ia32_vec_ext_v16qi (v16qi, const int) | |
7993 | Generates the @code{pextrb} machine instruction. | |
7994 | @item v16qi __builtin_ia32_vec_set_v16qi (v16qi, int, const int) | |
7995 | Generates the @code{pinsrb} machine instruction. | |
7996 | @item v4si __builtin_ia32_vec_set_v4si (v4si, int, const int) | |
7997 | Generates the @code{pinsrd} machine instruction. | |
7998 | @item v2di __builtin_ia32_vec_set_v2di (v2di, long long, const int) | |
7999 | Generates the @code{pinsrq} machine instruction in 64bit mode. | |
8000 | @end table | |
8001 | ||
8002 | The following built-in functions are changed to generate new SSE4.1 | |
8003 | instructions when @option{-msse4.1} is used. | |
8004 | ||
8005 | @table @code | |
8006 | @item float __builtin_ia32_vec_ext_v4sf (v4sf, const int) | |
8007 | Generates the @code{extractps} machine instruction. | |
8008 | @item int __builtin_ia32_vec_ext_v4si (v4si, const int) | |
8009 | Generates the @code{pextrd} machine instruction. | |
8010 | @item long long __builtin_ia32_vec_ext_v2di (v2di, const int) | |
8011 | Generates the @code{pextrq} machine instruction in 64bit mode. | |
8012 | @end table | |
8013 | ||
3b8dd071 L |
8014 | The following built-in functions are available when @option{-msse4.2} is |
8015 | used. All of them generate the machine instruction that is part of the | |
8016 | name. | |
8017 | ||
8018 | @smallexample | |
8019 | v16qi __builtin_ia32_pcmpestrm128 (v16qi, int, v16qi, int, const int) | |
8020 | int __builtin_ia32_pcmpestri128 (v16qi, int, v16qi, int, const int) | |
8021 | int __builtin_ia32_pcmpestria128 (v16qi, int, v16qi, int, const int) | |
8022 | int __builtin_ia32_pcmpestric128 (v16qi, int, v16qi, int, const int) | |
8023 | int __builtin_ia32_pcmpestrio128 (v16qi, int, v16qi, int, const int) | |
8024 | int __builtin_ia32_pcmpestris128 (v16qi, int, v16qi, int, const int) | |
8025 | int __builtin_ia32_pcmpestriz128 (v16qi, int, v16qi, int, const int) | |
8026 | v16qi __builtin_ia32_pcmpistrm128 (v16qi, v16qi, const int) | |
8027 | int __builtin_ia32_pcmpistri128 (v16qi, v16qi, const int) | |
8028 | int __builtin_ia32_pcmpistria128 (v16qi, v16qi, const int) | |
8029 | int __builtin_ia32_pcmpistric128 (v16qi, v16qi, const int) | |
8030 | int __builtin_ia32_pcmpistrio128 (v16qi, v16qi, const int) | |
8031 | int __builtin_ia32_pcmpistris128 (v16qi, v16qi, const int) | |
8032 | int __builtin_ia32_pcmpistriz128 (v16qi, v16qi, const int) | |
291d9a2d | 8033 | v2di __builtin_ia32_pcmpgtq (v2di, v2di) |
3b8dd071 L |
8034 | @end smallexample |
8035 | ||
8036 | The following built-in functions are available when @option{-msse4.2} is | |
8037 | used. | |
8038 | ||
8039 | @table @code | |
291d9a2d | 8040 | @item unsigned int __builtin_ia32_crc32qi (unsigned int, unsigned char) |
3b8dd071 | 8041 | Generates the @code{crc32b} machine instruction. |
291d9a2d | 8042 | @item unsigned int __builtin_ia32_crc32hi (unsigned int, unsigned short) |
3b8dd071 | 8043 | Generates the @code{crc32w} machine instruction. |
291d9a2d | 8044 | @item unsigned int __builtin_ia32_crc32si (unsigned int, unsigned int) |
3b8dd071 | 8045 | Generates the @code{crc32l} machine instruction. |
a44acfb9 | 8046 | @item unsigned long long __builtin_ia32_crc32di (unsigned long long, unsigned long long) |
3b8dd071 L |
8047 | @end table |
8048 | ||
8049 | The following built-in functions are changed to generate new SSE4.2 | |
8050 | instructions when @option{-msse4.2} is used. | |
8051 | ||
8052 | @table @code | |
291d9a2d | 8053 | @item int __builtin_popcount (unsigned int) |
3b8dd071 | 8054 | Generates the @code{popcntl} machine instruction. |
291d9a2d | 8055 | @item int __builtin_popcountl (unsigned long) |
3b8dd071 L |
8056 | Generates the @code{popcntl} or @code{popcntq} machine instruction, |
8057 | depending on the size of @code{unsigned long}. | |
291d9a2d | 8058 | @item int __builtin_popcountll (unsigned long long) |
3b8dd071 L |
8059 | Generates the @code{popcntq} machine instruction. |
8060 | @end table | |
8061 | ||
8b96a312 L |
8062 | The following built-in functions are available when @option{-maes} is |
8063 | used. All of them generate the machine instruction that is part of the | |
8064 | name. | |
8065 | ||
8066 | @smallexample | |
8067 | v2di __builtin_ia32_aesenc128 (v2di, v2di) | |
8068 | v2di __builtin_ia32_aesenclast128 (v2di, v2di) | |
8069 | v2di __builtin_ia32_aesdec128 (v2di, v2di) | |
8070 | v2di __builtin_ia32_aesdeclast128 (v2di, v2di) | |
8071 | v2di __builtin_ia32_aeskeygenassist128 (v2di, const int) | |
8072 | v2di __builtin_ia32_aesimc128 (v2di) | |
8073 | @end smallexample | |
8074 | ||
8075 | The following built-in function is available when @option{-mpclmul} is | |
8076 | used. | |
8077 | ||
8078 | @table @code | |
8079 | @item v2di __builtin_ia32_pclmulqdq128 (v2di, v2di, const int) | |
8080 | Generates the @code{pclmulqdq} machine instruction. | |
8081 | @end table | |
8082 | ||
21efb4d4 | 8083 | The following built-in functions are available when @option{-msse4a} is used. |
291d9a2d | 8084 | All of them generate the machine instruction that is part of the name. |
21efb4d4 HJ |
8085 | |
8086 | @smallexample | |
291d9a2d UB |
8087 | void __builtin_ia32_movntsd (double *, v2df) |
8088 | void __builtin_ia32_movntss (float *, v4sf) | |
8089 | v2di __builtin_ia32_extrq (v2di, v16qi) | |
8090 | v2di __builtin_ia32_extrqi (v2di, const unsigned int, const unsigned int) | |
8091 | v2di __builtin_ia32_insertq (v2di, v2di) | |
8092 | v2di __builtin_ia32_insertqi (v2di, v2di, const unsigned int, const unsigned int) | |
21efb4d4 HJ |
8093 | @end smallexample |
8094 | ||
04e1d06b MM |
8095 | The following built-in functions are available when @option{-msse5} is used. |
8096 | All of them generate the machine instruction that is part of the name | |
8097 | with MMX registers. | |
8098 | ||
8099 | @smallexample | |
8100 | v2df __builtin_ia32_comeqpd (v2df, v2df) | |
8101 | v2df __builtin_ia32_comeqps (v2df, v2df) | |
8102 | v4sf __builtin_ia32_comeqsd (v4sf, v4sf) | |
8103 | v4sf __builtin_ia32_comeqss (v4sf, v4sf) | |
8104 | v2df __builtin_ia32_comfalsepd (v2df, v2df) | |
8105 | v2df __builtin_ia32_comfalseps (v2df, v2df) | |
8106 | v4sf __builtin_ia32_comfalsesd (v4sf, v4sf) | |
8107 | v4sf __builtin_ia32_comfalsess (v4sf, v4sf) | |
8108 | v2df __builtin_ia32_comgepd (v2df, v2df) | |
8109 | v2df __builtin_ia32_comgeps (v2df, v2df) | |
8110 | v4sf __builtin_ia32_comgesd (v4sf, v4sf) | |
8111 | v4sf __builtin_ia32_comgess (v4sf, v4sf) | |
8112 | v2df __builtin_ia32_comgtpd (v2df, v2df) | |
8113 | v2df __builtin_ia32_comgtps (v2df, v2df) | |
8114 | v4sf __builtin_ia32_comgtsd (v4sf, v4sf) | |
8115 | v4sf __builtin_ia32_comgtss (v4sf, v4sf) | |
8116 | v2df __builtin_ia32_comlepd (v2df, v2df) | |
8117 | v2df __builtin_ia32_comleps (v2df, v2df) | |
8118 | v4sf __builtin_ia32_comlesd (v4sf, v4sf) | |
8119 | v4sf __builtin_ia32_comless (v4sf, v4sf) | |
8120 | v2df __builtin_ia32_comltpd (v2df, v2df) | |
8121 | v2df __builtin_ia32_comltps (v2df, v2df) | |
8122 | v4sf __builtin_ia32_comltsd (v4sf, v4sf) | |
8123 | v4sf __builtin_ia32_comltss (v4sf, v4sf) | |
8124 | v2df __builtin_ia32_comnepd (v2df, v2df) | |
8125 | v2df __builtin_ia32_comneps (v2df, v2df) | |
8126 | v4sf __builtin_ia32_comnesd (v4sf, v4sf) | |
8127 | v4sf __builtin_ia32_comness (v4sf, v4sf) | |
8128 | v2df __builtin_ia32_comordpd (v2df, v2df) | |
8129 | v2df __builtin_ia32_comordps (v2df, v2df) | |
8130 | v4sf __builtin_ia32_comordsd (v4sf, v4sf) | |
8131 | v4sf __builtin_ia32_comordss (v4sf, v4sf) | |
8132 | v2df __builtin_ia32_comtruepd (v2df, v2df) | |
8133 | v2df __builtin_ia32_comtrueps (v2df, v2df) | |
8134 | v4sf __builtin_ia32_comtruesd (v4sf, v4sf) | |
8135 | v4sf __builtin_ia32_comtruess (v4sf, v4sf) | |
8136 | v2df __builtin_ia32_comueqpd (v2df, v2df) | |
8137 | v2df __builtin_ia32_comueqps (v2df, v2df) | |
8138 | v4sf __builtin_ia32_comueqsd (v4sf, v4sf) | |
8139 | v4sf __builtin_ia32_comueqss (v4sf, v4sf) | |
8140 | v2df __builtin_ia32_comugepd (v2df, v2df) | |
8141 | v2df __builtin_ia32_comugeps (v2df, v2df) | |
8142 | v4sf __builtin_ia32_comugesd (v4sf, v4sf) | |
8143 | v4sf __builtin_ia32_comugess (v4sf, v4sf) | |
8144 | v2df __builtin_ia32_comugtpd (v2df, v2df) | |
8145 | v2df __builtin_ia32_comugtps (v2df, v2df) | |
8146 | v4sf __builtin_ia32_comugtsd (v4sf, v4sf) | |
8147 | v4sf __builtin_ia32_comugtss (v4sf, v4sf) | |
8148 | v2df __builtin_ia32_comulepd (v2df, v2df) | |
8149 | v2df __builtin_ia32_comuleps (v2df, v2df) | |
8150 | v4sf __builtin_ia32_comulesd (v4sf, v4sf) | |
8151 | v4sf __builtin_ia32_comuless (v4sf, v4sf) | |
8152 | v2df __builtin_ia32_comultpd (v2df, v2df) | |
8153 | v2df __builtin_ia32_comultps (v2df, v2df) | |
8154 | v4sf __builtin_ia32_comultsd (v4sf, v4sf) | |
8155 | v4sf __builtin_ia32_comultss (v4sf, v4sf) | |
8156 | v2df __builtin_ia32_comunepd (v2df, v2df) | |
8157 | v2df __builtin_ia32_comuneps (v2df, v2df) | |
8158 | v4sf __builtin_ia32_comunesd (v4sf, v4sf) | |
8159 | v4sf __builtin_ia32_comuness (v4sf, v4sf) | |
8160 | v2df __builtin_ia32_comunordpd (v2df, v2df) | |
8161 | v2df __builtin_ia32_comunordps (v2df, v2df) | |
8162 | v4sf __builtin_ia32_comunordsd (v4sf, v4sf) | |
8163 | v4sf __builtin_ia32_comunordss (v4sf, v4sf) | |
8164 | v2df __builtin_ia32_fmaddpd (v2df, v2df, v2df) | |
8165 | v4sf __builtin_ia32_fmaddps (v4sf, v4sf, v4sf) | |
8166 | v2df __builtin_ia32_fmaddsd (v2df, v2df, v2df) | |
8167 | v4sf __builtin_ia32_fmaddss (v4sf, v4sf, v4sf) | |
8168 | v2df __builtin_ia32_fmsubpd (v2df, v2df, v2df) | |
8169 | v4sf __builtin_ia32_fmsubps (v4sf, v4sf, v4sf) | |
8170 | v2df __builtin_ia32_fmsubsd (v2df, v2df, v2df) | |
8171 | v4sf __builtin_ia32_fmsubss (v4sf, v4sf, v4sf) | |
8172 | v2df __builtin_ia32_fnmaddpd (v2df, v2df, v2df) | |
8173 | v4sf __builtin_ia32_fnmaddps (v4sf, v4sf, v4sf) | |
8174 | v2df __builtin_ia32_fnmaddsd (v2df, v2df, v2df) | |
8175 | v4sf __builtin_ia32_fnmaddss (v4sf, v4sf, v4sf) | |
8176 | v2df __builtin_ia32_fnmsubpd (v2df, v2df, v2df) | |
8177 | v4sf __builtin_ia32_fnmsubps (v4sf, v4sf, v4sf) | |
8178 | v2df __builtin_ia32_fnmsubsd (v2df, v2df, v2df) | |
8179 | v4sf __builtin_ia32_fnmsubss (v4sf, v4sf, v4sf) | |
8180 | v2df __builtin_ia32_frczpd (v2df) | |
8181 | v4sf __builtin_ia32_frczps (v4sf) | |
8182 | v2df __builtin_ia32_frczsd (v2df, v2df) | |
8183 | v4sf __builtin_ia32_frczss (v4sf, v4sf) | |
8184 | v2di __builtin_ia32_pcmov (v2di, v2di, v2di) | |
8185 | v2di __builtin_ia32_pcmov_v2di (v2di, v2di, v2di) | |
8186 | v4si __builtin_ia32_pcmov_v4si (v4si, v4si, v4si) | |
8187 | v8hi __builtin_ia32_pcmov_v8hi (v8hi, v8hi, v8hi) | |
8188 | v16qi __builtin_ia32_pcmov_v16qi (v16qi, v16qi, v16qi) | |
8189 | v2df __builtin_ia32_pcmov_v2df (v2df, v2df, v2df) | |
8190 | v4sf __builtin_ia32_pcmov_v4sf (v4sf, v4sf, v4sf) | |
8191 | v16qi __builtin_ia32_pcomeqb (v16qi, v16qi) | |
8192 | v8hi __builtin_ia32_pcomeqw (v8hi, v8hi) | |
8193 | v4si __builtin_ia32_pcomeqd (v4si, v4si) | |
8194 | v2di __builtin_ia32_pcomeqq (v2di, v2di) | |
8195 | v16qi __builtin_ia32_pcomequb (v16qi, v16qi) | |
8196 | v4si __builtin_ia32_pcomequd (v4si, v4si) | |
8197 | v2di __builtin_ia32_pcomequq (v2di, v2di) | |
8198 | v8hi __builtin_ia32_pcomequw (v8hi, v8hi) | |
8199 | v8hi __builtin_ia32_pcomeqw (v8hi, v8hi) | |
8200 | v16qi __builtin_ia32_pcomfalseb (v16qi, v16qi) | |
8201 | v4si __builtin_ia32_pcomfalsed (v4si, v4si) | |
8202 | v2di __builtin_ia32_pcomfalseq (v2di, v2di) | |
8203 | v16qi __builtin_ia32_pcomfalseub (v16qi, v16qi) | |
8204 | v4si __builtin_ia32_pcomfalseud (v4si, v4si) | |
8205 | v2di __builtin_ia32_pcomfalseuq (v2di, v2di) | |
8206 | v8hi __builtin_ia32_pcomfalseuw (v8hi, v8hi) | |
8207 | v8hi __builtin_ia32_pcomfalsew (v8hi, v8hi) | |
8208 | v16qi __builtin_ia32_pcomgeb (v16qi, v16qi) | |
8209 | v4si __builtin_ia32_pcomged (v4si, v4si) | |
8210 | v2di __builtin_ia32_pcomgeq (v2di, v2di) | |
8211 | v16qi __builtin_ia32_pcomgeub (v16qi, v16qi) | |
8212 | v4si __builtin_ia32_pcomgeud (v4si, v4si) | |
8213 | v2di __builtin_ia32_pcomgeuq (v2di, v2di) | |
8214 | v8hi __builtin_ia32_pcomgeuw (v8hi, v8hi) | |
8215 | v8hi __builtin_ia32_pcomgew (v8hi, v8hi) | |
8216 | v16qi __builtin_ia32_pcomgtb (v16qi, v16qi) | |
8217 | v4si __builtin_ia32_pcomgtd (v4si, v4si) | |
8218 | v2di __builtin_ia32_pcomgtq (v2di, v2di) | |
8219 | v16qi __builtin_ia32_pcomgtub (v16qi, v16qi) | |
8220 | v4si __builtin_ia32_pcomgtud (v4si, v4si) | |
8221 | v2di __builtin_ia32_pcomgtuq (v2di, v2di) | |
8222 | v8hi __builtin_ia32_pcomgtuw (v8hi, v8hi) | |
8223 | v8hi __builtin_ia32_pcomgtw (v8hi, v8hi) | |
8224 | v16qi __builtin_ia32_pcomleb (v16qi, v16qi) | |
8225 | v4si __builtin_ia32_pcomled (v4si, v4si) | |
8226 | v2di __builtin_ia32_pcomleq (v2di, v2di) | |
8227 | v16qi __builtin_ia32_pcomleub (v16qi, v16qi) | |
8228 | v4si __builtin_ia32_pcomleud (v4si, v4si) | |
8229 | v2di __builtin_ia32_pcomleuq (v2di, v2di) | |
8230 | v8hi __builtin_ia32_pcomleuw (v8hi, v8hi) | |
8231 | v8hi __builtin_ia32_pcomlew (v8hi, v8hi) | |
8232 | v16qi __builtin_ia32_pcomltb (v16qi, v16qi) | |
8233 | v4si __builtin_ia32_pcomltd (v4si, v4si) | |
8234 | v2di __builtin_ia32_pcomltq (v2di, v2di) | |
8235 | v16qi __builtin_ia32_pcomltub (v16qi, v16qi) | |
8236 | v4si __builtin_ia32_pcomltud (v4si, v4si) | |
8237 | v2di __builtin_ia32_pcomltuq (v2di, v2di) | |
8238 | v8hi __builtin_ia32_pcomltuw (v8hi, v8hi) | |
8239 | v8hi __builtin_ia32_pcomltw (v8hi, v8hi) | |
8240 | v16qi __builtin_ia32_pcomneb (v16qi, v16qi) | |
8241 | v4si __builtin_ia32_pcomned (v4si, v4si) | |
8242 | v2di __builtin_ia32_pcomneq (v2di, v2di) | |
8243 | v16qi __builtin_ia32_pcomneub (v16qi, v16qi) | |
8244 | v4si __builtin_ia32_pcomneud (v4si, v4si) | |
8245 | v2di __builtin_ia32_pcomneuq (v2di, v2di) | |
8246 | v8hi __builtin_ia32_pcomneuw (v8hi, v8hi) | |
8247 | v8hi __builtin_ia32_pcomnew (v8hi, v8hi) | |
8248 | v16qi __builtin_ia32_pcomtrueb (v16qi, v16qi) | |
8249 | v4si __builtin_ia32_pcomtrued (v4si, v4si) | |
8250 | v2di __builtin_ia32_pcomtrueq (v2di, v2di) | |
8251 | v16qi __builtin_ia32_pcomtrueub (v16qi, v16qi) | |
8252 | v4si __builtin_ia32_pcomtrueud (v4si, v4si) | |
8253 | v2di __builtin_ia32_pcomtrueuq (v2di, v2di) | |
8254 | v8hi __builtin_ia32_pcomtrueuw (v8hi, v8hi) | |
8255 | v8hi __builtin_ia32_pcomtruew (v8hi, v8hi) | |
8256 | v4df __builtin_ia32_permpd (v2df, v2df, v16qi) | |
8257 | v4sf __builtin_ia32_permps (v4sf, v4sf, v16qi) | |
8258 | v4si __builtin_ia32_phaddbd (v16qi) | |
8259 | v2di __builtin_ia32_phaddbq (v16qi) | |
8260 | v8hi __builtin_ia32_phaddbw (v16qi) | |
8261 | v2di __builtin_ia32_phadddq (v4si) | |
8262 | v4si __builtin_ia32_phaddubd (v16qi) | |
8263 | v2di __builtin_ia32_phaddubq (v16qi) | |
8264 | v8hi __builtin_ia32_phaddubw (v16qi) | |
8265 | v2di __builtin_ia32_phaddudq (v4si) | |
8266 | v4si __builtin_ia32_phadduwd (v8hi) | |
8267 | v2di __builtin_ia32_phadduwq (v8hi) | |
8268 | v4si __builtin_ia32_phaddwd (v8hi) | |
8269 | v2di __builtin_ia32_phaddwq (v8hi) | |
8270 | v8hi __builtin_ia32_phsubbw (v16qi) | |
8271 | v2di __builtin_ia32_phsubdq (v4si) | |
8272 | v4si __builtin_ia32_phsubwd (v8hi) | |
8273 | v4si __builtin_ia32_pmacsdd (v4si, v4si, v4si) | |
8274 | v2di __builtin_ia32_pmacsdqh (v4si, v4si, v2di) | |
8275 | v2di __builtin_ia32_pmacsdql (v4si, v4si, v2di) | |
8276 | v4si __builtin_ia32_pmacssdd (v4si, v4si, v4si) | |
8277 | v2di __builtin_ia32_pmacssdqh (v4si, v4si, v2di) | |
8278 | v2di __builtin_ia32_pmacssdql (v4si, v4si, v2di) | |
8279 | v4si __builtin_ia32_pmacsswd (v8hi, v8hi, v4si) | |
8280 | v8hi __builtin_ia32_pmacssww (v8hi, v8hi, v8hi) | |
8281 | v4si __builtin_ia32_pmacswd (v8hi, v8hi, v4si) | |
8282 | v8hi __builtin_ia32_pmacsww (v8hi, v8hi, v8hi) | |
8283 | v4si __builtin_ia32_pmadcsswd (v8hi, v8hi, v4si) | |
8284 | v4si __builtin_ia32_pmadcswd (v8hi, v8hi, v4si) | |
8285 | v16qi __builtin_ia32_pperm (v16qi, v16qi, v16qi) | |
8286 | v16qi __builtin_ia32_protb (v16qi, v16qi) | |
8287 | v4si __builtin_ia32_protd (v4si, v4si) | |
8288 | v2di __builtin_ia32_protq (v2di, v2di) | |
8289 | v8hi __builtin_ia32_protw (v8hi, v8hi) | |
8290 | v16qi __builtin_ia32_pshab (v16qi, v16qi) | |
8291 | v4si __builtin_ia32_pshad (v4si, v4si) | |
8292 | v2di __builtin_ia32_pshaq (v2di, v2di) | |
8293 | v8hi __builtin_ia32_pshaw (v8hi, v8hi) | |
8294 | v16qi __builtin_ia32_pshlb (v16qi, v16qi) | |
8295 | v4si __builtin_ia32_pshld (v4si, v4si) | |
8296 | v2di __builtin_ia32_pshlq (v2di, v2di) | |
8297 | v8hi __builtin_ia32_pshlw (v8hi, v8hi) | |
8298 | @end smallexample | |
8299 | ||
84fbffb2 | 8300 | The following builtin-in functions are available when @option{-msse5} |
04e1d06b MM |
8301 | is used. The second argument must be an integer constant and generate |
8302 | the machine instruction that is part of the name with the @samp{_imm} | |
8303 | suffix removed. | |
8304 | ||
8305 | @smallexample | |
8306 | v16qi __builtin_ia32_protb_imm (v16qi, int) | |
8307 | v4si __builtin_ia32_protd_imm (v4si, int) | |
8308 | v2di __builtin_ia32_protq_imm (v2di, int) | |
8309 | v8hi __builtin_ia32_protw_imm (v8hi, int) | |
8310 | @end smallexample | |
8311 | ||
0975678f JM |
8312 | The following built-in functions are available when @option{-m3dnow} is used. |
8313 | All of them generate the machine instruction that is part of the name. | |
8314 | ||
3ab51846 | 8315 | @smallexample |
0975678f JM |
8316 | void __builtin_ia32_femms (void) |
8317 | v8qi __builtin_ia32_pavgusb (v8qi, v8qi) | |
8318 | v2si __builtin_ia32_pf2id (v2sf) | |
8319 | v2sf __builtin_ia32_pfacc (v2sf, v2sf) | |
8320 | v2sf __builtin_ia32_pfadd (v2sf, v2sf) | |
8321 | v2si __builtin_ia32_pfcmpeq (v2sf, v2sf) | |
8322 | v2si __builtin_ia32_pfcmpge (v2sf, v2sf) | |
8323 | v2si __builtin_ia32_pfcmpgt (v2sf, v2sf) | |
8324 | v2sf __builtin_ia32_pfmax (v2sf, v2sf) | |
8325 | v2sf __builtin_ia32_pfmin (v2sf, v2sf) | |
8326 | v2sf __builtin_ia32_pfmul (v2sf, v2sf) | |
8327 | v2sf __builtin_ia32_pfrcp (v2sf) | |
8328 | v2sf __builtin_ia32_pfrcpit1 (v2sf, v2sf) | |
8329 | v2sf __builtin_ia32_pfrcpit2 (v2sf, v2sf) | |
8330 | v2sf __builtin_ia32_pfrsqrt (v2sf) | |
8331 | v2sf __builtin_ia32_pfrsqrtit1 (v2sf, v2sf) | |
8332 | v2sf __builtin_ia32_pfsub (v2sf, v2sf) | |
8333 | v2sf __builtin_ia32_pfsubr (v2sf, v2sf) | |
8334 | v2sf __builtin_ia32_pi2fd (v2si) | |
8335 | v4hi __builtin_ia32_pmulhrw (v4hi, v4hi) | |
3ab51846 | 8336 | @end smallexample |
0975678f JM |
8337 | |
8338 | The following built-in functions are available when both @option{-m3dnow} | |
8339 | and @option{-march=athlon} are used. All of them generate the machine | |
8340 | instruction that is part of the name. | |
8341 | ||
3ab51846 | 8342 | @smallexample |
0975678f JM |
8343 | v2si __builtin_ia32_pf2iw (v2sf) |
8344 | v2sf __builtin_ia32_pfnacc (v2sf, v2sf) | |
8345 | v2sf __builtin_ia32_pfpnacc (v2sf, v2sf) | |
8346 | v2sf __builtin_ia32_pi2fw (v2si) | |
8347 | v2sf __builtin_ia32_pswapdsf (v2sf) | |
8348 | v2si __builtin_ia32_pswapdsi (v2si) | |
3ab51846 | 8349 | @end smallexample |
0975678f | 8350 | |
118ea793 CF |
8351 | @node MIPS DSP Built-in Functions |
8352 | @subsection MIPS DSP Built-in Functions | |
8353 | ||
8354 | The MIPS DSP Application-Specific Extension (ASE) includes new | |
8355 | instructions that are designed to improve the performance of DSP and | |
8356 | media applications. It provides instructions that operate on packed | |
32041385 | 8357 | 8-bit/16-bit integer data, Q7, Q15 and Q31 fractional data. |
118ea793 CF |
8358 | |
8359 | GCC supports MIPS DSP operations using both the generic | |
8360 | vector extensions (@pxref{Vector Extensions}) and a collection of | |
8361 | MIPS-specific built-in functions. Both kinds of support are | |
8362 | enabled by the @option{-mdsp} command-line option. | |
8363 | ||
32041385 CF |
8364 | Revision 2 of the ASE was introduced in the second half of 2006. |
8365 | This revision adds extra instructions to the original ASE, but is | |
8366 | otherwise backwards-compatible with it. You can select revision 2 | |
8367 | using the command-line option @option{-mdspr2}; this option implies | |
8368 | @option{-mdsp}. | |
8369 | ||
118ea793 CF |
8370 | At present, GCC only provides support for operations on 32-bit |
8371 | vectors. The vector type associated with 8-bit integer data is | |
32041385 CF |
8372 | usually called @code{v4i8}, the vector type associated with Q7 |
8373 | is usually called @code{v4q7}, the vector type associated with 16-bit | |
8374 | integer data is usually called @code{v2i16}, and the vector type | |
8375 | associated with Q15 is usually called @code{v2q15}. They can be | |
8376 | defined in C as follows: | |
118ea793 CF |
8377 | |
8378 | @smallexample | |
32041385 CF |
8379 | typedef signed char v4i8 __attribute__ ((vector_size(4))); |
8380 | typedef signed char v4q7 __attribute__ ((vector_size(4))); | |
8381 | typedef short v2i16 __attribute__ ((vector_size(4))); | |
118ea793 CF |
8382 | typedef short v2q15 __attribute__ ((vector_size(4))); |
8383 | @end smallexample | |
8384 | ||
32041385 CF |
8385 | @code{v4i8}, @code{v4q7}, @code{v2i16} and @code{v2q15} values are |
8386 | initialized in the same way as aggregates. For example: | |
118ea793 CF |
8387 | |
8388 | @smallexample | |
8389 | v4i8 a = @{1, 2, 3, 4@}; | |
8390 | v4i8 b; | |
8391 | b = (v4i8) @{5, 6, 7, 8@}; | |
8392 | ||
8393 | v2q15 c = @{0x0fcb, 0x3a75@}; | |
8394 | v2q15 d; | |
8395 | d = (v2q15) @{0.1234 * 0x1.0p15, 0.4567 * 0x1.0p15@}; | |
8396 | @end smallexample | |
8397 | ||
8398 | @emph{Note:} The CPU's endianness determines the order in which values | |
8399 | are packed. On little-endian targets, the first value is the least | |
8400 | significant and the last value is the most significant. The opposite | |
8401 | order applies to big-endian targets. For example, the code above will | |
8402 | set the lowest byte of @code{a} to @code{1} on little-endian targets | |
8403 | and @code{4} on big-endian targets. | |
8404 | ||
32041385 | 8405 | @emph{Note:} Q7, Q15 and Q31 values must be initialized with their integer |
118ea793 | 8406 | representation. As shown in this example, the integer representation |
32041385 CF |
8407 | of a Q7 value can be obtained by multiplying the fractional value by |
8408 | @code{0x1.0p7}. The equivalent for Q15 values is to multiply by | |
118ea793 CF |
8409 | @code{0x1.0p15}. The equivalent for Q31 values is to multiply by |
8410 | @code{0x1.0p31}. | |
8411 | ||
8412 | The table below lists the @code{v4i8} and @code{v2q15} operations for which | |
8413 | hardware support exists. @code{a} and @code{b} are @code{v4i8} values, | |
8414 | and @code{c} and @code{d} are @code{v2q15} values. | |
8415 | ||
8416 | @multitable @columnfractions .50 .50 | |
8417 | @item C code @tab MIPS instruction | |
8418 | @item @code{a + b} @tab @code{addu.qb} | |
8419 | @item @code{c + d} @tab @code{addq.ph} | |
8420 | @item @code{a - b} @tab @code{subu.qb} | |
8421 | @item @code{c - d} @tab @code{subq.ph} | |
8422 | @end multitable | |
8423 | ||
32041385 CF |
8424 | The table below lists the @code{v2i16} operation for which |
8425 | hardware support exists for the DSP ASE REV 2. @code{e} and @code{f} are | |
8426 | @code{v2i16} values. | |
8427 | ||
8428 | @multitable @columnfractions .50 .50 | |
8429 | @item C code @tab MIPS instruction | |
8430 | @item @code{e * f} @tab @code{mul.ph} | |
8431 | @end multitable | |
8432 | ||
118ea793 CF |
8433 | It is easier to describe the DSP built-in functions if we first define |
8434 | the following types: | |
8435 | ||
8436 | @smallexample | |
8437 | typedef int q31; | |
8438 | typedef int i32; | |
32041385 | 8439 | typedef unsigned int ui32; |
118ea793 CF |
8440 | typedef long long a64; |
8441 | @end smallexample | |
8442 | ||
8443 | @code{q31} and @code{i32} are actually the same as @code{int}, but we | |
8444 | use @code{q31} to indicate a Q31 fractional value and @code{i32} to | |
8445 | indicate a 32-bit integer value. Similarly, @code{a64} is the same as | |
8446 | @code{long long}, but we use @code{a64} to indicate values that will | |
8447 | be placed in one of the four DSP accumulators (@code{$ac0}, | |
8448 | @code{$ac1}, @code{$ac2} or @code{$ac3}). | |
8449 | ||
8450 | Also, some built-in functions prefer or require immediate numbers as | |
8451 | parameters, because the corresponding DSP instructions accept both immediate | |
8452 | numbers and register operands, or accept immediate numbers only. The | |
8453 | immediate parameters are listed as follows. | |
8454 | ||
8455 | @smallexample | |
32041385 | 8456 | imm0_3: 0 to 3. |
118ea793 CF |
8457 | imm0_7: 0 to 7. |
8458 | imm0_15: 0 to 15. | |
8459 | imm0_31: 0 to 31. | |
8460 | imm0_63: 0 to 63. | |
8461 | imm0_255: 0 to 255. | |
8462 | imm_n32_31: -32 to 31. | |
8463 | imm_n512_511: -512 to 511. | |
8464 | @end smallexample | |
8465 | ||
8466 | The following built-in functions map directly to a particular MIPS DSP | |
8467 | instruction. Please refer to the architecture specification | |
8468 | for details on what each instruction does. | |
8469 | ||
8470 | @smallexample | |
8471 | v2q15 __builtin_mips_addq_ph (v2q15, v2q15) | |
8472 | v2q15 __builtin_mips_addq_s_ph (v2q15, v2q15) | |
8473 | q31 __builtin_mips_addq_s_w (q31, q31) | |
8474 | v4i8 __builtin_mips_addu_qb (v4i8, v4i8) | |
8475 | v4i8 __builtin_mips_addu_s_qb (v4i8, v4i8) | |
8476 | v2q15 __builtin_mips_subq_ph (v2q15, v2q15) | |
8477 | v2q15 __builtin_mips_subq_s_ph (v2q15, v2q15) | |
8478 | q31 __builtin_mips_subq_s_w (q31, q31) | |
8479 | v4i8 __builtin_mips_subu_qb (v4i8, v4i8) | |
8480 | v4i8 __builtin_mips_subu_s_qb (v4i8, v4i8) | |
8481 | i32 __builtin_mips_addsc (i32, i32) | |
8482 | i32 __builtin_mips_addwc (i32, i32) | |
8483 | i32 __builtin_mips_modsub (i32, i32) | |
8484 | i32 __builtin_mips_raddu_w_qb (v4i8) | |
8485 | v2q15 __builtin_mips_absq_s_ph (v2q15) | |
8486 | q31 __builtin_mips_absq_s_w (q31) | |
8487 | v4i8 __builtin_mips_precrq_qb_ph (v2q15, v2q15) | |
8488 | v2q15 __builtin_mips_precrq_ph_w (q31, q31) | |
8489 | v2q15 __builtin_mips_precrq_rs_ph_w (q31, q31) | |
8490 | v4i8 __builtin_mips_precrqu_s_qb_ph (v2q15, v2q15) | |
8491 | q31 __builtin_mips_preceq_w_phl (v2q15) | |
8492 | q31 __builtin_mips_preceq_w_phr (v2q15) | |
8493 | v2q15 __builtin_mips_precequ_ph_qbl (v4i8) | |
8494 | v2q15 __builtin_mips_precequ_ph_qbr (v4i8) | |
8495 | v2q15 __builtin_mips_precequ_ph_qbla (v4i8) | |
8496 | v2q15 __builtin_mips_precequ_ph_qbra (v4i8) | |
8497 | v2q15 __builtin_mips_preceu_ph_qbl (v4i8) | |
8498 | v2q15 __builtin_mips_preceu_ph_qbr (v4i8) | |
8499 | v2q15 __builtin_mips_preceu_ph_qbla (v4i8) | |
8500 | v2q15 __builtin_mips_preceu_ph_qbra (v4i8) | |
8501 | v4i8 __builtin_mips_shll_qb (v4i8, imm0_7) | |
8502 | v4i8 __builtin_mips_shll_qb (v4i8, i32) | |
8503 | v2q15 __builtin_mips_shll_ph (v2q15, imm0_15) | |
8504 | v2q15 __builtin_mips_shll_ph (v2q15, i32) | |
8505 | v2q15 __builtin_mips_shll_s_ph (v2q15, imm0_15) | |
8506 | v2q15 __builtin_mips_shll_s_ph (v2q15, i32) | |
8507 | q31 __builtin_mips_shll_s_w (q31, imm0_31) | |
8508 | q31 __builtin_mips_shll_s_w (q31, i32) | |
8509 | v4i8 __builtin_mips_shrl_qb (v4i8, imm0_7) | |
8510 | v4i8 __builtin_mips_shrl_qb (v4i8, i32) | |
8511 | v2q15 __builtin_mips_shra_ph (v2q15, imm0_15) | |
8512 | v2q15 __builtin_mips_shra_ph (v2q15, i32) | |
8513 | v2q15 __builtin_mips_shra_r_ph (v2q15, imm0_15) | |
8514 | v2q15 __builtin_mips_shra_r_ph (v2q15, i32) | |
8515 | q31 __builtin_mips_shra_r_w (q31, imm0_31) | |
8516 | q31 __builtin_mips_shra_r_w (q31, i32) | |
8517 | v2q15 __builtin_mips_muleu_s_ph_qbl (v4i8, v2q15) | |
8518 | v2q15 __builtin_mips_muleu_s_ph_qbr (v4i8, v2q15) | |
8519 | v2q15 __builtin_mips_mulq_rs_ph (v2q15, v2q15) | |
8520 | q31 __builtin_mips_muleq_s_w_phl (v2q15, v2q15) | |
8521 | q31 __builtin_mips_muleq_s_w_phr (v2q15, v2q15) | |
8522 | a64 __builtin_mips_dpau_h_qbl (a64, v4i8, v4i8) | |
8523 | a64 __builtin_mips_dpau_h_qbr (a64, v4i8, v4i8) | |
8524 | a64 __builtin_mips_dpsu_h_qbl (a64, v4i8, v4i8) | |
8525 | a64 __builtin_mips_dpsu_h_qbr (a64, v4i8, v4i8) | |
8526 | a64 __builtin_mips_dpaq_s_w_ph (a64, v2q15, v2q15) | |
8527 | a64 __builtin_mips_dpaq_sa_l_w (a64, q31, q31) | |
8528 | a64 __builtin_mips_dpsq_s_w_ph (a64, v2q15, v2q15) | |
8529 | a64 __builtin_mips_dpsq_sa_l_w (a64, q31, q31) | |
8530 | a64 __builtin_mips_mulsaq_s_w_ph (a64, v2q15, v2q15) | |
8531 | a64 __builtin_mips_maq_s_w_phl (a64, v2q15, v2q15) | |
8532 | a64 __builtin_mips_maq_s_w_phr (a64, v2q15, v2q15) | |
8533 | a64 __builtin_mips_maq_sa_w_phl (a64, v2q15, v2q15) | |
8534 | a64 __builtin_mips_maq_sa_w_phr (a64, v2q15, v2q15) | |
8535 | i32 __builtin_mips_bitrev (i32) | |
8536 | i32 __builtin_mips_insv (i32, i32) | |
8537 | v4i8 __builtin_mips_repl_qb (imm0_255) | |
8538 | v4i8 __builtin_mips_repl_qb (i32) | |
8539 | v2q15 __builtin_mips_repl_ph (imm_n512_511) | |
8540 | v2q15 __builtin_mips_repl_ph (i32) | |
8541 | void __builtin_mips_cmpu_eq_qb (v4i8, v4i8) | |
8542 | void __builtin_mips_cmpu_lt_qb (v4i8, v4i8) | |
8543 | void __builtin_mips_cmpu_le_qb (v4i8, v4i8) | |
8544 | i32 __builtin_mips_cmpgu_eq_qb (v4i8, v4i8) | |
8545 | i32 __builtin_mips_cmpgu_lt_qb (v4i8, v4i8) | |
8546 | i32 __builtin_mips_cmpgu_le_qb (v4i8, v4i8) | |
8547 | void __builtin_mips_cmp_eq_ph (v2q15, v2q15) | |
8548 | void __builtin_mips_cmp_lt_ph (v2q15, v2q15) | |
8549 | void __builtin_mips_cmp_le_ph (v2q15, v2q15) | |
8550 | v4i8 __builtin_mips_pick_qb (v4i8, v4i8) | |
8551 | v2q15 __builtin_mips_pick_ph (v2q15, v2q15) | |
8552 | v2q15 __builtin_mips_packrl_ph (v2q15, v2q15) | |
8553 | i32 __builtin_mips_extr_w (a64, imm0_31) | |
8554 | i32 __builtin_mips_extr_w (a64, i32) | |
8555 | i32 __builtin_mips_extr_r_w (a64, imm0_31) | |
8556 | i32 __builtin_mips_extr_s_h (a64, i32) | |
8557 | i32 __builtin_mips_extr_rs_w (a64, imm0_31) | |
8558 | i32 __builtin_mips_extr_rs_w (a64, i32) | |
8559 | i32 __builtin_mips_extr_s_h (a64, imm0_31) | |
8560 | i32 __builtin_mips_extr_r_w (a64, i32) | |
8561 | i32 __builtin_mips_extp (a64, imm0_31) | |
8562 | i32 __builtin_mips_extp (a64, i32) | |
8563 | i32 __builtin_mips_extpdp (a64, imm0_31) | |
8564 | i32 __builtin_mips_extpdp (a64, i32) | |
8565 | a64 __builtin_mips_shilo (a64, imm_n32_31) | |
8566 | a64 __builtin_mips_shilo (a64, i32) | |
8567 | a64 __builtin_mips_mthlip (a64, i32) | |
8568 | void __builtin_mips_wrdsp (i32, imm0_63) | |
8569 | i32 __builtin_mips_rddsp (imm0_63) | |
8570 | i32 __builtin_mips_lbux (void *, i32) | |
8571 | i32 __builtin_mips_lhx (void *, i32) | |
8572 | i32 __builtin_mips_lwx (void *, i32) | |
8573 | i32 __builtin_mips_bposge32 (void) | |
8574 | @end smallexample | |
8575 | ||
32041385 CF |
8576 | The following built-in functions map directly to a particular MIPS DSP REV 2 |
8577 | instruction. Please refer to the architecture specification | |
8578 | for details on what each instruction does. | |
8579 | ||
8580 | @smallexample | |
8581 | v4q7 __builtin_mips_absq_s_qb (v4q7); | |
8582 | v2i16 __builtin_mips_addu_ph (v2i16, v2i16); | |
8583 | v2i16 __builtin_mips_addu_s_ph (v2i16, v2i16); | |
8584 | v4i8 __builtin_mips_adduh_qb (v4i8, v4i8); | |
8585 | v4i8 __builtin_mips_adduh_r_qb (v4i8, v4i8); | |
8586 | i32 __builtin_mips_append (i32, i32, imm0_31); | |
8587 | i32 __builtin_mips_balign (i32, i32, imm0_3); | |
8588 | i32 __builtin_mips_cmpgdu_eq_qb (v4i8, v4i8); | |
8589 | i32 __builtin_mips_cmpgdu_lt_qb (v4i8, v4i8); | |
8590 | i32 __builtin_mips_cmpgdu_le_qb (v4i8, v4i8); | |
8591 | a64 __builtin_mips_dpa_w_ph (a64, v2i16, v2i16); | |
8592 | a64 __builtin_mips_dps_w_ph (a64, v2i16, v2i16); | |
8593 | a64 __builtin_mips_madd (a64, i32, i32); | |
8594 | a64 __builtin_mips_maddu (a64, ui32, ui32); | |
8595 | a64 __builtin_mips_msub (a64, i32, i32); | |
8596 | a64 __builtin_mips_msubu (a64, ui32, ui32); | |
8597 | v2i16 __builtin_mips_mul_ph (v2i16, v2i16); | |
8598 | v2i16 __builtin_mips_mul_s_ph (v2i16, v2i16); | |
8599 | q31 __builtin_mips_mulq_rs_w (q31, q31); | |
8600 | v2q15 __builtin_mips_mulq_s_ph (v2q15, v2q15); | |
8601 | q31 __builtin_mips_mulq_s_w (q31, q31); | |
8602 | a64 __builtin_mips_mulsa_w_ph (a64, v2i16, v2i16); | |
8603 | a64 __builtin_mips_mult (i32, i32); | |
8604 | a64 __builtin_mips_multu (ui32, ui32); | |
8605 | v4i8 __builtin_mips_precr_qb_ph (v2i16, v2i16); | |
8606 | v2i16 __builtin_mips_precr_sra_ph_w (i32, i32, imm0_31); | |
8607 | v2i16 __builtin_mips_precr_sra_r_ph_w (i32, i32, imm0_31); | |
8608 | i32 __builtin_mips_prepend (i32, i32, imm0_31); | |
8609 | v4i8 __builtin_mips_shra_qb (v4i8, imm0_7); | |
8610 | v4i8 __builtin_mips_shra_r_qb (v4i8, imm0_7); | |
8611 | v4i8 __builtin_mips_shra_qb (v4i8, i32); | |
8612 | v4i8 __builtin_mips_shra_r_qb (v4i8, i32); | |
8613 | v2i16 __builtin_mips_shrl_ph (v2i16, imm0_15); | |
8614 | v2i16 __builtin_mips_shrl_ph (v2i16, i32); | |
8615 | v2i16 __builtin_mips_subu_ph (v2i16, v2i16); | |
8616 | v2i16 __builtin_mips_subu_s_ph (v2i16, v2i16); | |
8617 | v4i8 __builtin_mips_subuh_qb (v4i8, v4i8); | |
8618 | v4i8 __builtin_mips_subuh_r_qb (v4i8, v4i8); | |
8619 | v2q15 __builtin_mips_addqh_ph (v2q15, v2q15); | |
8620 | v2q15 __builtin_mips_addqh_r_ph (v2q15, v2q15); | |
8621 | q31 __builtin_mips_addqh_w (q31, q31); | |
8622 | q31 __builtin_mips_addqh_r_w (q31, q31); | |
8623 | v2q15 __builtin_mips_subqh_ph (v2q15, v2q15); | |
8624 | v2q15 __builtin_mips_subqh_r_ph (v2q15, v2q15); | |
8625 | q31 __builtin_mips_subqh_w (q31, q31); | |
8626 | q31 __builtin_mips_subqh_r_w (q31, q31); | |
8627 | a64 __builtin_mips_dpax_w_ph (a64, v2i16, v2i16); | |
8628 | a64 __builtin_mips_dpsx_w_ph (a64, v2i16, v2i16); | |
8629 | a64 __builtin_mips_dpaqx_s_w_ph (a64, v2q15, v2q15); | |
8630 | a64 __builtin_mips_dpaqx_sa_w_ph (a64, v2q15, v2q15); | |
8631 | a64 __builtin_mips_dpsqx_s_w_ph (a64, v2q15, v2q15); | |
8632 | a64 __builtin_mips_dpsqx_sa_w_ph (a64, v2q15, v2q15); | |
8633 | @end smallexample | |
8634 | ||
8635 | ||
d840bfd3 CF |
8636 | @node MIPS Paired-Single Support |
8637 | @subsection MIPS Paired-Single Support | |
8638 | ||
8639 | The MIPS64 architecture includes a number of instructions that | |
8640 | operate on pairs of single-precision floating-point values. | |
8641 | Each pair is packed into a 64-bit floating-point register, | |
8642 | with one element being designated the ``upper half'' and | |
8643 | the other being designated the ``lower half''. | |
8644 | ||
8645 | GCC supports paired-single operations using both the generic | |
8646 | vector extensions (@pxref{Vector Extensions}) and a collection of | |
8647 | MIPS-specific built-in functions. Both kinds of support are | |
8648 | enabled by the @option{-mpaired-single} command-line option. | |
8649 | ||
8650 | The vector type associated with paired-single values is usually | |
8651 | called @code{v2sf}. It can be defined in C as follows: | |
8652 | ||
8653 | @smallexample | |
8654 | typedef float v2sf __attribute__ ((vector_size (8))); | |
8655 | @end smallexample | |
8656 | ||
8657 | @code{v2sf} values are initialized in the same way as aggregates. | |
8658 | For example: | |
8659 | ||
8660 | @smallexample | |
8661 | v2sf a = @{1.5, 9.1@}; | |
8662 | v2sf b; | |
8663 | float e, f; | |
8664 | b = (v2sf) @{e, f@}; | |
8665 | @end smallexample | |
8666 | ||
8667 | @emph{Note:} The CPU's endianness determines which value is stored in | |
8668 | the upper half of a register and which value is stored in the lower half. | |
8669 | On little-endian targets, the first value is the lower one and the second | |
8670 | value is the upper one. The opposite order applies to big-endian targets. | |
8671 | For example, the code above will set the lower half of @code{a} to | |
8672 | @code{1.5} on little-endian targets and @code{9.1} on big-endian targets. | |
8673 | ||
93581857 MS |
8674 | @node MIPS Loongson Built-in Functions |
8675 | @subsection MIPS Loongson Built-in Functions | |
8676 | ||
8677 | GCC provides intrinsics to access the SIMD instructions provided by the | |
8678 | ST Microelectronics Loongson-2E and -2F processors. These intrinsics, | |
8679 | available after inclusion of the @code{loongson.h} header file, | |
8680 | operate on the following 64-bit vector types: | |
8681 | ||
8682 | @itemize | |
8683 | @item @code{uint8x8_t}, a vector of eight unsigned 8-bit integers; | |
8684 | @item @code{uint16x4_t}, a vector of four unsigned 16-bit integers; | |
8685 | @item @code{uint32x2_t}, a vector of two unsigned 32-bit integers; | |
8686 | @item @code{int8x8_t}, a vector of eight signed 8-bit integers; | |
8687 | @item @code{int16x4_t}, a vector of four signed 16-bit integers; | |
8688 | @item @code{int32x2_t}, a vector of two signed 32-bit integers. | |
8689 | @end itemize | |
8690 | ||
8691 | The intrinsics provided are listed below; each is named after the | |
8692 | machine instruction to which it corresponds, with suffixes added as | |
8693 | appropriate to distinguish intrinsics that expand to the same machine | |
8694 | instruction yet have different argument types. Refer to the architecture | |
8695 | documentation for a description of the functionality of each | |
8696 | instruction. | |
8697 | ||
8698 | @smallexample | |
8699 | int16x4_t packsswh (int32x2_t s, int32x2_t t); | |
8700 | int8x8_t packsshb (int16x4_t s, int16x4_t t); | |
8701 | uint8x8_t packushb (uint16x4_t s, uint16x4_t t); | |
8702 | uint32x2_t paddw_u (uint32x2_t s, uint32x2_t t); | |
8703 | uint16x4_t paddh_u (uint16x4_t s, uint16x4_t t); | |
8704 | uint8x8_t paddb_u (uint8x8_t s, uint8x8_t t); | |
8705 | int32x2_t paddw_s (int32x2_t s, int32x2_t t); | |
8706 | int16x4_t paddh_s (int16x4_t s, int16x4_t t); | |
8707 | int8x8_t paddb_s (int8x8_t s, int8x8_t t); | |
8708 | uint64_t paddd_u (uint64_t s, uint64_t t); | |
8709 | int64_t paddd_s (int64_t s, int64_t t); | |
8710 | int16x4_t paddsh (int16x4_t s, int16x4_t t); | |
8711 | int8x8_t paddsb (int8x8_t s, int8x8_t t); | |
8712 | uint16x4_t paddush (uint16x4_t s, uint16x4_t t); | |
8713 | uint8x8_t paddusb (uint8x8_t s, uint8x8_t t); | |
8714 | uint64_t pandn_ud (uint64_t s, uint64_t t); | |
8715 | uint32x2_t pandn_uw (uint32x2_t s, uint32x2_t t); | |
8716 | uint16x4_t pandn_uh (uint16x4_t s, uint16x4_t t); | |
8717 | uint8x8_t pandn_ub (uint8x8_t s, uint8x8_t t); | |
8718 | int64_t pandn_sd (int64_t s, int64_t t); | |
8719 | int32x2_t pandn_sw (int32x2_t s, int32x2_t t); | |
8720 | int16x4_t pandn_sh (int16x4_t s, int16x4_t t); | |
8721 | int8x8_t pandn_sb (int8x8_t s, int8x8_t t); | |
8722 | uint16x4_t pavgh (uint16x4_t s, uint16x4_t t); | |
8723 | uint8x8_t pavgb (uint8x8_t s, uint8x8_t t); | |
8724 | uint32x2_t pcmpeqw_u (uint32x2_t s, uint32x2_t t); | |
8725 | uint16x4_t pcmpeqh_u (uint16x4_t s, uint16x4_t t); | |
8726 | uint8x8_t pcmpeqb_u (uint8x8_t s, uint8x8_t t); | |
8727 | int32x2_t pcmpeqw_s (int32x2_t s, int32x2_t t); | |
8728 | int16x4_t pcmpeqh_s (int16x4_t s, int16x4_t t); | |
8729 | int8x8_t pcmpeqb_s (int8x8_t s, int8x8_t t); | |
8730 | uint32x2_t pcmpgtw_u (uint32x2_t s, uint32x2_t t); | |
8731 | uint16x4_t pcmpgth_u (uint16x4_t s, uint16x4_t t); | |
8732 | uint8x8_t pcmpgtb_u (uint8x8_t s, uint8x8_t t); | |
8733 | int32x2_t pcmpgtw_s (int32x2_t s, int32x2_t t); | |
8734 | int16x4_t pcmpgth_s (int16x4_t s, int16x4_t t); | |
8735 | int8x8_t pcmpgtb_s (int8x8_t s, int8x8_t t); | |
8736 | uint16x4_t pextrh_u (uint16x4_t s, int field); | |
8737 | int16x4_t pextrh_s (int16x4_t s, int field); | |
8738 | uint16x4_t pinsrh_0_u (uint16x4_t s, uint16x4_t t); | |
8739 | uint16x4_t pinsrh_1_u (uint16x4_t s, uint16x4_t t); | |
8740 | uint16x4_t pinsrh_2_u (uint16x4_t s, uint16x4_t t); | |
8741 | uint16x4_t pinsrh_3_u (uint16x4_t s, uint16x4_t t); | |
8742 | int16x4_t pinsrh_0_s (int16x4_t s, int16x4_t t); | |
8743 | int16x4_t pinsrh_1_s (int16x4_t s, int16x4_t t); | |
8744 | int16x4_t pinsrh_2_s (int16x4_t s, int16x4_t t); | |
8745 | int16x4_t pinsrh_3_s (int16x4_t s, int16x4_t t); | |
8746 | int32x2_t pmaddhw (int16x4_t s, int16x4_t t); | |
8747 | int16x4_t pmaxsh (int16x4_t s, int16x4_t t); | |
8748 | uint8x8_t pmaxub (uint8x8_t s, uint8x8_t t); | |
8749 | int16x4_t pminsh (int16x4_t s, int16x4_t t); | |
8750 | uint8x8_t pminub (uint8x8_t s, uint8x8_t t); | |
8751 | uint8x8_t pmovmskb_u (uint8x8_t s); | |
8752 | int8x8_t pmovmskb_s (int8x8_t s); | |
8753 | uint16x4_t pmulhuh (uint16x4_t s, uint16x4_t t); | |
8754 | int16x4_t pmulhh (int16x4_t s, int16x4_t t); | |
8755 | int16x4_t pmullh (int16x4_t s, int16x4_t t); | |
8756 | int64_t pmuluw (uint32x2_t s, uint32x2_t t); | |
8757 | uint8x8_t pasubub (uint8x8_t s, uint8x8_t t); | |
8758 | uint16x4_t biadd (uint8x8_t s); | |
8759 | uint16x4_t psadbh (uint8x8_t s, uint8x8_t t); | |
8760 | uint16x4_t pshufh_u (uint16x4_t dest, uint16x4_t s, uint8_t order); | |
8761 | int16x4_t pshufh_s (int16x4_t dest, int16x4_t s, uint8_t order); | |
8762 | uint16x4_t psllh_u (uint16x4_t s, uint8_t amount); | |
8763 | int16x4_t psllh_s (int16x4_t s, uint8_t amount); | |
8764 | uint32x2_t psllw_u (uint32x2_t s, uint8_t amount); | |
8765 | int32x2_t psllw_s (int32x2_t s, uint8_t amount); | |
8766 | uint16x4_t psrlh_u (uint16x4_t s, uint8_t amount); | |
8767 | int16x4_t psrlh_s (int16x4_t s, uint8_t amount); | |
8768 | uint32x2_t psrlw_u (uint32x2_t s, uint8_t amount); | |
8769 | int32x2_t psrlw_s (int32x2_t s, uint8_t amount); | |
8770 | uint16x4_t psrah_u (uint16x4_t s, uint8_t amount); | |
8771 | int16x4_t psrah_s (int16x4_t s, uint8_t amount); | |
8772 | uint32x2_t psraw_u (uint32x2_t s, uint8_t amount); | |
8773 | int32x2_t psraw_s (int32x2_t s, uint8_t amount); | |
8774 | uint32x2_t psubw_u (uint32x2_t s, uint32x2_t t); | |
8775 | uint16x4_t psubh_u (uint16x4_t s, uint16x4_t t); | |
8776 | uint8x8_t psubb_u (uint8x8_t s, uint8x8_t t); | |
8777 | int32x2_t psubw_s (int32x2_t s, int32x2_t t); | |
8778 | int16x4_t psubh_s (int16x4_t s, int16x4_t t); | |
8779 | int8x8_t psubb_s (int8x8_t s, int8x8_t t); | |
8780 | uint64_t psubd_u (uint64_t s, uint64_t t); | |
8781 | int64_t psubd_s (int64_t s, int64_t t); | |
8782 | int16x4_t psubsh (int16x4_t s, int16x4_t t); | |
8783 | int8x8_t psubsb (int8x8_t s, int8x8_t t); | |
8784 | uint16x4_t psubush (uint16x4_t s, uint16x4_t t); | |
8785 | uint8x8_t psubusb (uint8x8_t s, uint8x8_t t); | |
8786 | uint32x2_t punpckhwd_u (uint32x2_t s, uint32x2_t t); | |
8787 | uint16x4_t punpckhhw_u (uint16x4_t s, uint16x4_t t); | |
8788 | uint8x8_t punpckhbh_u (uint8x8_t s, uint8x8_t t); | |
8789 | int32x2_t punpckhwd_s (int32x2_t s, int32x2_t t); | |
8790 | int16x4_t punpckhhw_s (int16x4_t s, int16x4_t t); | |
8791 | int8x8_t punpckhbh_s (int8x8_t s, int8x8_t t); | |
8792 | uint32x2_t punpcklwd_u (uint32x2_t s, uint32x2_t t); | |
8793 | uint16x4_t punpcklhw_u (uint16x4_t s, uint16x4_t t); | |
8794 | uint8x8_t punpcklbh_u (uint8x8_t s, uint8x8_t t); | |
8795 | int32x2_t punpcklwd_s (int32x2_t s, int32x2_t t); | |
8796 | int16x4_t punpcklhw_s (int16x4_t s, int16x4_t t); | |
8797 | int8x8_t punpcklbh_s (int8x8_t s, int8x8_t t); | |
8798 | @end smallexample | |
8799 | ||
d840bfd3 CF |
8800 | @menu |
8801 | * Paired-Single Arithmetic:: | |
8802 | * Paired-Single Built-in Functions:: | |
8803 | * MIPS-3D Built-in Functions:: | |
8804 | @end menu | |
8805 | ||
8806 | @node Paired-Single Arithmetic | |
8807 | @subsubsection Paired-Single Arithmetic | |
8808 | ||
8809 | The table below lists the @code{v2sf} operations for which hardware | |
8810 | support exists. @code{a}, @code{b} and @code{c} are @code{v2sf} | |
8811 | values and @code{x} is an integral value. | |
8812 | ||
8813 | @multitable @columnfractions .50 .50 | |
8814 | @item C code @tab MIPS instruction | |
8815 | @item @code{a + b} @tab @code{add.ps} | |
8816 | @item @code{a - b} @tab @code{sub.ps} | |
8817 | @item @code{-a} @tab @code{neg.ps} | |
8818 | @item @code{a * b} @tab @code{mul.ps} | |
8819 | @item @code{a * b + c} @tab @code{madd.ps} | |
8820 | @item @code{a * b - c} @tab @code{msub.ps} | |
8821 | @item @code{-(a * b + c)} @tab @code{nmadd.ps} | |
8822 | @item @code{-(a * b - c)} @tab @code{nmsub.ps} | |
8823 | @item @code{x ? a : b} @tab @code{movn.ps}/@code{movz.ps} | |
8824 | @end multitable | |
8825 | ||
8826 | Note that the multiply-accumulate instructions can be disabled | |
8827 | using the command-line option @code{-mno-fused-madd}. | |
8828 | ||
8829 | @node Paired-Single Built-in Functions | |
8830 | @subsubsection Paired-Single Built-in Functions | |
8831 | ||
8832 | The following paired-single functions map directly to a particular | |
8833 | MIPS instruction. Please refer to the architecture specification | |
8834 | for details on what each instruction does. | |
8835 | ||
8836 | @table @code | |
8837 | @item v2sf __builtin_mips_pll_ps (v2sf, v2sf) | |
8838 | Pair lower lower (@code{pll.ps}). | |
8839 | ||
8840 | @item v2sf __builtin_mips_pul_ps (v2sf, v2sf) | |
8841 | Pair upper lower (@code{pul.ps}). | |
8842 | ||
8843 | @item v2sf __builtin_mips_plu_ps (v2sf, v2sf) | |
8844 | Pair lower upper (@code{plu.ps}). | |
8845 | ||
8846 | @item v2sf __builtin_mips_puu_ps (v2sf, v2sf) | |
8847 | Pair upper upper (@code{puu.ps}). | |
8848 | ||
8849 | @item v2sf __builtin_mips_cvt_ps_s (float, float) | |
8850 | Convert pair to paired single (@code{cvt.ps.s}). | |
8851 | ||
8852 | @item float __builtin_mips_cvt_s_pl (v2sf) | |
8853 | Convert pair lower to single (@code{cvt.s.pl}). | |
8854 | ||
8855 | @item float __builtin_mips_cvt_s_pu (v2sf) | |
8856 | Convert pair upper to single (@code{cvt.s.pu}). | |
8857 | ||
8858 | @item v2sf __builtin_mips_abs_ps (v2sf) | |
8859 | Absolute value (@code{abs.ps}). | |
8860 | ||
8861 | @item v2sf __builtin_mips_alnv_ps (v2sf, v2sf, int) | |
8862 | Align variable (@code{alnv.ps}). | |
8863 | ||
8864 | @emph{Note:} The value of the third parameter must be 0 or 4 | |
8865 | modulo 8, otherwise the result will be unpredictable. Please read the | |
8866 | instruction description for details. | |
8867 | @end table | |
8868 | ||
8869 | The following multi-instruction functions are also available. | |
8870 | In each case, @var{cond} can be any of the 16 floating-point conditions: | |
8871 | @code{f}, @code{un}, @code{eq}, @code{ueq}, @code{olt}, @code{ult}, | |
8872 | @code{ole}, @code{ule}, @code{sf}, @code{ngle}, @code{seq}, @code{ngl}, | |
8873 | @code{lt}, @code{nge}, @code{le} or @code{ngt}. | |
8874 | ||
8875 | @table @code | |
8876 | @item v2sf __builtin_mips_movt_c_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}, v2sf @var{c}, v2sf @var{d}) | |
8877 | @itemx v2sf __builtin_mips_movf_c_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}, v2sf @var{c}, v2sf @var{d}) | |
8878 | Conditional move based on floating point comparison (@code{c.@var{cond}.ps}, | |
8879 | @code{movt.ps}/@code{movf.ps}). | |
8880 | ||
8881 | The @code{movt} functions return the value @var{x} computed by: | |
8882 | ||
8883 | @smallexample | |
8884 | c.@var{cond}.ps @var{cc},@var{a},@var{b} | |
8885 | mov.ps @var{x},@var{c} | |
8886 | movt.ps @var{x},@var{d},@var{cc} | |
8887 | @end smallexample | |
8888 | ||
8889 | The @code{movf} functions are similar but use @code{movf.ps} instead | |
8890 | of @code{movt.ps}. | |
8891 | ||
8892 | @item int __builtin_mips_upper_c_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}) | |
8893 | @itemx int __builtin_mips_lower_c_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}) | |
8894 | Comparison of two paired-single values (@code{c.@var{cond}.ps}, | |
8895 | @code{bc1t}/@code{bc1f}). | |
8896 | ||
8897 | These functions compare @var{a} and @var{b} using @code{c.@var{cond}.ps} | |
8898 | and return either the upper or lower half of the result. For example: | |
8899 | ||
8900 | @smallexample | |
8901 | v2sf a, b; | |
8902 | if (__builtin_mips_upper_c_eq_ps (a, b)) | |
8903 | upper_halves_are_equal (); | |
8904 | else | |
8905 | upper_halves_are_unequal (); | |
8906 | ||
8907 | if (__builtin_mips_lower_c_eq_ps (a, b)) | |
8908 | lower_halves_are_equal (); | |
8909 | else | |
8910 | lower_halves_are_unequal (); | |
8911 | @end smallexample | |
8912 | @end table | |
8913 | ||
8914 | @node MIPS-3D Built-in Functions | |
8915 | @subsubsection MIPS-3D Built-in Functions | |
8916 | ||
8917 | The MIPS-3D Application-Specific Extension (ASE) includes additional | |
8918 | paired-single instructions that are designed to improve the performance | |
8919 | of 3D graphics operations. Support for these instructions is controlled | |
8920 | by the @option{-mips3d} command-line option. | |
8921 | ||
8922 | The functions listed below map directly to a particular MIPS-3D | |
8923 | instruction. Please refer to the architecture specification for | |
8924 | more details on what each instruction does. | |
8925 | ||
8926 | @table @code | |
8927 | @item v2sf __builtin_mips_addr_ps (v2sf, v2sf) | |
8928 | Reduction add (@code{addr.ps}). | |
8929 | ||
8930 | @item v2sf __builtin_mips_mulr_ps (v2sf, v2sf) | |
8931 | Reduction multiply (@code{mulr.ps}). | |
8932 | ||
8933 | @item v2sf __builtin_mips_cvt_pw_ps (v2sf) | |
8934 | Convert paired single to paired word (@code{cvt.pw.ps}). | |
8935 | ||
8936 | @item v2sf __builtin_mips_cvt_ps_pw (v2sf) | |
8937 | Convert paired word to paired single (@code{cvt.ps.pw}). | |
8938 | ||
8939 | @item float __builtin_mips_recip1_s (float) | |
8940 | @itemx double __builtin_mips_recip1_d (double) | |
8941 | @itemx v2sf __builtin_mips_recip1_ps (v2sf) | |
8942 | Reduced precision reciprocal (sequence step 1) (@code{recip1.@var{fmt}}). | |
8943 | ||
8944 | @item float __builtin_mips_recip2_s (float, float) | |
8945 | @itemx double __builtin_mips_recip2_d (double, double) | |
8946 | @itemx v2sf __builtin_mips_recip2_ps (v2sf, v2sf) | |
8947 | Reduced precision reciprocal (sequence step 2) (@code{recip2.@var{fmt}}). | |
8948 | ||
8949 | @item float __builtin_mips_rsqrt1_s (float) | |
8950 | @itemx double __builtin_mips_rsqrt1_d (double) | |
8951 | @itemx v2sf __builtin_mips_rsqrt1_ps (v2sf) | |
8952 | Reduced precision reciprocal square root (sequence step 1) | |
8953 | (@code{rsqrt1.@var{fmt}}). | |
8954 | ||
8955 | @item float __builtin_mips_rsqrt2_s (float, float) | |
8956 | @itemx double __builtin_mips_rsqrt2_d (double, double) | |
8957 | @itemx v2sf __builtin_mips_rsqrt2_ps (v2sf, v2sf) | |
8958 | Reduced precision reciprocal square root (sequence step 2) | |
8959 | (@code{rsqrt2.@var{fmt}}). | |
8960 | @end table | |
8961 | ||
8962 | The following multi-instruction functions are also available. | |
8963 | In each case, @var{cond} can be any of the 16 floating-point conditions: | |
8964 | @code{f}, @code{un}, @code{eq}, @code{ueq}, @code{olt}, @code{ult}, | |
8965 | @code{ole}, @code{ule}, @code{sf}, @code{ngle}, @code{seq}, | |
8966 | @code{ngl}, @code{lt}, @code{nge}, @code{le} or @code{ngt}. | |
8967 | ||
8968 | @table @code | |
8969 | @item int __builtin_mips_cabs_@var{cond}_s (float @var{a}, float @var{b}) | |
8970 | @itemx int __builtin_mips_cabs_@var{cond}_d (double @var{a}, double @var{b}) | |
8971 | Absolute comparison of two scalar values (@code{cabs.@var{cond}.@var{fmt}}, | |
8972 | @code{bc1t}/@code{bc1f}). | |
8973 | ||
8974 | These functions compare @var{a} and @var{b} using @code{cabs.@var{cond}.s} | |
8975 | or @code{cabs.@var{cond}.d} and return the result as a boolean value. | |
8976 | For example: | |
8977 | ||
8978 | @smallexample | |
8979 | float a, b; | |
8980 | if (__builtin_mips_cabs_eq_s (a, b)) | |
8981 | true (); | |
8982 | else | |
8983 | false (); | |
8984 | @end smallexample | |
8985 | ||
8986 | @item int __builtin_mips_upper_cabs_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}) | |
8987 | @itemx int __builtin_mips_lower_cabs_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}) | |
8988 | Absolute comparison of two paired-single values (@code{cabs.@var{cond}.ps}, | |
8989 | @code{bc1t}/@code{bc1f}). | |
8990 | ||
8991 | These functions compare @var{a} and @var{b} using @code{cabs.@var{cond}.ps} | |
8992 | and return either the upper or lower half of the result. For example: | |
8993 | ||
8994 | @smallexample | |
8995 | v2sf a, b; | |
8996 | if (__builtin_mips_upper_cabs_eq_ps (a, b)) | |
8997 | upper_halves_are_equal (); | |
8998 | else | |
8999 | upper_halves_are_unequal (); | |
9000 | ||
9001 | if (__builtin_mips_lower_cabs_eq_ps (a, b)) | |
9002 | lower_halves_are_equal (); | |
9003 | else | |
9004 | lower_halves_are_unequal (); | |
9005 | @end smallexample | |
9006 | ||
9007 | @item v2sf __builtin_mips_movt_cabs_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}, v2sf @var{c}, v2sf @var{d}) | |
9008 | @itemx v2sf __builtin_mips_movf_cabs_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}, v2sf @var{c}, v2sf @var{d}) | |
9009 | Conditional move based on absolute comparison (@code{cabs.@var{cond}.ps}, | |
9010 | @code{movt.ps}/@code{movf.ps}). | |
9011 | ||
9012 | The @code{movt} functions return the value @var{x} computed by: | |
9013 | ||
9014 | @smallexample | |
9015 | cabs.@var{cond}.ps @var{cc},@var{a},@var{b} | |
9016 | mov.ps @var{x},@var{c} | |
9017 | movt.ps @var{x},@var{d},@var{cc} | |
9018 | @end smallexample | |
9019 | ||
9020 | The @code{movf} functions are similar but use @code{movf.ps} instead | |
9021 | of @code{movt.ps}. | |
9022 | ||
9023 | @item int __builtin_mips_any_c_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}) | |
9024 | @itemx int __builtin_mips_all_c_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}) | |
9025 | @itemx int __builtin_mips_any_cabs_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}) | |
9026 | @itemx int __builtin_mips_all_cabs_@var{cond}_ps (v2sf @var{a}, v2sf @var{b}) | |
9027 | Comparison of two paired-single values | |
9028 | (@code{c.@var{cond}.ps}/@code{cabs.@var{cond}.ps}, | |
9029 | @code{bc1any2t}/@code{bc1any2f}). | |
9030 | ||
9031 | These functions compare @var{a} and @var{b} using @code{c.@var{cond}.ps} | |
9032 | or @code{cabs.@var{cond}.ps}. The @code{any} forms return true if either | |
9033 | result is true and the @code{all} forms return true if both results are true. | |
9034 | For example: | |
9035 | ||
9036 | @smallexample | |
9037 | v2sf a, b; | |
9038 | if (__builtin_mips_any_c_eq_ps (a, b)) | |
9039 | one_is_true (); | |
9040 | else | |
9041 | both_are_false (); | |
9042 | ||
9043 | if (__builtin_mips_all_c_eq_ps (a, b)) | |
9044 | both_are_true (); | |
9045 | else | |
9046 | one_is_false (); | |
9047 | @end smallexample | |
9048 | ||
9049 | @item int __builtin_mips_any_c_@var{cond}_4s (v2sf @var{a}, v2sf @var{b}, v2sf @var{c}, v2sf @var{d}) | |
9050 | @itemx int __builtin_mips_all_c_@var{cond}_4s (v2sf @var{a}, v2sf @var{b}, v2sf @var{c}, v2sf @var{d}) | |
9051 | @itemx int __builtin_mips_any_cabs_@var{cond}_4s (v2sf @var{a}, v2sf @var{b}, v2sf @var{c}, v2sf @var{d}) | |
9052 | @itemx int __builtin_mips_all_cabs_@var{cond}_4s (v2sf @var{a}, v2sf @var{b}, v2sf @var{c}, v2sf @var{d}) | |
9053 | Comparison of four paired-single values | |
9054 | (@code{c.@var{cond}.ps}/@code{cabs.@var{cond}.ps}, | |
9055 | @code{bc1any4t}/@code{bc1any4f}). | |
9056 | ||
9057 | These functions use @code{c.@var{cond}.ps} or @code{cabs.@var{cond}.ps} | |
9058 | to compare @var{a} with @var{b} and to compare @var{c} with @var{d}. | |
9059 | The @code{any} forms return true if any of the four results are true | |
9060 | and the @code{all} forms return true if all four results are true. | |
9061 | For example: | |
9062 | ||
9063 | @smallexample | |
9064 | v2sf a, b, c, d; | |
9065 | if (__builtin_mips_any_c_eq_4s (a, b, c, d)) | |
9066 | some_are_true (); | |
9067 | else | |
9068 | all_are_false (); | |
9069 | ||
9070 | if (__builtin_mips_all_c_eq_4s (a, b, c, d)) | |
9071 | all_are_true (); | |
9072 | else | |
9073 | some_are_false (); | |
9074 | @end smallexample | |
9075 | @end table | |
9076 | ||
333c8841 AH |
9077 | @node PowerPC AltiVec Built-in Functions |
9078 | @subsection PowerPC AltiVec Built-in Functions | |
9079 | ||
b0b343db JJ |
9080 | GCC provides an interface for the PowerPC family of processors to access |
9081 | the AltiVec operations described in Motorola's AltiVec Programming | |
9082 | Interface Manual. The interface is made available by including | |
9083 | @code{<altivec.h>} and using @option{-maltivec} and | |
9084 | @option{-mabi=altivec}. The interface supports the following vector | |
9085 | types. | |
333c8841 | 9086 | |
b0b343db JJ |
9087 | @smallexample |
9088 | vector unsigned char | |
9089 | vector signed char | |
9090 | vector bool char | |
333c8841 | 9091 | |
b0b343db JJ |
9092 | vector unsigned short |
9093 | vector signed short | |
9094 | vector bool short | |
9095 | vector pixel | |
9096 | ||
9097 | vector unsigned int | |
9098 | vector signed int | |
9099 | vector bool int | |
9100 | vector float | |
9101 | @end smallexample | |
9102 | ||
9103 | GCC's implementation of the high-level language interface available from | |
9104 | C and C++ code differs from Motorola's documentation in several ways. | |
9105 | ||
9106 | @itemize @bullet | |
9107 | ||
9108 | @item | |
9109 | A vector constant is a list of constant expressions within curly braces. | |
9110 | ||
9111 | @item | |
9112 | A vector initializer requires no cast if the vector constant is of the | |
9113 | same type as the variable it is initializing. | |
333c8841 | 9114 | |
b0b343db | 9115 | @item |
5edea4c6 JJ |
9116 | If @code{signed} or @code{unsigned} is omitted, the signedness of the |
9117 | vector type is the default signedness of the base type. The default | |
9118 | varies depending on the operating system, so a portable program should | |
9119 | always specify the signedness. | |
4e6e4e4c JJ |
9120 | |
9121 | @item | |
9122 | Compiling with @option{-maltivec} adds keywords @code{__vector}, | |
5950c3c9 BE |
9123 | @code{vector}, @code{__pixel}, @code{pixel}, @code{__bool} and |
9124 | @code{bool}. When compiling ISO C, the context-sensitive substitution | |
9125 | of the keywords @code{vector}, @code{pixel} and @code{bool} is | |
9126 | disabled. To use them, you must include @code{<altivec.h>} instead. | |
4e6e4e4c JJ |
9127 | |
9128 | @item | |
9129 | GCC allows using a @code{typedef} name as the type specifier for a | |
9130 | vector type. | |
b0b343db JJ |
9131 | |
9132 | @item | |
9133 | For C, overloaded functions are implemented with macros so the following | |
9134 | does not work: | |
90989b26 AH |
9135 | |
9136 | @smallexample | |
8254cb45 | 9137 | vec_add ((vector signed int)@{1, 2, 3, 4@}, foo); |
90989b26 AH |
9138 | @end smallexample |
9139 | ||
b0b343db JJ |
9140 | Since @code{vec_add} is a macro, the vector constant in the example |
9141 | is treated as four separate arguments. Wrap the entire argument in | |
9142 | parentheses for this to work. | |
9143 | @end itemize | |
90989b26 | 9144 | |
ae4b4a02 AH |
9145 | @emph{Note:} Only the @code{<altivec.h>} interface is supported. |
9146 | Internally, GCC uses built-in functions to achieve the functionality in | |
9147 | the aforementioned header file, but they are not supported and are | |
9148 | subject to change without notice. | |
9149 | ||
b0b343db JJ |
9150 | The following interfaces are supported for the generic and specific |
9151 | AltiVec operations and the AltiVec predicates. In cases where there | |
9152 | is a direct mapping between generic and specific operations, only the | |
9153 | generic names are shown here, although the specific operations can also | |
9154 | be used. | |
333c8841 | 9155 | |
b0b343db JJ |
9156 | Arguments that are documented as @code{const int} require literal |
9157 | integral values within the range required for that operation. | |
333c8841 | 9158 | |
b0b343db JJ |
9159 | @smallexample |
9160 | vector signed char vec_abs (vector signed char); | |
9161 | vector signed short vec_abs (vector signed short); | |
9162 | vector signed int vec_abs (vector signed int); | |
9163 | vector float vec_abs (vector float); | |
333c8841 | 9164 | |
b0b343db JJ |
9165 | vector signed char vec_abss (vector signed char); |
9166 | vector signed short vec_abss (vector signed short); | |
9167 | vector signed int vec_abss (vector signed int); | |
333c8841 | 9168 | |
b0b343db JJ |
9169 | vector signed char vec_add (vector bool char, vector signed char); |
9170 | vector signed char vec_add (vector signed char, vector bool char); | |
9171 | vector signed char vec_add (vector signed char, vector signed char); | |
9172 | vector unsigned char vec_add (vector bool char, vector unsigned char); | |
9173 | vector unsigned char vec_add (vector unsigned char, vector bool char); | |
924fcc4e JM |
9174 | vector unsigned char vec_add (vector unsigned char, |
9175 | vector unsigned char); | |
b0b343db JJ |
9176 | vector signed short vec_add (vector bool short, vector signed short); |
9177 | vector signed short vec_add (vector signed short, vector bool short); | |
333c8841 | 9178 | vector signed short vec_add (vector signed short, vector signed short); |
b0b343db | 9179 | vector unsigned short vec_add (vector bool short, |
924fcc4e JM |
9180 | vector unsigned short); |
9181 | vector unsigned short vec_add (vector unsigned short, | |
b0b343db | 9182 | vector bool short); |
6e5bb5ad JM |
9183 | vector unsigned short vec_add (vector unsigned short, |
9184 | vector unsigned short); | |
b0b343db JJ |
9185 | vector signed int vec_add (vector bool int, vector signed int); |
9186 | vector signed int vec_add (vector signed int, vector bool int); | |
333c8841 | 9187 | vector signed int vec_add (vector signed int, vector signed int); |
b0b343db JJ |
9188 | vector unsigned int vec_add (vector bool int, vector unsigned int); |
9189 | vector unsigned int vec_add (vector unsigned int, vector bool int); | |
333c8841 AH |
9190 | vector unsigned int vec_add (vector unsigned int, vector unsigned int); |
9191 | vector float vec_add (vector float, vector float); | |
9192 | ||
b0b343db JJ |
9193 | vector float vec_vaddfp (vector float, vector float); |
9194 | ||
9195 | vector signed int vec_vadduwm (vector bool int, vector signed int); | |
9196 | vector signed int vec_vadduwm (vector signed int, vector bool int); | |
9197 | vector signed int vec_vadduwm (vector signed int, vector signed int); | |
9198 | vector unsigned int vec_vadduwm (vector bool int, vector unsigned int); | |
9199 | vector unsigned int vec_vadduwm (vector unsigned int, vector bool int); | |
9200 | vector unsigned int vec_vadduwm (vector unsigned int, | |
9201 | vector unsigned int); | |
9202 | ||
9203 | vector signed short vec_vadduhm (vector bool short, | |
9204 | vector signed short); | |
9205 | vector signed short vec_vadduhm (vector signed short, | |
9206 | vector bool short); | |
9207 | vector signed short vec_vadduhm (vector signed short, | |
9208 | vector signed short); | |
9209 | vector unsigned short vec_vadduhm (vector bool short, | |
9210 | vector unsigned short); | |
9211 | vector unsigned short vec_vadduhm (vector unsigned short, | |
9212 | vector bool short); | |
9213 | vector unsigned short vec_vadduhm (vector unsigned short, | |
9214 | vector unsigned short); | |
9215 | ||
9216 | vector signed char vec_vaddubm (vector bool char, vector signed char); | |
9217 | vector signed char vec_vaddubm (vector signed char, vector bool char); | |
9218 | vector signed char vec_vaddubm (vector signed char, vector signed char); | |
9219 | vector unsigned char vec_vaddubm (vector bool char, | |
9220 | vector unsigned char); | |
9221 | vector unsigned char vec_vaddubm (vector unsigned char, | |
9222 | vector bool char); | |
9223 | vector unsigned char vec_vaddubm (vector unsigned char, | |
9224 | vector unsigned char); | |
9225 | ||
333c8841 AH |
9226 | vector unsigned int vec_addc (vector unsigned int, vector unsigned int); |
9227 | ||
b0b343db JJ |
9228 | vector unsigned char vec_adds (vector bool char, vector unsigned char); |
9229 | vector unsigned char vec_adds (vector unsigned char, vector bool char); | |
924fcc4e JM |
9230 | vector unsigned char vec_adds (vector unsigned char, |
9231 | vector unsigned char); | |
b0b343db JJ |
9232 | vector signed char vec_adds (vector bool char, vector signed char); |
9233 | vector signed char vec_adds (vector signed char, vector bool char); | |
333c8841 | 9234 | vector signed char vec_adds (vector signed char, vector signed char); |
b0b343db | 9235 | vector unsigned short vec_adds (vector bool short, |
924fcc4e JM |
9236 | vector unsigned short); |
9237 | vector unsigned short vec_adds (vector unsigned short, | |
b0b343db | 9238 | vector bool short); |
6e5bb5ad JM |
9239 | vector unsigned short vec_adds (vector unsigned short, |
9240 | vector unsigned short); | |
b0b343db JJ |
9241 | vector signed short vec_adds (vector bool short, vector signed short); |
9242 | vector signed short vec_adds (vector signed short, vector bool short); | |
333c8841 | 9243 | vector signed short vec_adds (vector signed short, vector signed short); |
b0b343db JJ |
9244 | vector unsigned int vec_adds (vector bool int, vector unsigned int); |
9245 | vector unsigned int vec_adds (vector unsigned int, vector bool int); | |
333c8841 | 9246 | vector unsigned int vec_adds (vector unsigned int, vector unsigned int); |
b0b343db JJ |
9247 | vector signed int vec_adds (vector bool int, vector signed int); |
9248 | vector signed int vec_adds (vector signed int, vector bool int); | |
333c8841 AH |
9249 | vector signed int vec_adds (vector signed int, vector signed int); |
9250 | ||
b0b343db JJ |
9251 | vector signed int vec_vaddsws (vector bool int, vector signed int); |
9252 | vector signed int vec_vaddsws (vector signed int, vector bool int); | |
9253 | vector signed int vec_vaddsws (vector signed int, vector signed int); | |
9254 | ||
9255 | vector unsigned int vec_vadduws (vector bool int, vector unsigned int); | |
9256 | vector unsigned int vec_vadduws (vector unsigned int, vector bool int); | |
9257 | vector unsigned int vec_vadduws (vector unsigned int, | |
9258 | vector unsigned int); | |
9259 | ||
9260 | vector signed short vec_vaddshs (vector bool short, | |
9261 | vector signed short); | |
9262 | vector signed short vec_vaddshs (vector signed short, | |
9263 | vector bool short); | |
9264 | vector signed short vec_vaddshs (vector signed short, | |
9265 | vector signed short); | |
9266 | ||
9267 | vector unsigned short vec_vadduhs (vector bool short, | |
9268 | vector unsigned short); | |
9269 | vector unsigned short vec_vadduhs (vector unsigned short, | |
9270 | vector bool short); | |
9271 | vector unsigned short vec_vadduhs (vector unsigned short, | |
9272 | vector unsigned short); | |
9273 | ||
9274 | vector signed char vec_vaddsbs (vector bool char, vector signed char); | |
9275 | vector signed char vec_vaddsbs (vector signed char, vector bool char); | |
9276 | vector signed char vec_vaddsbs (vector signed char, vector signed char); | |
9277 | ||
9278 | vector unsigned char vec_vaddubs (vector bool char, | |
9279 | vector unsigned char); | |
9280 | vector unsigned char vec_vaddubs (vector unsigned char, | |
9281 | vector bool char); | |
9282 | vector unsigned char vec_vaddubs (vector unsigned char, | |
9283 | vector unsigned char); | |
9284 | ||
333c8841 | 9285 | vector float vec_and (vector float, vector float); |
b0b343db JJ |
9286 | vector float vec_and (vector float, vector bool int); |
9287 | vector float vec_and (vector bool int, vector float); | |
9288 | vector bool int vec_and (vector bool int, vector bool int); | |
9289 | vector signed int vec_and (vector bool int, vector signed int); | |
9290 | vector signed int vec_and (vector signed int, vector bool int); | |
333c8841 | 9291 | vector signed int vec_and (vector signed int, vector signed int); |
b0b343db JJ |
9292 | vector unsigned int vec_and (vector bool int, vector unsigned int); |
9293 | vector unsigned int vec_and (vector unsigned int, vector bool int); | |
333c8841 | 9294 | vector unsigned int vec_and (vector unsigned int, vector unsigned int); |
b0b343db JJ |
9295 | vector bool short vec_and (vector bool short, vector bool short); |
9296 | vector signed short vec_and (vector bool short, vector signed short); | |
9297 | vector signed short vec_and (vector signed short, vector bool short); | |
333c8841 | 9298 | vector signed short vec_and (vector signed short, vector signed short); |
b0b343db | 9299 | vector unsigned short vec_and (vector bool short, |
924fcc4e JM |
9300 | vector unsigned short); |
9301 | vector unsigned short vec_and (vector unsigned short, | |
b0b343db | 9302 | vector bool short); |
6e5bb5ad JM |
9303 | vector unsigned short vec_and (vector unsigned short, |
9304 | vector unsigned short); | |
b0b343db JJ |
9305 | vector signed char vec_and (vector bool char, vector signed char); |
9306 | vector bool char vec_and (vector bool char, vector bool char); | |
9307 | vector signed char vec_and (vector signed char, vector bool char); | |
333c8841 | 9308 | vector signed char vec_and (vector signed char, vector signed char); |
b0b343db JJ |
9309 | vector unsigned char vec_and (vector bool char, vector unsigned char); |
9310 | vector unsigned char vec_and (vector unsigned char, vector bool char); | |
924fcc4e JM |
9311 | vector unsigned char vec_and (vector unsigned char, |
9312 | vector unsigned char); | |
333c8841 AH |
9313 | |
9314 | vector float vec_andc (vector float, vector float); | |
b0b343db JJ |
9315 | vector float vec_andc (vector float, vector bool int); |
9316 | vector float vec_andc (vector bool int, vector float); | |
9317 | vector bool int vec_andc (vector bool int, vector bool int); | |
9318 | vector signed int vec_andc (vector bool int, vector signed int); | |
9319 | vector signed int vec_andc (vector signed int, vector bool int); | |
333c8841 | 9320 | vector signed int vec_andc (vector signed int, vector signed int); |
b0b343db JJ |
9321 | vector unsigned int vec_andc (vector bool int, vector unsigned int); |
9322 | vector unsigned int vec_andc (vector unsigned int, vector bool int); | |
333c8841 | 9323 | vector unsigned int vec_andc (vector unsigned int, vector unsigned int); |
b0b343db JJ |
9324 | vector bool short vec_andc (vector bool short, vector bool short); |
9325 | vector signed short vec_andc (vector bool short, vector signed short); | |
9326 | vector signed short vec_andc (vector signed short, vector bool short); | |
333c8841 | 9327 | vector signed short vec_andc (vector signed short, vector signed short); |
b0b343db | 9328 | vector unsigned short vec_andc (vector bool short, |
924fcc4e JM |
9329 | vector unsigned short); |
9330 | vector unsigned short vec_andc (vector unsigned short, | |
b0b343db | 9331 | vector bool short); |
6e5bb5ad JM |
9332 | vector unsigned short vec_andc (vector unsigned short, |
9333 | vector unsigned short); | |
b0b343db JJ |
9334 | vector signed char vec_andc (vector bool char, vector signed char); |
9335 | vector bool char vec_andc (vector bool char, vector bool char); | |
9336 | vector signed char vec_andc (vector signed char, vector bool char); | |
333c8841 | 9337 | vector signed char vec_andc (vector signed char, vector signed char); |
b0b343db JJ |
9338 | vector unsigned char vec_andc (vector bool char, vector unsigned char); |
9339 | vector unsigned char vec_andc (vector unsigned char, vector bool char); | |
924fcc4e JM |
9340 | vector unsigned char vec_andc (vector unsigned char, |
9341 | vector unsigned char); | |
333c8841 | 9342 | |
924fcc4e JM |
9343 | vector unsigned char vec_avg (vector unsigned char, |
9344 | vector unsigned char); | |
333c8841 | 9345 | vector signed char vec_avg (vector signed char, vector signed char); |
6e5bb5ad JM |
9346 | vector unsigned short vec_avg (vector unsigned short, |
9347 | vector unsigned short); | |
333c8841 AH |
9348 | vector signed short vec_avg (vector signed short, vector signed short); |
9349 | vector unsigned int vec_avg (vector unsigned int, vector unsigned int); | |
9350 | vector signed int vec_avg (vector signed int, vector signed int); | |
9351 | ||
b0b343db JJ |
9352 | vector signed int vec_vavgsw (vector signed int, vector signed int); |
9353 | ||
9354 | vector unsigned int vec_vavguw (vector unsigned int, | |
9355 | vector unsigned int); | |
9356 | ||
9357 | vector signed short vec_vavgsh (vector signed short, | |
9358 | vector signed short); | |
9359 | ||
9360 | vector unsigned short vec_vavguh (vector unsigned short, | |
9361 | vector unsigned short); | |
9362 | ||
9363 | vector signed char vec_vavgsb (vector signed char, vector signed char); | |
9364 | ||
9365 | vector unsigned char vec_vavgub (vector unsigned char, | |
9366 | vector unsigned char); | |
9367 | ||
333c8841 AH |
9368 | vector float vec_ceil (vector float); |
9369 | ||
9370 | vector signed int vec_cmpb (vector float, vector float); | |
9371 | ||
b0b343db JJ |
9372 | vector bool char vec_cmpeq (vector signed char, vector signed char); |
9373 | vector bool char vec_cmpeq (vector unsigned char, vector unsigned char); | |
9374 | vector bool short vec_cmpeq (vector signed short, vector signed short); | |
9375 | vector bool short vec_cmpeq (vector unsigned short, | |
9376 | vector unsigned short); | |
9377 | vector bool int vec_cmpeq (vector signed int, vector signed int); | |
9378 | vector bool int vec_cmpeq (vector unsigned int, vector unsigned int); | |
9379 | vector bool int vec_cmpeq (vector float, vector float); | |
333c8841 | 9380 | |
b0b343db | 9381 | vector bool int vec_vcmpeqfp (vector float, vector float); |
333c8841 | 9382 | |
b0b343db JJ |
9383 | vector bool int vec_vcmpequw (vector signed int, vector signed int); |
9384 | vector bool int vec_vcmpequw (vector unsigned int, vector unsigned int); | |
9385 | ||
9386 | vector bool short vec_vcmpequh (vector signed short, | |
9387 | vector signed short); | |
9388 | vector bool short vec_vcmpequh (vector unsigned short, | |
9389 | vector unsigned short); | |
333c8841 | 9390 | |
b0b343db JJ |
9391 | vector bool char vec_vcmpequb (vector signed char, vector signed char); |
9392 | vector bool char vec_vcmpequb (vector unsigned char, | |
9393 | vector unsigned char); | |
333c8841 | 9394 | |
b0b343db | 9395 | vector bool int vec_cmpge (vector float, vector float); |
333c8841 | 9396 | |
b0b343db JJ |
9397 | vector bool char vec_cmpgt (vector unsigned char, vector unsigned char); |
9398 | vector bool char vec_cmpgt (vector signed char, vector signed char); | |
9399 | vector bool short vec_cmpgt (vector unsigned short, | |
9400 | vector unsigned short); | |
9401 | vector bool short vec_cmpgt (vector signed short, vector signed short); | |
9402 | vector bool int vec_cmpgt (vector unsigned int, vector unsigned int); | |
9403 | vector bool int vec_cmpgt (vector signed int, vector signed int); | |
9404 | vector bool int vec_cmpgt (vector float, vector float); | |
9405 | ||
9406 | vector bool int vec_vcmpgtfp (vector float, vector float); | |
9407 | ||
9408 | vector bool int vec_vcmpgtsw (vector signed int, vector signed int); | |
9409 | ||
9410 | vector bool int vec_vcmpgtuw (vector unsigned int, vector unsigned int); | |
333c8841 | 9411 | |
b0b343db JJ |
9412 | vector bool short vec_vcmpgtsh (vector signed short, |
9413 | vector signed short); | |
9414 | ||
9415 | vector bool short vec_vcmpgtuh (vector unsigned short, | |
9416 | vector unsigned short); | |
9417 | ||
9418 | vector bool char vec_vcmpgtsb (vector signed char, vector signed char); | |
9419 | ||
9420 | vector bool char vec_vcmpgtub (vector unsigned char, | |
9421 | vector unsigned char); | |
9422 | ||
9423 | vector bool int vec_cmple (vector float, vector float); | |
9424 | ||
9425 | vector bool char vec_cmplt (vector unsigned char, vector unsigned char); | |
9426 | vector bool char vec_cmplt (vector signed char, vector signed char); | |
9427 | vector bool short vec_cmplt (vector unsigned short, | |
9428 | vector unsigned short); | |
9429 | vector bool short vec_cmplt (vector signed short, vector signed short); | |
9430 | vector bool int vec_cmplt (vector unsigned int, vector unsigned int); | |
9431 | vector bool int vec_cmplt (vector signed int, vector signed int); | |
9432 | vector bool int vec_cmplt (vector float, vector float); | |
333c8841 | 9433 | |
b0b343db JJ |
9434 | vector float vec_ctf (vector unsigned int, const int); |
9435 | vector float vec_ctf (vector signed int, const int); | |
333c8841 | 9436 | |
b0b343db JJ |
9437 | vector float vec_vcfsx (vector signed int, const int); |
9438 | ||
9439 | vector float vec_vcfux (vector unsigned int, const int); | |
9440 | ||
9441 | vector signed int vec_cts (vector float, const int); | |
9442 | ||
9443 | vector unsigned int vec_ctu (vector float, const int); | |
9444 | ||
9445 | void vec_dss (const int); | |
333c8841 AH |
9446 | |
9447 | void vec_dssall (void); | |
9448 | ||
b0b343db JJ |
9449 | void vec_dst (const vector unsigned char *, int, const int); |
9450 | void vec_dst (const vector signed char *, int, const int); | |
9451 | void vec_dst (const vector bool char *, int, const int); | |
9452 | void vec_dst (const vector unsigned short *, int, const int); | |
9453 | void vec_dst (const vector signed short *, int, const int); | |
9454 | void vec_dst (const vector bool short *, int, const int); | |
9455 | void vec_dst (const vector pixel *, int, const int); | |
9456 | void vec_dst (const vector unsigned int *, int, const int); | |
9457 | void vec_dst (const vector signed int *, int, const int); | |
9458 | void vec_dst (const vector bool int *, int, const int); | |
9459 | void vec_dst (const vector float *, int, const int); | |
9460 | void vec_dst (const unsigned char *, int, const int); | |
9461 | void vec_dst (const signed char *, int, const int); | |
9462 | void vec_dst (const unsigned short *, int, const int); | |
9463 | void vec_dst (const short *, int, const int); | |
9464 | void vec_dst (const unsigned int *, int, const int); | |
9465 | void vec_dst (const int *, int, const int); | |
9466 | void vec_dst (const unsigned long *, int, const int); | |
9467 | void vec_dst (const long *, int, const int); | |
9468 | void vec_dst (const float *, int, const int); | |
9469 | ||
9470 | void vec_dstst (const vector unsigned char *, int, const int); | |
9471 | void vec_dstst (const vector signed char *, int, const int); | |
9472 | void vec_dstst (const vector bool char *, int, const int); | |
9473 | void vec_dstst (const vector unsigned short *, int, const int); | |
9474 | void vec_dstst (const vector signed short *, int, const int); | |
9475 | void vec_dstst (const vector bool short *, int, const int); | |
9476 | void vec_dstst (const vector pixel *, int, const int); | |
9477 | void vec_dstst (const vector unsigned int *, int, const int); | |
9478 | void vec_dstst (const vector signed int *, int, const int); | |
9479 | void vec_dstst (const vector bool int *, int, const int); | |
9480 | void vec_dstst (const vector float *, int, const int); | |
9481 | void vec_dstst (const unsigned char *, int, const int); | |
9482 | void vec_dstst (const signed char *, int, const int); | |
9483 | void vec_dstst (const unsigned short *, int, const int); | |
9484 | void vec_dstst (const short *, int, const int); | |
9485 | void vec_dstst (const unsigned int *, int, const int); | |
9486 | void vec_dstst (const int *, int, const int); | |
9487 | void vec_dstst (const unsigned long *, int, const int); | |
9488 | void vec_dstst (const long *, int, const int); | |
9489 | void vec_dstst (const float *, int, const int); | |
9490 | ||
9491 | void vec_dststt (const vector unsigned char *, int, const int); | |
9492 | void vec_dststt (const vector signed char *, int, const int); | |
9493 | void vec_dststt (const vector bool char *, int, const int); | |
9494 | void vec_dststt (const vector unsigned short *, int, const int); | |
9495 | void vec_dststt (const vector signed short *, int, const int); | |
9496 | void vec_dststt (const vector bool short *, int, const int); | |
9497 | void vec_dststt (const vector pixel *, int, const int); | |
9498 | void vec_dststt (const vector unsigned int *, int, const int); | |
9499 | void vec_dststt (const vector signed int *, int, const int); | |
9500 | void vec_dststt (const vector bool int *, int, const int); | |
9501 | void vec_dststt (const vector float *, int, const int); | |
9502 | void vec_dststt (const unsigned char *, int, const int); | |
9503 | void vec_dststt (const signed char *, int, const int); | |
9504 | void vec_dststt (const unsigned short *, int, const int); | |
9505 | void vec_dststt (const short *, int, const int); | |
9506 | void vec_dststt (const unsigned int *, int, const int); | |
9507 | void vec_dststt (const int *, int, const int); | |
9508 | void vec_dststt (const unsigned long *, int, const int); | |
9509 | void vec_dststt (const long *, int, const int); | |
9510 | void vec_dststt (const float *, int, const int); | |
9511 | ||
9512 | void vec_dstt (const vector unsigned char *, int, const int); | |
9513 | void vec_dstt (const vector signed char *, int, const int); | |
9514 | void vec_dstt (const vector bool char *, int, const int); | |
9515 | void vec_dstt (const vector unsigned short *, int, const int); | |
9516 | void vec_dstt (const vector signed short *, int, const int); | |
9517 | void vec_dstt (const vector bool short *, int, const int); | |
9518 | void vec_dstt (const vector pixel *, int, const int); | |
9519 | void vec_dstt (const vector unsigned int *, int, const int); | |
9520 | void vec_dstt (const vector signed int *, int, const int); | |
9521 | void vec_dstt (const vector bool int *, int, const int); | |
9522 | void vec_dstt (const vector float *, int, const int); | |
9523 | void vec_dstt (const unsigned char *, int, const int); | |
9524 | void vec_dstt (const signed char *, int, const int); | |
9525 | void vec_dstt (const unsigned short *, int, const int); | |
9526 | void vec_dstt (const short *, int, const int); | |
9527 | void vec_dstt (const unsigned int *, int, const int); | |
9528 | void vec_dstt (const int *, int, const int); | |
9529 | void vec_dstt (const unsigned long *, int, const int); | |
9530 | void vec_dstt (const long *, int, const int); | |
9531 | void vec_dstt (const float *, int, const int); | |
9532 | ||
9533 | vector float vec_expte (vector float); | |
9534 | ||
9535 | vector float vec_floor (vector float); | |
9536 | ||
9537 | vector float vec_ld (int, const vector float *); | |
9538 | vector float vec_ld (int, const float *); | |
9539 | vector bool int vec_ld (int, const vector bool int *); | |
9540 | vector signed int vec_ld (int, const vector signed int *); | |
9541 | vector signed int vec_ld (int, const int *); | |
9542 | vector signed int vec_ld (int, const long *); | |
9543 | vector unsigned int vec_ld (int, const vector unsigned int *); | |
9544 | vector unsigned int vec_ld (int, const unsigned int *); | |
9545 | vector unsigned int vec_ld (int, const unsigned long *); | |
9546 | vector bool short vec_ld (int, const vector bool short *); | |
9547 | vector pixel vec_ld (int, const vector pixel *); | |
9548 | vector signed short vec_ld (int, const vector signed short *); | |
9549 | vector signed short vec_ld (int, const short *); | |
9550 | vector unsigned short vec_ld (int, const vector unsigned short *); | |
9551 | vector unsigned short vec_ld (int, const unsigned short *); | |
9552 | vector bool char vec_ld (int, const vector bool char *); | |
9553 | vector signed char vec_ld (int, const vector signed char *); | |
9554 | vector signed char vec_ld (int, const signed char *); | |
9555 | vector unsigned char vec_ld (int, const vector unsigned char *); | |
9556 | vector unsigned char vec_ld (int, const unsigned char *); | |
9557 | ||
9558 | vector signed char vec_lde (int, const signed char *); | |
9559 | vector unsigned char vec_lde (int, const unsigned char *); | |
9560 | vector signed short vec_lde (int, const short *); | |
9561 | vector unsigned short vec_lde (int, const unsigned short *); | |
9562 | vector float vec_lde (int, const float *); | |
9563 | vector signed int vec_lde (int, const int *); | |
9564 | vector unsigned int vec_lde (int, const unsigned int *); | |
9565 | vector signed int vec_lde (int, const long *); | |
9566 | vector unsigned int vec_lde (int, const unsigned long *); | |
9567 | ||
9568 | vector float vec_lvewx (int, float *); | |
9569 | vector signed int vec_lvewx (int, int *); | |
9570 | vector unsigned int vec_lvewx (int, unsigned int *); | |
9571 | vector signed int vec_lvewx (int, long *); | |
9572 | vector unsigned int vec_lvewx (int, unsigned long *); | |
9573 | ||
9574 | vector signed short vec_lvehx (int, short *); | |
9575 | vector unsigned short vec_lvehx (int, unsigned short *); | |
9576 | ||
9577 | vector signed char vec_lvebx (int, char *); | |
9578 | vector unsigned char vec_lvebx (int, unsigned char *); | |
9579 | ||
9580 | vector float vec_ldl (int, const vector float *); | |
9581 | vector float vec_ldl (int, const float *); | |
9582 | vector bool int vec_ldl (int, const vector bool int *); | |
9583 | vector signed int vec_ldl (int, const vector signed int *); | |
9584 | vector signed int vec_ldl (int, const int *); | |
9585 | vector signed int vec_ldl (int, const long *); | |
9586 | vector unsigned int vec_ldl (int, const vector unsigned int *); | |
9587 | vector unsigned int vec_ldl (int, const unsigned int *); | |
9588 | vector unsigned int vec_ldl (int, const unsigned long *); | |
9589 | vector bool short vec_ldl (int, const vector bool short *); | |
9590 | vector pixel vec_ldl (int, const vector pixel *); | |
9591 | vector signed short vec_ldl (int, const vector signed short *); | |
9592 | vector signed short vec_ldl (int, const short *); | |
9593 | vector unsigned short vec_ldl (int, const vector unsigned short *); | |
9594 | vector unsigned short vec_ldl (int, const unsigned short *); | |
9595 | vector bool char vec_ldl (int, const vector bool char *); | |
9596 | vector signed char vec_ldl (int, const vector signed char *); | |
9597 | vector signed char vec_ldl (int, const signed char *); | |
9598 | vector unsigned char vec_ldl (int, const vector unsigned char *); | |
9599 | vector unsigned char vec_ldl (int, const unsigned char *); | |
333c8841 AH |
9600 | |
9601 | vector float vec_loge (vector float); | |
9602 | ||
b0b343db JJ |
9603 | vector unsigned char vec_lvsl (int, const volatile unsigned char *); |
9604 | vector unsigned char vec_lvsl (int, const volatile signed char *); | |
9605 | vector unsigned char vec_lvsl (int, const volatile unsigned short *); | |
9606 | vector unsigned char vec_lvsl (int, const volatile short *); | |
9607 | vector unsigned char vec_lvsl (int, const volatile unsigned int *); | |
9608 | vector unsigned char vec_lvsl (int, const volatile int *); | |
9609 | vector unsigned char vec_lvsl (int, const volatile unsigned long *); | |
9610 | vector unsigned char vec_lvsl (int, const volatile long *); | |
9611 | vector unsigned char vec_lvsl (int, const volatile float *); | |
9612 | ||
9613 | vector unsigned char vec_lvsr (int, const volatile unsigned char *); | |
9614 | vector unsigned char vec_lvsr (int, const volatile signed char *); | |
9615 | vector unsigned char vec_lvsr (int, const volatile unsigned short *); | |
9616 | vector unsigned char vec_lvsr (int, const volatile short *); | |
9617 | vector unsigned char vec_lvsr (int, const volatile unsigned int *); | |
9618 | vector unsigned char vec_lvsr (int, const volatile int *); | |
9619 | vector unsigned char vec_lvsr (int, const volatile unsigned long *); | |
9620 | vector unsigned char vec_lvsr (int, const volatile long *); | |
9621 | vector unsigned char vec_lvsr (int, const volatile float *); | |
333c8841 AH |
9622 | |
9623 | vector float vec_madd (vector float, vector float, vector float); | |
9624 | ||
b0b343db JJ |
9625 | vector signed short vec_madds (vector signed short, |
9626 | vector signed short, | |
6e5bb5ad | 9627 | vector signed short); |
333c8841 | 9628 | |
b0b343db JJ |
9629 | vector unsigned char vec_max (vector bool char, vector unsigned char); |
9630 | vector unsigned char vec_max (vector unsigned char, vector bool char); | |
924fcc4e JM |
9631 | vector unsigned char vec_max (vector unsigned char, |
9632 | vector unsigned char); | |
b0b343db JJ |
9633 | vector signed char vec_max (vector bool char, vector signed char); |
9634 | vector signed char vec_max (vector signed char, vector bool char); | |
333c8841 | 9635 | vector signed char vec_max (vector signed char, vector signed char); |
b0b343db | 9636 | vector unsigned short vec_max (vector bool short, |
924fcc4e JM |
9637 | vector unsigned short); |
9638 | vector unsigned short vec_max (vector unsigned short, | |
b0b343db | 9639 | vector bool short); |
6e5bb5ad JM |
9640 | vector unsigned short vec_max (vector unsigned short, |
9641 | vector unsigned short); | |
b0b343db JJ |
9642 | vector signed short vec_max (vector bool short, vector signed short); |
9643 | vector signed short vec_max (vector signed short, vector bool short); | |
333c8841 | 9644 | vector signed short vec_max (vector signed short, vector signed short); |
b0b343db JJ |
9645 | vector unsigned int vec_max (vector bool int, vector unsigned int); |
9646 | vector unsigned int vec_max (vector unsigned int, vector bool int); | |
333c8841 | 9647 | vector unsigned int vec_max (vector unsigned int, vector unsigned int); |
b0b343db JJ |
9648 | vector signed int vec_max (vector bool int, vector signed int); |
9649 | vector signed int vec_max (vector signed int, vector bool int); | |
333c8841 AH |
9650 | vector signed int vec_max (vector signed int, vector signed int); |
9651 | vector float vec_max (vector float, vector float); | |
9652 | ||
b0b343db JJ |
9653 | vector float vec_vmaxfp (vector float, vector float); |
9654 | ||
9655 | vector signed int vec_vmaxsw (vector bool int, vector signed int); | |
9656 | vector signed int vec_vmaxsw (vector signed int, vector bool int); | |
9657 | vector signed int vec_vmaxsw (vector signed int, vector signed int); | |
9658 | ||
9659 | vector unsigned int vec_vmaxuw (vector bool int, vector unsigned int); | |
9660 | vector unsigned int vec_vmaxuw (vector unsigned int, vector bool int); | |
9661 | vector unsigned int vec_vmaxuw (vector unsigned int, | |
9662 | vector unsigned int); | |
9663 | ||
9664 | vector signed short vec_vmaxsh (vector bool short, vector signed short); | |
9665 | vector signed short vec_vmaxsh (vector signed short, vector bool short); | |
9666 | vector signed short vec_vmaxsh (vector signed short, | |
9667 | vector signed short); | |
9668 | ||
9669 | vector unsigned short vec_vmaxuh (vector bool short, | |
9670 | vector unsigned short); | |
9671 | vector unsigned short vec_vmaxuh (vector unsigned short, | |
9672 | vector bool short); | |
9673 | vector unsigned short vec_vmaxuh (vector unsigned short, | |
9674 | vector unsigned short); | |
9675 | ||
9676 | vector signed char vec_vmaxsb (vector bool char, vector signed char); | |
9677 | vector signed char vec_vmaxsb (vector signed char, vector bool char); | |
9678 | vector signed char vec_vmaxsb (vector signed char, vector signed char); | |
9679 | ||
9680 | vector unsigned char vec_vmaxub (vector bool char, | |
9681 | vector unsigned char); | |
9682 | vector unsigned char vec_vmaxub (vector unsigned char, | |
9683 | vector bool char); | |
9684 | vector unsigned char vec_vmaxub (vector unsigned char, | |
9685 | vector unsigned char); | |
9686 | ||
9687 | vector bool char vec_mergeh (vector bool char, vector bool char); | |
333c8841 | 9688 | vector signed char vec_mergeh (vector signed char, vector signed char); |
6e5bb5ad JM |
9689 | vector unsigned char vec_mergeh (vector unsigned char, |
9690 | vector unsigned char); | |
b0b343db JJ |
9691 | vector bool short vec_mergeh (vector bool short, vector bool short); |
9692 | vector pixel vec_mergeh (vector pixel, vector pixel); | |
924fcc4e JM |
9693 | vector signed short vec_mergeh (vector signed short, |
9694 | vector signed short); | |
6e5bb5ad JM |
9695 | vector unsigned short vec_mergeh (vector unsigned short, |
9696 | vector unsigned short); | |
333c8841 | 9697 | vector float vec_mergeh (vector float, vector float); |
b0b343db | 9698 | vector bool int vec_mergeh (vector bool int, vector bool int); |
333c8841 | 9699 | vector signed int vec_mergeh (vector signed int, vector signed int); |
924fcc4e JM |
9700 | vector unsigned int vec_mergeh (vector unsigned int, |
9701 | vector unsigned int); | |
333c8841 | 9702 | |
b0b343db JJ |
9703 | vector float vec_vmrghw (vector float, vector float); |
9704 | vector bool int vec_vmrghw (vector bool int, vector bool int); | |
9705 | vector signed int vec_vmrghw (vector signed int, vector signed int); | |
9706 | vector unsigned int vec_vmrghw (vector unsigned int, | |
9707 | vector unsigned int); | |
9708 | ||
9709 | vector bool short vec_vmrghh (vector bool short, vector bool short); | |
9710 | vector signed short vec_vmrghh (vector signed short, | |
9711 | vector signed short); | |
9712 | vector unsigned short vec_vmrghh (vector unsigned short, | |
9713 | vector unsigned short); | |
9714 | vector pixel vec_vmrghh (vector pixel, vector pixel); | |
9715 | ||
9716 | vector bool char vec_vmrghb (vector bool char, vector bool char); | |
9717 | vector signed char vec_vmrghb (vector signed char, vector signed char); | |
9718 | vector unsigned char vec_vmrghb (vector unsigned char, | |
9719 | vector unsigned char); | |
9720 | ||
9721 | vector bool char vec_mergel (vector bool char, vector bool char); | |
333c8841 | 9722 | vector signed char vec_mergel (vector signed char, vector signed char); |
6e5bb5ad JM |
9723 | vector unsigned char vec_mergel (vector unsigned char, |
9724 | vector unsigned char); | |
b0b343db JJ |
9725 | vector bool short vec_mergel (vector bool short, vector bool short); |
9726 | vector pixel vec_mergel (vector pixel, vector pixel); | |
924fcc4e JM |
9727 | vector signed short vec_mergel (vector signed short, |
9728 | vector signed short); | |
6e5bb5ad JM |
9729 | vector unsigned short vec_mergel (vector unsigned short, |
9730 | vector unsigned short); | |
333c8841 | 9731 | vector float vec_mergel (vector float, vector float); |
b0b343db | 9732 | vector bool int vec_mergel (vector bool int, vector bool int); |
333c8841 | 9733 | vector signed int vec_mergel (vector signed int, vector signed int); |
924fcc4e JM |
9734 | vector unsigned int vec_mergel (vector unsigned int, |
9735 | vector unsigned int); | |
333c8841 | 9736 | |
b0b343db JJ |
9737 | vector float vec_vmrglw (vector float, vector float); |
9738 | vector signed int vec_vmrglw (vector signed int, vector signed int); | |
9739 | vector unsigned int vec_vmrglw (vector unsigned int, | |
9740 | vector unsigned int); | |
9741 | vector bool int vec_vmrglw (vector bool int, vector bool int); | |
333c8841 | 9742 | |
b0b343db JJ |
9743 | vector bool short vec_vmrglh (vector bool short, vector bool short); |
9744 | vector signed short vec_vmrglh (vector signed short, | |
9745 | vector signed short); | |
9746 | vector unsigned short vec_vmrglh (vector unsigned short, | |
9747 | vector unsigned short); | |
9748 | vector pixel vec_vmrglh (vector pixel, vector pixel); | |
9749 | ||
9750 | vector bool char vec_vmrglb (vector bool char, vector bool char); | |
9751 | vector signed char vec_vmrglb (vector signed char, vector signed char); | |
9752 | vector unsigned char vec_vmrglb (vector unsigned char, | |
9753 | vector unsigned char); | |
333c8841 | 9754 | |
b0b343db | 9755 | vector unsigned short vec_mfvscr (void); |
333c8841 | 9756 | |
b0b343db JJ |
9757 | vector unsigned char vec_min (vector bool char, vector unsigned char); |
9758 | vector unsigned char vec_min (vector unsigned char, vector bool char); | |
924fcc4e JM |
9759 | vector unsigned char vec_min (vector unsigned char, |
9760 | vector unsigned char); | |
b0b343db JJ |
9761 | vector signed char vec_min (vector bool char, vector signed char); |
9762 | vector signed char vec_min (vector signed char, vector bool char); | |
333c8841 | 9763 | vector signed char vec_min (vector signed char, vector signed char); |
b0b343db | 9764 | vector unsigned short vec_min (vector bool short, |
924fcc4e JM |
9765 | vector unsigned short); |
9766 | vector unsigned short vec_min (vector unsigned short, | |
b0b343db | 9767 | vector bool short); |
6e5bb5ad JM |
9768 | vector unsigned short vec_min (vector unsigned short, |
9769 | vector unsigned short); | |
b0b343db JJ |
9770 | vector signed short vec_min (vector bool short, vector signed short); |
9771 | vector signed short vec_min (vector signed short, vector bool short); | |
333c8841 | 9772 | vector signed short vec_min (vector signed short, vector signed short); |
b0b343db JJ |
9773 | vector unsigned int vec_min (vector bool int, vector unsigned int); |
9774 | vector unsigned int vec_min (vector unsigned int, vector bool int); | |
333c8841 | 9775 | vector unsigned int vec_min (vector unsigned int, vector unsigned int); |
b0b343db JJ |
9776 | vector signed int vec_min (vector bool int, vector signed int); |
9777 | vector signed int vec_min (vector signed int, vector bool int); | |
333c8841 AH |
9778 | vector signed int vec_min (vector signed int, vector signed int); |
9779 | vector float vec_min (vector float, vector float); | |
9780 | ||
b0b343db JJ |
9781 | vector float vec_vminfp (vector float, vector float); |
9782 | ||
9783 | vector signed int vec_vminsw (vector bool int, vector signed int); | |
9784 | vector signed int vec_vminsw (vector signed int, vector bool int); | |
9785 | vector signed int vec_vminsw (vector signed int, vector signed int); | |
9786 | ||
9787 | vector unsigned int vec_vminuw (vector bool int, vector unsigned int); | |
9788 | vector unsigned int vec_vminuw (vector unsigned int, vector bool int); | |
9789 | vector unsigned int vec_vminuw (vector unsigned int, | |
9790 | vector unsigned int); | |
9791 | ||
9792 | vector signed short vec_vminsh (vector bool short, vector signed short); | |
9793 | vector signed short vec_vminsh (vector signed short, vector bool short); | |
9794 | vector signed short vec_vminsh (vector signed short, | |
9795 | vector signed short); | |
9796 | ||
9797 | vector unsigned short vec_vminuh (vector bool short, | |
9798 | vector unsigned short); | |
9799 | vector unsigned short vec_vminuh (vector unsigned short, | |
9800 | vector bool short); | |
9801 | vector unsigned short vec_vminuh (vector unsigned short, | |
9802 | vector unsigned short); | |
9803 | ||
9804 | vector signed char vec_vminsb (vector bool char, vector signed char); | |
9805 | vector signed char vec_vminsb (vector signed char, vector bool char); | |
9806 | vector signed char vec_vminsb (vector signed char, vector signed char); | |
9807 | ||
9808 | vector unsigned char vec_vminub (vector bool char, | |
9809 | vector unsigned char); | |
9810 | vector unsigned char vec_vminub (vector unsigned char, | |
9811 | vector bool char); | |
9812 | vector unsigned char vec_vminub (vector unsigned char, | |
9813 | vector unsigned char); | |
9814 | ||
9815 | vector signed short vec_mladd (vector signed short, | |
9816 | vector signed short, | |
6e5bb5ad | 9817 | vector signed short); |
924fcc4e JM |
9818 | vector signed short vec_mladd (vector signed short, |
9819 | vector unsigned short, | |
6e5bb5ad | 9820 | vector unsigned short); |
924fcc4e JM |
9821 | vector signed short vec_mladd (vector unsigned short, |
9822 | vector signed short, | |
6e5bb5ad JM |
9823 | vector signed short); |
9824 | vector unsigned short vec_mladd (vector unsigned short, | |
9825 | vector unsigned short, | |
9826 | vector unsigned short); | |
9827 | ||
924fcc4e JM |
9828 | vector signed short vec_mradds (vector signed short, |
9829 | vector signed short, | |
6e5bb5ad JM |
9830 | vector signed short); |
9831 | ||
924fcc4e JM |
9832 | vector unsigned int vec_msum (vector unsigned char, |
9833 | vector unsigned char, | |
6e5bb5ad | 9834 | vector unsigned int); |
b0b343db JJ |
9835 | vector signed int vec_msum (vector signed char, |
9836 | vector unsigned char, | |
6e5bb5ad | 9837 | vector signed int); |
924fcc4e JM |
9838 | vector unsigned int vec_msum (vector unsigned short, |
9839 | vector unsigned short, | |
6e5bb5ad | 9840 | vector unsigned int); |
b0b343db JJ |
9841 | vector signed int vec_msum (vector signed short, |
9842 | vector signed short, | |
6e5bb5ad JM |
9843 | vector signed int); |
9844 | ||
b0b343db JJ |
9845 | vector signed int vec_vmsumshm (vector signed short, |
9846 | vector signed short, | |
9847 | vector signed int); | |
9848 | ||
9849 | vector unsigned int vec_vmsumuhm (vector unsigned short, | |
9850 | vector unsigned short, | |
9851 | vector unsigned int); | |
9852 | ||
9853 | vector signed int vec_vmsummbm (vector signed char, | |
9854 | vector unsigned char, | |
9855 | vector signed int); | |
9856 | ||
9857 | vector unsigned int vec_vmsumubm (vector unsigned char, | |
9858 | vector unsigned char, | |
9859 | vector unsigned int); | |
9860 | ||
6e5bb5ad | 9861 | vector unsigned int vec_msums (vector unsigned short, |
924fcc4e JM |
9862 | vector unsigned short, |
9863 | vector unsigned int); | |
b0b343db JJ |
9864 | vector signed int vec_msums (vector signed short, |
9865 | vector signed short, | |
6e5bb5ad | 9866 | vector signed int); |
333c8841 | 9867 | |
b0b343db JJ |
9868 | vector signed int vec_vmsumshs (vector signed short, |
9869 | vector signed short, | |
9870 | vector signed int); | |
9871 | ||
9872 | vector unsigned int vec_vmsumuhs (vector unsigned short, | |
9873 | vector unsigned short, | |
9874 | vector unsigned int); | |
9875 | ||
333c8841 AH |
9876 | void vec_mtvscr (vector signed int); |
9877 | void vec_mtvscr (vector unsigned int); | |
b0b343db | 9878 | void vec_mtvscr (vector bool int); |
333c8841 AH |
9879 | void vec_mtvscr (vector signed short); |
9880 | void vec_mtvscr (vector unsigned short); | |
b0b343db JJ |
9881 | void vec_mtvscr (vector bool short); |
9882 | void vec_mtvscr (vector pixel); | |
333c8841 AH |
9883 | void vec_mtvscr (vector signed char); |
9884 | void vec_mtvscr (vector unsigned char); | |
b0b343db | 9885 | void vec_mtvscr (vector bool char); |
333c8841 | 9886 | |
924fcc4e JM |
9887 | vector unsigned short vec_mule (vector unsigned char, |
9888 | vector unsigned char); | |
b0b343db JJ |
9889 | vector signed short vec_mule (vector signed char, |
9890 | vector signed char); | |
924fcc4e JM |
9891 | vector unsigned int vec_mule (vector unsigned short, |
9892 | vector unsigned short); | |
333c8841 AH |
9893 | vector signed int vec_mule (vector signed short, vector signed short); |
9894 | ||
b0b343db JJ |
9895 | vector signed int vec_vmulesh (vector signed short, |
9896 | vector signed short); | |
9897 | ||
9898 | vector unsigned int vec_vmuleuh (vector unsigned short, | |
9899 | vector unsigned short); | |
9900 | ||
9901 | vector signed short vec_vmulesb (vector signed char, | |
9902 | vector signed char); | |
9903 | ||
9904 | vector unsigned short vec_vmuleub (vector unsigned char, | |
9905 | vector unsigned char); | |
9906 | ||
924fcc4e JM |
9907 | vector unsigned short vec_mulo (vector unsigned char, |
9908 | vector unsigned char); | |
333c8841 | 9909 | vector signed short vec_mulo (vector signed char, vector signed char); |
924fcc4e JM |
9910 | vector unsigned int vec_mulo (vector unsigned short, |
9911 | vector unsigned short); | |
333c8841 AH |
9912 | vector signed int vec_mulo (vector signed short, vector signed short); |
9913 | ||
b0b343db JJ |
9914 | vector signed int vec_vmulosh (vector signed short, |
9915 | vector signed short); | |
9916 | ||
9917 | vector unsigned int vec_vmulouh (vector unsigned short, | |
9918 | vector unsigned short); | |
9919 | ||
9920 | vector signed short vec_vmulosb (vector signed char, | |
9921 | vector signed char); | |
9922 | ||
9923 | vector unsigned short vec_vmuloub (vector unsigned char, | |
9924 | vector unsigned char); | |
9925 | ||
333c8841 AH |
9926 | vector float vec_nmsub (vector float, vector float, vector float); |
9927 | ||
9928 | vector float vec_nor (vector float, vector float); | |
9929 | vector signed int vec_nor (vector signed int, vector signed int); | |
9930 | vector unsigned int vec_nor (vector unsigned int, vector unsigned int); | |
b0b343db | 9931 | vector bool int vec_nor (vector bool int, vector bool int); |
333c8841 | 9932 | vector signed short vec_nor (vector signed short, vector signed short); |
6e5bb5ad JM |
9933 | vector unsigned short vec_nor (vector unsigned short, |
9934 | vector unsigned short); | |
b0b343db | 9935 | vector bool short vec_nor (vector bool short, vector bool short); |
333c8841 | 9936 | vector signed char vec_nor (vector signed char, vector signed char); |
924fcc4e JM |
9937 | vector unsigned char vec_nor (vector unsigned char, |
9938 | vector unsigned char); | |
b0b343db | 9939 | vector bool char vec_nor (vector bool char, vector bool char); |
333c8841 AH |
9940 | |
9941 | vector float vec_or (vector float, vector float); | |
b0b343db JJ |
9942 | vector float vec_or (vector float, vector bool int); |
9943 | vector float vec_or (vector bool int, vector float); | |
9944 | vector bool int vec_or (vector bool int, vector bool int); | |
9945 | vector signed int vec_or (vector bool int, vector signed int); | |
9946 | vector signed int vec_or (vector signed int, vector bool int); | |
333c8841 | 9947 | vector signed int vec_or (vector signed int, vector signed int); |
b0b343db JJ |
9948 | vector unsigned int vec_or (vector bool int, vector unsigned int); |
9949 | vector unsigned int vec_or (vector unsigned int, vector bool int); | |
333c8841 | 9950 | vector unsigned int vec_or (vector unsigned int, vector unsigned int); |
b0b343db JJ |
9951 | vector bool short vec_or (vector bool short, vector bool short); |
9952 | vector signed short vec_or (vector bool short, vector signed short); | |
9953 | vector signed short vec_or (vector signed short, vector bool short); | |
333c8841 | 9954 | vector signed short vec_or (vector signed short, vector signed short); |
b0b343db JJ |
9955 | vector unsigned short vec_or (vector bool short, vector unsigned short); |
9956 | vector unsigned short vec_or (vector unsigned short, vector bool short); | |
924fcc4e JM |
9957 | vector unsigned short vec_or (vector unsigned short, |
9958 | vector unsigned short); | |
b0b343db JJ |
9959 | vector signed char vec_or (vector bool char, vector signed char); |
9960 | vector bool char vec_or (vector bool char, vector bool char); | |
9961 | vector signed char vec_or (vector signed char, vector bool char); | |
333c8841 | 9962 | vector signed char vec_or (vector signed char, vector signed char); |
b0b343db JJ |
9963 | vector unsigned char vec_or (vector bool char, vector unsigned char); |
9964 | vector unsigned char vec_or (vector unsigned char, vector bool char); | |
924fcc4e JM |
9965 | vector unsigned char vec_or (vector unsigned char, |
9966 | vector unsigned char); | |
333c8841 AH |
9967 | |
9968 | vector signed char vec_pack (vector signed short, vector signed short); | |
6e5bb5ad JM |
9969 | vector unsigned char vec_pack (vector unsigned short, |
9970 | vector unsigned short); | |
b0b343db | 9971 | vector bool char vec_pack (vector bool short, vector bool short); |
333c8841 | 9972 | vector signed short vec_pack (vector signed int, vector signed int); |
924fcc4e JM |
9973 | vector unsigned short vec_pack (vector unsigned int, |
9974 | vector unsigned int); | |
b0b343db | 9975 | vector bool short vec_pack (vector bool int, vector bool int); |
333c8841 | 9976 | |
b0b343db JJ |
9977 | vector bool short vec_vpkuwum (vector bool int, vector bool int); |
9978 | vector signed short vec_vpkuwum (vector signed int, vector signed int); | |
9979 | vector unsigned short vec_vpkuwum (vector unsigned int, | |
9980 | vector unsigned int); | |
9981 | ||
9982 | vector bool char vec_vpkuhum (vector bool short, vector bool short); | |
9983 | vector signed char vec_vpkuhum (vector signed short, | |
9984 | vector signed short); | |
9985 | vector unsigned char vec_vpkuhum (vector unsigned short, | |
9986 | vector unsigned short); | |
9987 | ||
9988 | vector pixel vec_packpx (vector unsigned int, vector unsigned int); | |
333c8841 | 9989 | |
6e5bb5ad JM |
9990 | vector unsigned char vec_packs (vector unsigned short, |
9991 | vector unsigned short); | |
333c8841 | 9992 | vector signed char vec_packs (vector signed short, vector signed short); |
924fcc4e JM |
9993 | vector unsigned short vec_packs (vector unsigned int, |
9994 | vector unsigned int); | |
333c8841 AH |
9995 | vector signed short vec_packs (vector signed int, vector signed int); |
9996 | ||
b0b343db JJ |
9997 | vector signed short vec_vpkswss (vector signed int, vector signed int); |
9998 | ||
9999 | vector unsigned short vec_vpkuwus (vector unsigned int, | |
10000 | vector unsigned int); | |
10001 | ||
10002 | vector signed char vec_vpkshss (vector signed short, | |
10003 | vector signed short); | |
10004 | ||
10005 | vector unsigned char vec_vpkuhus (vector unsigned short, | |
10006 | vector unsigned short); | |
10007 | ||
6e5bb5ad JM |
10008 | vector unsigned char vec_packsu (vector unsigned short, |
10009 | vector unsigned short); | |
924fcc4e JM |
10010 | vector unsigned char vec_packsu (vector signed short, |
10011 | vector signed short); | |
10012 | vector unsigned short vec_packsu (vector unsigned int, | |
10013 | vector unsigned int); | |
333c8841 AH |
10014 | vector unsigned short vec_packsu (vector signed int, vector signed int); |
10015 | ||
b0b343db JJ |
10016 | vector unsigned short vec_vpkswus (vector signed int, |
10017 | vector signed int); | |
10018 | ||
10019 | vector unsigned char vec_vpkshus (vector signed short, | |
10020 | vector signed short); | |
10021 | ||
10022 | vector float vec_perm (vector float, | |
10023 | vector float, | |
924fcc4e | 10024 | vector unsigned char); |
b0b343db JJ |
10025 | vector signed int vec_perm (vector signed int, |
10026 | vector signed int, | |
6e5bb5ad | 10027 | vector unsigned char); |
b0b343db JJ |
10028 | vector unsigned int vec_perm (vector unsigned int, |
10029 | vector unsigned int, | |
6e5bb5ad | 10030 | vector unsigned char); |
b0b343db JJ |
10031 | vector bool int vec_perm (vector bool int, |
10032 | vector bool int, | |
10033 | vector unsigned char); | |
10034 | vector signed short vec_perm (vector signed short, | |
10035 | vector signed short, | |
6e5bb5ad JM |
10036 | vector unsigned char); |
10037 | vector unsigned short vec_perm (vector unsigned short, | |
10038 | vector unsigned short, | |
10039 | vector unsigned char); | |
b0b343db JJ |
10040 | vector bool short vec_perm (vector bool short, |
10041 | vector bool short, | |
10042 | vector unsigned char); | |
10043 | vector pixel vec_perm (vector pixel, | |
10044 | vector pixel, | |
10045 | vector unsigned char); | |
10046 | vector signed char vec_perm (vector signed char, | |
10047 | vector signed char, | |
6e5bb5ad | 10048 | vector unsigned char); |
924fcc4e JM |
10049 | vector unsigned char vec_perm (vector unsigned char, |
10050 | vector unsigned char, | |
6e5bb5ad | 10051 | vector unsigned char); |
b0b343db JJ |
10052 | vector bool char vec_perm (vector bool char, |
10053 | vector bool char, | |
10054 | vector unsigned char); | |
333c8841 AH |
10055 | |
10056 | vector float vec_re (vector float); | |
10057 | ||
b0b343db JJ |
10058 | vector signed char vec_rl (vector signed char, |
10059 | vector unsigned char); | |
924fcc4e JM |
10060 | vector unsigned char vec_rl (vector unsigned char, |
10061 | vector unsigned char); | |
333c8841 | 10062 | vector signed short vec_rl (vector signed short, vector unsigned short); |
924fcc4e JM |
10063 | vector unsigned short vec_rl (vector unsigned short, |
10064 | vector unsigned short); | |
333c8841 AH |
10065 | vector signed int vec_rl (vector signed int, vector unsigned int); |
10066 | vector unsigned int vec_rl (vector unsigned int, vector unsigned int); | |
10067 | ||
b0b343db JJ |
10068 | vector signed int vec_vrlw (vector signed int, vector unsigned int); |
10069 | vector unsigned int vec_vrlw (vector unsigned int, vector unsigned int); | |
10070 | ||
10071 | vector signed short vec_vrlh (vector signed short, | |
10072 | vector unsigned short); | |
10073 | vector unsigned short vec_vrlh (vector unsigned short, | |
10074 | vector unsigned short); | |
10075 | ||
10076 | vector signed char vec_vrlb (vector signed char, vector unsigned char); | |
10077 | vector unsigned char vec_vrlb (vector unsigned char, | |
10078 | vector unsigned char); | |
10079 | ||
333c8841 AH |
10080 | vector float vec_round (vector float); |
10081 | ||
10082 | vector float vec_rsqrte (vector float); | |
10083 | ||
b0b343db | 10084 | vector float vec_sel (vector float, vector float, vector bool int); |
333c8841 | 10085 | vector float vec_sel (vector float, vector float, vector unsigned int); |
b0b343db JJ |
10086 | vector signed int vec_sel (vector signed int, |
10087 | vector signed int, | |
10088 | vector bool int); | |
10089 | vector signed int vec_sel (vector signed int, | |
10090 | vector signed int, | |
6e5bb5ad | 10091 | vector unsigned int); |
b0b343db JJ |
10092 | vector unsigned int vec_sel (vector unsigned int, |
10093 | vector unsigned int, | |
10094 | vector bool int); | |
10095 | vector unsigned int vec_sel (vector unsigned int, | |
10096 | vector unsigned int, | |
6e5bb5ad | 10097 | vector unsigned int); |
b0b343db JJ |
10098 | vector bool int vec_sel (vector bool int, |
10099 | vector bool int, | |
10100 | vector bool int); | |
10101 | vector bool int vec_sel (vector bool int, | |
10102 | vector bool int, | |
10103 | vector unsigned int); | |
10104 | vector signed short vec_sel (vector signed short, | |
10105 | vector signed short, | |
10106 | vector bool short); | |
10107 | vector signed short vec_sel (vector signed short, | |
10108 | vector signed short, | |
6e5bb5ad JM |
10109 | vector unsigned short); |
10110 | vector unsigned short vec_sel (vector unsigned short, | |
924fcc4e | 10111 | vector unsigned short, |
b0b343db | 10112 | vector bool short); |
6e5bb5ad JM |
10113 | vector unsigned short vec_sel (vector unsigned short, |
10114 | vector unsigned short, | |
10115 | vector unsigned short); | |
b0b343db JJ |
10116 | vector bool short vec_sel (vector bool short, |
10117 | vector bool short, | |
10118 | vector bool short); | |
10119 | vector bool short vec_sel (vector bool short, | |
10120 | vector bool short, | |
10121 | vector unsigned short); | |
10122 | vector signed char vec_sel (vector signed char, | |
10123 | vector signed char, | |
10124 | vector bool char); | |
10125 | vector signed char vec_sel (vector signed char, | |
10126 | vector signed char, | |
6e5bb5ad | 10127 | vector unsigned char); |
924fcc4e JM |
10128 | vector unsigned char vec_sel (vector unsigned char, |
10129 | vector unsigned char, | |
b0b343db | 10130 | vector bool char); |
924fcc4e JM |
10131 | vector unsigned char vec_sel (vector unsigned char, |
10132 | vector unsigned char, | |
6e5bb5ad | 10133 | vector unsigned char); |
b0b343db JJ |
10134 | vector bool char vec_sel (vector bool char, |
10135 | vector bool char, | |
10136 | vector bool char); | |
10137 | vector bool char vec_sel (vector bool char, | |
10138 | vector bool char, | |
10139 | vector unsigned char); | |
10140 | ||
10141 | vector signed char vec_sl (vector signed char, | |
10142 | vector unsigned char); | |
924fcc4e JM |
10143 | vector unsigned char vec_sl (vector unsigned char, |
10144 | vector unsigned char); | |
333c8841 | 10145 | vector signed short vec_sl (vector signed short, vector unsigned short); |
924fcc4e JM |
10146 | vector unsigned short vec_sl (vector unsigned short, |
10147 | vector unsigned short); | |
333c8841 AH |
10148 | vector signed int vec_sl (vector signed int, vector unsigned int); |
10149 | vector unsigned int vec_sl (vector unsigned int, vector unsigned int); | |
10150 | ||
b0b343db JJ |
10151 | vector signed int vec_vslw (vector signed int, vector unsigned int); |
10152 | vector unsigned int vec_vslw (vector unsigned int, vector unsigned int); | |
10153 | ||
10154 | vector signed short vec_vslh (vector signed short, | |
10155 | vector unsigned short); | |
10156 | vector unsigned short vec_vslh (vector unsigned short, | |
10157 | vector unsigned short); | |
10158 | ||
10159 | vector signed char vec_vslb (vector signed char, vector unsigned char); | |
10160 | vector unsigned char vec_vslb (vector unsigned char, | |
10161 | vector unsigned char); | |
10162 | ||
10163 | vector float vec_sld (vector float, vector float, const int); | |
10164 | vector signed int vec_sld (vector signed int, | |
10165 | vector signed int, | |
10166 | const int); | |
10167 | vector unsigned int vec_sld (vector unsigned int, | |
10168 | vector unsigned int, | |
10169 | const int); | |
10170 | vector bool int vec_sld (vector bool int, | |
10171 | vector bool int, | |
10172 | const int); | |
10173 | vector signed short vec_sld (vector signed short, | |
10174 | vector signed short, | |
10175 | const int); | |
6e5bb5ad | 10176 | vector unsigned short vec_sld (vector unsigned short, |
b0b343db JJ |
10177 | vector unsigned short, |
10178 | const int); | |
10179 | vector bool short vec_sld (vector bool short, | |
10180 | vector bool short, | |
10181 | const int); | |
10182 | vector pixel vec_sld (vector pixel, | |
10183 | vector pixel, | |
10184 | const int); | |
10185 | vector signed char vec_sld (vector signed char, | |
10186 | vector signed char, | |
10187 | const int); | |
924fcc4e JM |
10188 | vector unsigned char vec_sld (vector unsigned char, |
10189 | vector unsigned char, | |
b0b343db JJ |
10190 | const int); |
10191 | vector bool char vec_sld (vector bool char, | |
10192 | vector bool char, | |
10193 | const int); | |
333c8841 | 10194 | |
b0b343db JJ |
10195 | vector signed int vec_sll (vector signed int, |
10196 | vector unsigned int); | |
10197 | vector signed int vec_sll (vector signed int, | |
10198 | vector unsigned short); | |
10199 | vector signed int vec_sll (vector signed int, | |
10200 | vector unsigned char); | |
10201 | vector unsigned int vec_sll (vector unsigned int, | |
10202 | vector unsigned int); | |
924fcc4e JM |
10203 | vector unsigned int vec_sll (vector unsigned int, |
10204 | vector unsigned short); | |
b0b343db JJ |
10205 | vector unsigned int vec_sll (vector unsigned int, |
10206 | vector unsigned char); | |
10207 | vector bool int vec_sll (vector bool int, | |
10208 | vector unsigned int); | |
10209 | vector bool int vec_sll (vector bool int, | |
10210 | vector unsigned short); | |
10211 | vector bool int vec_sll (vector bool int, | |
10212 | vector unsigned char); | |
10213 | vector signed short vec_sll (vector signed short, | |
10214 | vector unsigned int); | |
924fcc4e JM |
10215 | vector signed short vec_sll (vector signed short, |
10216 | vector unsigned short); | |
b0b343db JJ |
10217 | vector signed short vec_sll (vector signed short, |
10218 | vector unsigned char); | |
924fcc4e JM |
10219 | vector unsigned short vec_sll (vector unsigned short, |
10220 | vector unsigned int); | |
6e5bb5ad JM |
10221 | vector unsigned short vec_sll (vector unsigned short, |
10222 | vector unsigned short); | |
924fcc4e JM |
10223 | vector unsigned short vec_sll (vector unsigned short, |
10224 | vector unsigned char); | |
b0b343db JJ |
10225 | vector bool short vec_sll (vector bool short, vector unsigned int); |
10226 | vector bool short vec_sll (vector bool short, vector unsigned short); | |
10227 | vector bool short vec_sll (vector bool short, vector unsigned char); | |
10228 | vector pixel vec_sll (vector pixel, vector unsigned int); | |
10229 | vector pixel vec_sll (vector pixel, vector unsigned short); | |
10230 | vector pixel vec_sll (vector pixel, vector unsigned char); | |
333c8841 AH |
10231 | vector signed char vec_sll (vector signed char, vector unsigned int); |
10232 | vector signed char vec_sll (vector signed char, vector unsigned short); | |
10233 | vector signed char vec_sll (vector signed char, vector unsigned char); | |
924fcc4e JM |
10234 | vector unsigned char vec_sll (vector unsigned char, |
10235 | vector unsigned int); | |
10236 | vector unsigned char vec_sll (vector unsigned char, | |
10237 | vector unsigned short); | |
10238 | vector unsigned char vec_sll (vector unsigned char, | |
10239 | vector unsigned char); | |
b0b343db JJ |
10240 | vector bool char vec_sll (vector bool char, vector unsigned int); |
10241 | vector bool char vec_sll (vector bool char, vector unsigned short); | |
10242 | vector bool char vec_sll (vector bool char, vector unsigned char); | |
333c8841 AH |
10243 | |
10244 | vector float vec_slo (vector float, vector signed char); | |
10245 | vector float vec_slo (vector float, vector unsigned char); | |
10246 | vector signed int vec_slo (vector signed int, vector signed char); | |
10247 | vector signed int vec_slo (vector signed int, vector unsigned char); | |
10248 | vector unsigned int vec_slo (vector unsigned int, vector signed char); | |
10249 | vector unsigned int vec_slo (vector unsigned int, vector unsigned char); | |
333c8841 AH |
10250 | vector signed short vec_slo (vector signed short, vector signed char); |
10251 | vector signed short vec_slo (vector signed short, vector unsigned char); | |
924fcc4e JM |
10252 | vector unsigned short vec_slo (vector unsigned short, |
10253 | vector signed char); | |
10254 | vector unsigned short vec_slo (vector unsigned short, | |
10255 | vector unsigned char); | |
b0b343db JJ |
10256 | vector pixel vec_slo (vector pixel, vector signed char); |
10257 | vector pixel vec_slo (vector pixel, vector unsigned char); | |
333c8841 AH |
10258 | vector signed char vec_slo (vector signed char, vector signed char); |
10259 | vector signed char vec_slo (vector signed char, vector unsigned char); | |
10260 | vector unsigned char vec_slo (vector unsigned char, vector signed char); | |
924fcc4e JM |
10261 | vector unsigned char vec_slo (vector unsigned char, |
10262 | vector unsigned char); | |
333c8841 | 10263 | |
b0b343db JJ |
10264 | vector signed char vec_splat (vector signed char, const int); |
10265 | vector unsigned char vec_splat (vector unsigned char, const int); | |
10266 | vector bool char vec_splat (vector bool char, const int); | |
10267 | vector signed short vec_splat (vector signed short, const int); | |
10268 | vector unsigned short vec_splat (vector unsigned short, const int); | |
10269 | vector bool short vec_splat (vector bool short, const int); | |
10270 | vector pixel vec_splat (vector pixel, const int); | |
10271 | vector float vec_splat (vector float, const int); | |
10272 | vector signed int vec_splat (vector signed int, const int); | |
10273 | vector unsigned int vec_splat (vector unsigned int, const int); | |
10274 | vector bool int vec_splat (vector bool int, const int); | |
10275 | ||
10276 | vector float vec_vspltw (vector float, const int); | |
10277 | vector signed int vec_vspltw (vector signed int, const int); | |
10278 | vector unsigned int vec_vspltw (vector unsigned int, const int); | |
10279 | vector bool int vec_vspltw (vector bool int, const int); | |
10280 | ||
10281 | vector bool short vec_vsplth (vector bool short, const int); | |
10282 | vector signed short vec_vsplth (vector signed short, const int); | |
10283 | vector unsigned short vec_vsplth (vector unsigned short, const int); | |
10284 | vector pixel vec_vsplth (vector pixel, const int); | |
10285 | ||
10286 | vector signed char vec_vspltb (vector signed char, const int); | |
10287 | vector unsigned char vec_vspltb (vector unsigned char, const int); | |
10288 | vector bool char vec_vspltb (vector bool char, const int); | |
333c8841 | 10289 | |
b0b343db | 10290 | vector signed char vec_splat_s8 (const int); |
333c8841 | 10291 | |
b0b343db | 10292 | vector signed short vec_splat_s16 (const int); |
333c8841 | 10293 | |
b0b343db | 10294 | vector signed int vec_splat_s32 (const int); |
333c8841 | 10295 | |
b0b343db | 10296 | vector unsigned char vec_splat_u8 (const int); |
333c8841 | 10297 | |
b0b343db | 10298 | vector unsigned short vec_splat_u16 (const int); |
333c8841 | 10299 | |
b0b343db | 10300 | vector unsigned int vec_splat_u32 (const int); |
333c8841 AH |
10301 | |
10302 | vector signed char vec_sr (vector signed char, vector unsigned char); | |
924fcc4e JM |
10303 | vector unsigned char vec_sr (vector unsigned char, |
10304 | vector unsigned char); | |
b0b343db JJ |
10305 | vector signed short vec_sr (vector signed short, |
10306 | vector unsigned short); | |
924fcc4e JM |
10307 | vector unsigned short vec_sr (vector unsigned short, |
10308 | vector unsigned short); | |
333c8841 AH |
10309 | vector signed int vec_sr (vector signed int, vector unsigned int); |
10310 | vector unsigned int vec_sr (vector unsigned int, vector unsigned int); | |
10311 | ||
b0b343db JJ |
10312 | vector signed int vec_vsrw (vector signed int, vector unsigned int); |
10313 | vector unsigned int vec_vsrw (vector unsigned int, vector unsigned int); | |
10314 | ||
10315 | vector signed short vec_vsrh (vector signed short, | |
10316 | vector unsigned short); | |
10317 | vector unsigned short vec_vsrh (vector unsigned short, | |
10318 | vector unsigned short); | |
10319 | ||
10320 | vector signed char vec_vsrb (vector signed char, vector unsigned char); | |
10321 | vector unsigned char vec_vsrb (vector unsigned char, | |
10322 | vector unsigned char); | |
10323 | ||
333c8841 | 10324 | vector signed char vec_sra (vector signed char, vector unsigned char); |
924fcc4e JM |
10325 | vector unsigned char vec_sra (vector unsigned char, |
10326 | vector unsigned char); | |
10327 | vector signed short vec_sra (vector signed short, | |
10328 | vector unsigned short); | |
6e5bb5ad JM |
10329 | vector unsigned short vec_sra (vector unsigned short, |
10330 | vector unsigned short); | |
333c8841 AH |
10331 | vector signed int vec_sra (vector signed int, vector unsigned int); |
10332 | vector unsigned int vec_sra (vector unsigned int, vector unsigned int); | |
10333 | ||
b0b343db JJ |
10334 | vector signed int vec_vsraw (vector signed int, vector unsigned int); |
10335 | vector unsigned int vec_vsraw (vector unsigned int, | |
10336 | vector unsigned int); | |
10337 | ||
10338 | vector signed short vec_vsrah (vector signed short, | |
10339 | vector unsigned short); | |
10340 | vector unsigned short vec_vsrah (vector unsigned short, | |
10341 | vector unsigned short); | |
10342 | ||
10343 | vector signed char vec_vsrab (vector signed char, vector unsigned char); | |
10344 | vector unsigned char vec_vsrab (vector unsigned char, | |
10345 | vector unsigned char); | |
10346 | ||
333c8841 AH |
10347 | vector signed int vec_srl (vector signed int, vector unsigned int); |
10348 | vector signed int vec_srl (vector signed int, vector unsigned short); | |
10349 | vector signed int vec_srl (vector signed int, vector unsigned char); | |
10350 | vector unsigned int vec_srl (vector unsigned int, vector unsigned int); | |
924fcc4e JM |
10351 | vector unsigned int vec_srl (vector unsigned int, |
10352 | vector unsigned short); | |
333c8841 | 10353 | vector unsigned int vec_srl (vector unsigned int, vector unsigned char); |
b0b343db JJ |
10354 | vector bool int vec_srl (vector bool int, vector unsigned int); |
10355 | vector bool int vec_srl (vector bool int, vector unsigned short); | |
10356 | vector bool int vec_srl (vector bool int, vector unsigned char); | |
333c8841 | 10357 | vector signed short vec_srl (vector signed short, vector unsigned int); |
924fcc4e JM |
10358 | vector signed short vec_srl (vector signed short, |
10359 | vector unsigned short); | |
333c8841 | 10360 | vector signed short vec_srl (vector signed short, vector unsigned char); |
924fcc4e JM |
10361 | vector unsigned short vec_srl (vector unsigned short, |
10362 | vector unsigned int); | |
6e5bb5ad JM |
10363 | vector unsigned short vec_srl (vector unsigned short, |
10364 | vector unsigned short); | |
924fcc4e JM |
10365 | vector unsigned short vec_srl (vector unsigned short, |
10366 | vector unsigned char); | |
b0b343db JJ |
10367 | vector bool short vec_srl (vector bool short, vector unsigned int); |
10368 | vector bool short vec_srl (vector bool short, vector unsigned short); | |
10369 | vector bool short vec_srl (vector bool short, vector unsigned char); | |
10370 | vector pixel vec_srl (vector pixel, vector unsigned int); | |
10371 | vector pixel vec_srl (vector pixel, vector unsigned short); | |
10372 | vector pixel vec_srl (vector pixel, vector unsigned char); | |
333c8841 AH |
10373 | vector signed char vec_srl (vector signed char, vector unsigned int); |
10374 | vector signed char vec_srl (vector signed char, vector unsigned short); | |
10375 | vector signed char vec_srl (vector signed char, vector unsigned char); | |
924fcc4e JM |
10376 | vector unsigned char vec_srl (vector unsigned char, |
10377 | vector unsigned int); | |
10378 | vector unsigned char vec_srl (vector unsigned char, | |
10379 | vector unsigned short); | |
10380 | vector unsigned char vec_srl (vector unsigned char, | |
10381 | vector unsigned char); | |
b0b343db JJ |
10382 | vector bool char vec_srl (vector bool char, vector unsigned int); |
10383 | vector bool char vec_srl (vector bool char, vector unsigned short); | |
10384 | vector bool char vec_srl (vector bool char, vector unsigned char); | |
333c8841 AH |
10385 | |
10386 | vector float vec_sro (vector float, vector signed char); | |
10387 | vector float vec_sro (vector float, vector unsigned char); | |
10388 | vector signed int vec_sro (vector signed int, vector signed char); | |
10389 | vector signed int vec_sro (vector signed int, vector unsigned char); | |
10390 | vector unsigned int vec_sro (vector unsigned int, vector signed char); | |
10391 | vector unsigned int vec_sro (vector unsigned int, vector unsigned char); | |
333c8841 AH |
10392 | vector signed short vec_sro (vector signed short, vector signed char); |
10393 | vector signed short vec_sro (vector signed short, vector unsigned char); | |
924fcc4e JM |
10394 | vector unsigned short vec_sro (vector unsigned short, |
10395 | vector signed char); | |
10396 | vector unsigned short vec_sro (vector unsigned short, | |
10397 | vector unsigned char); | |
b0b343db JJ |
10398 | vector pixel vec_sro (vector pixel, vector signed char); |
10399 | vector pixel vec_sro (vector pixel, vector unsigned char); | |
333c8841 AH |
10400 | vector signed char vec_sro (vector signed char, vector signed char); |
10401 | vector signed char vec_sro (vector signed char, vector unsigned char); | |
10402 | vector unsigned char vec_sro (vector unsigned char, vector signed char); | |
924fcc4e JM |
10403 | vector unsigned char vec_sro (vector unsigned char, |
10404 | vector unsigned char); | |
333c8841 | 10405 | |
333c8841 | 10406 | void vec_st (vector float, int, vector float *); |
b0b343db JJ |
10407 | void vec_st (vector float, int, float *); |
10408 | void vec_st (vector signed int, int, vector signed int *); | |
333c8841 | 10409 | void vec_st (vector signed int, int, int *); |
333c8841 | 10410 | void vec_st (vector unsigned int, int, vector unsigned int *); |
b0b343db JJ |
10411 | void vec_st (vector unsigned int, int, unsigned int *); |
10412 | void vec_st (vector bool int, int, vector bool int *); | |
10413 | void vec_st (vector bool int, int, unsigned int *); | |
10414 | void vec_st (vector bool int, int, int *); | |
333c8841 | 10415 | void vec_st (vector signed short, int, vector signed short *); |
b0b343db | 10416 | void vec_st (vector signed short, int, short *); |
333c8841 | 10417 | void vec_st (vector unsigned short, int, vector unsigned short *); |
b0b343db JJ |
10418 | void vec_st (vector unsigned short, int, unsigned short *); |
10419 | void vec_st (vector bool short, int, vector bool short *); | |
10420 | void vec_st (vector bool short, int, unsigned short *); | |
10421 | void vec_st (vector pixel, int, vector pixel *); | |
10422 | void vec_st (vector pixel, int, unsigned short *); | |
10423 | void vec_st (vector pixel, int, short *); | |
10424 | void vec_st (vector bool short, int, short *); | |
333c8841 | 10425 | void vec_st (vector signed char, int, vector signed char *); |
b0b343db | 10426 | void vec_st (vector signed char, int, signed char *); |
333c8841 | 10427 | void vec_st (vector unsigned char, int, vector unsigned char *); |
b0b343db JJ |
10428 | void vec_st (vector unsigned char, int, unsigned char *); |
10429 | void vec_st (vector bool char, int, vector bool char *); | |
10430 | void vec_st (vector bool char, int, unsigned char *); | |
10431 | void vec_st (vector bool char, int, signed char *); | |
333c8841 | 10432 | |
333c8841 AH |
10433 | void vec_ste (vector signed char, int, signed char *); |
10434 | void vec_ste (vector unsigned char, int, unsigned char *); | |
b0b343db JJ |
10435 | void vec_ste (vector bool char, int, signed char *); |
10436 | void vec_ste (vector bool char, int, unsigned char *); | |
333c8841 | 10437 | void vec_ste (vector signed short, int, short *); |
b0b343db JJ |
10438 | void vec_ste (vector unsigned short, int, unsigned short *); |
10439 | void vec_ste (vector bool short, int, short *); | |
10440 | void vec_ste (vector bool short, int, unsigned short *); | |
10441 | void vec_ste (vector pixel, int, short *); | |
10442 | void vec_ste (vector pixel, int, unsigned short *); | |
10443 | void vec_ste (vector float, int, float *); | |
333c8841 AH |
10444 | void vec_ste (vector signed int, int, int *); |
10445 | void vec_ste (vector unsigned int, int, unsigned int *); | |
b0b343db JJ |
10446 | void vec_ste (vector bool int, int, int *); |
10447 | void vec_ste (vector bool int, int, unsigned int *); | |
10448 | ||
10449 | void vec_stvewx (vector float, int, float *); | |
10450 | void vec_stvewx (vector signed int, int, int *); | |
10451 | void vec_stvewx (vector unsigned int, int, unsigned int *); | |
10452 | void vec_stvewx (vector bool int, int, int *); | |
10453 | void vec_stvewx (vector bool int, int, unsigned int *); | |
10454 | ||
10455 | void vec_stvehx (vector signed short, int, short *); | |
10456 | void vec_stvehx (vector unsigned short, int, unsigned short *); | |
10457 | void vec_stvehx (vector bool short, int, short *); | |
10458 | void vec_stvehx (vector bool short, int, unsigned short *); | |
10459 | void vec_stvehx (vector pixel, int, short *); | |
10460 | void vec_stvehx (vector pixel, int, unsigned short *); | |
10461 | ||
10462 | void vec_stvebx (vector signed char, int, signed char *); | |
10463 | void vec_stvebx (vector unsigned char, int, unsigned char *); | |
10464 | void vec_stvebx (vector bool char, int, signed char *); | |
10465 | void vec_stvebx (vector bool char, int, unsigned char *); | |
333c8841 AH |
10466 | |
10467 | void vec_stl (vector float, int, vector float *); | |
10468 | void vec_stl (vector float, int, float *); | |
10469 | void vec_stl (vector signed int, int, vector signed int *); | |
10470 | void vec_stl (vector signed int, int, int *); | |
333c8841 AH |
10471 | void vec_stl (vector unsigned int, int, vector unsigned int *); |
10472 | void vec_stl (vector unsigned int, int, unsigned int *); | |
b0b343db JJ |
10473 | void vec_stl (vector bool int, int, vector bool int *); |
10474 | void vec_stl (vector bool int, int, unsigned int *); | |
10475 | void vec_stl (vector bool int, int, int *); | |
333c8841 | 10476 | void vec_stl (vector signed short, int, vector signed short *); |
b0b343db JJ |
10477 | void vec_stl (vector signed short, int, short *); |
10478 | void vec_stl (vector unsigned short, int, vector unsigned short *); | |
333c8841 | 10479 | void vec_stl (vector unsigned short, int, unsigned short *); |
b0b343db JJ |
10480 | void vec_stl (vector bool short, int, vector bool short *); |
10481 | void vec_stl (vector bool short, int, unsigned short *); | |
10482 | void vec_stl (vector bool short, int, short *); | |
10483 | void vec_stl (vector pixel, int, vector pixel *); | |
10484 | void vec_stl (vector pixel, int, unsigned short *); | |
10485 | void vec_stl (vector pixel, int, short *); | |
333c8841 | 10486 | void vec_stl (vector signed char, int, vector signed char *); |
b0b343db | 10487 | void vec_stl (vector signed char, int, signed char *); |
333c8841 | 10488 | void vec_stl (vector unsigned char, int, vector unsigned char *); |
b0b343db JJ |
10489 | void vec_stl (vector unsigned char, int, unsigned char *); |
10490 | void vec_stl (vector bool char, int, vector bool char *); | |
10491 | void vec_stl (vector bool char, int, unsigned char *); | |
10492 | void vec_stl (vector bool char, int, signed char *); | |
333c8841 | 10493 | |
b0b343db JJ |
10494 | vector signed char vec_sub (vector bool char, vector signed char); |
10495 | vector signed char vec_sub (vector signed char, vector bool char); | |
333c8841 | 10496 | vector signed char vec_sub (vector signed char, vector signed char); |
b0b343db JJ |
10497 | vector unsigned char vec_sub (vector bool char, vector unsigned char); |
10498 | vector unsigned char vec_sub (vector unsigned char, vector bool char); | |
924fcc4e JM |
10499 | vector unsigned char vec_sub (vector unsigned char, |
10500 | vector unsigned char); | |
b0b343db JJ |
10501 | vector signed short vec_sub (vector bool short, vector signed short); |
10502 | vector signed short vec_sub (vector signed short, vector bool short); | |
333c8841 | 10503 | vector signed short vec_sub (vector signed short, vector signed short); |
b0b343db | 10504 | vector unsigned short vec_sub (vector bool short, |
924fcc4e JM |
10505 | vector unsigned short); |
10506 | vector unsigned short vec_sub (vector unsigned short, | |
b0b343db | 10507 | vector bool short); |
6e5bb5ad JM |
10508 | vector unsigned short vec_sub (vector unsigned short, |
10509 | vector unsigned short); | |
b0b343db JJ |
10510 | vector signed int vec_sub (vector bool int, vector signed int); |
10511 | vector signed int vec_sub (vector signed int, vector bool int); | |
333c8841 | 10512 | vector signed int vec_sub (vector signed int, vector signed int); |
b0b343db JJ |
10513 | vector unsigned int vec_sub (vector bool int, vector unsigned int); |
10514 | vector unsigned int vec_sub (vector unsigned int, vector bool int); | |
333c8841 AH |
10515 | vector unsigned int vec_sub (vector unsigned int, vector unsigned int); |
10516 | vector float vec_sub (vector float, vector float); | |
10517 | ||
b0b343db JJ |
10518 | vector float vec_vsubfp (vector float, vector float); |
10519 | ||
10520 | vector signed int vec_vsubuwm (vector bool int, vector signed int); | |
10521 | vector signed int vec_vsubuwm (vector signed int, vector bool int); | |
10522 | vector signed int vec_vsubuwm (vector signed int, vector signed int); | |
10523 | vector unsigned int vec_vsubuwm (vector bool int, vector unsigned int); | |
10524 | vector unsigned int vec_vsubuwm (vector unsigned int, vector bool int); | |
10525 | vector unsigned int vec_vsubuwm (vector unsigned int, | |
10526 | vector unsigned int); | |
10527 | ||
10528 | vector signed short vec_vsubuhm (vector bool short, | |
10529 | vector signed short); | |
10530 | vector signed short vec_vsubuhm (vector signed short, | |
10531 | vector bool short); | |
10532 | vector signed short vec_vsubuhm (vector signed short, | |
10533 | vector signed short); | |
10534 | vector unsigned short vec_vsubuhm (vector bool short, | |
10535 | vector unsigned short); | |
10536 | vector unsigned short vec_vsubuhm (vector unsigned short, | |
10537 | vector bool short); | |
10538 | vector unsigned short vec_vsubuhm (vector unsigned short, | |
10539 | vector unsigned short); | |
10540 | ||
10541 | vector signed char vec_vsububm (vector bool char, vector signed char); | |
10542 | vector signed char vec_vsububm (vector signed char, vector bool char); | |
10543 | vector signed char vec_vsububm (vector signed char, vector signed char); | |
10544 | vector unsigned char vec_vsububm (vector bool char, | |
10545 | vector unsigned char); | |
10546 | vector unsigned char vec_vsububm (vector unsigned char, | |
10547 | vector bool char); | |
10548 | vector unsigned char vec_vsububm (vector unsigned char, | |
10549 | vector unsigned char); | |
10550 | ||
333c8841 AH |
10551 | vector unsigned int vec_subc (vector unsigned int, vector unsigned int); |
10552 | ||
b0b343db JJ |
10553 | vector unsigned char vec_subs (vector bool char, vector unsigned char); |
10554 | vector unsigned char vec_subs (vector unsigned char, vector bool char); | |
924fcc4e JM |
10555 | vector unsigned char vec_subs (vector unsigned char, |
10556 | vector unsigned char); | |
b0b343db JJ |
10557 | vector signed char vec_subs (vector bool char, vector signed char); |
10558 | vector signed char vec_subs (vector signed char, vector bool char); | |
333c8841 | 10559 | vector signed char vec_subs (vector signed char, vector signed char); |
b0b343db | 10560 | vector unsigned short vec_subs (vector bool short, |
924fcc4e JM |
10561 | vector unsigned short); |
10562 | vector unsigned short vec_subs (vector unsigned short, | |
b0b343db | 10563 | vector bool short); |
6e5bb5ad JM |
10564 | vector unsigned short vec_subs (vector unsigned short, |
10565 | vector unsigned short); | |
b0b343db JJ |
10566 | vector signed short vec_subs (vector bool short, vector signed short); |
10567 | vector signed short vec_subs (vector signed short, vector bool short); | |
333c8841 | 10568 | vector signed short vec_subs (vector signed short, vector signed short); |
b0b343db JJ |
10569 | vector unsigned int vec_subs (vector bool int, vector unsigned int); |
10570 | vector unsigned int vec_subs (vector unsigned int, vector bool int); | |
333c8841 | 10571 | vector unsigned int vec_subs (vector unsigned int, vector unsigned int); |
b0b343db JJ |
10572 | vector signed int vec_subs (vector bool int, vector signed int); |
10573 | vector signed int vec_subs (vector signed int, vector bool int); | |
333c8841 AH |
10574 | vector signed int vec_subs (vector signed int, vector signed int); |
10575 | ||
b0b343db JJ |
10576 | vector signed int vec_vsubsws (vector bool int, vector signed int); |
10577 | vector signed int vec_vsubsws (vector signed int, vector bool int); | |
10578 | vector signed int vec_vsubsws (vector signed int, vector signed int); | |
10579 | ||
10580 | vector unsigned int vec_vsubuws (vector bool int, vector unsigned int); | |
10581 | vector unsigned int vec_vsubuws (vector unsigned int, vector bool int); | |
10582 | vector unsigned int vec_vsubuws (vector unsigned int, | |
10583 | vector unsigned int); | |
10584 | ||
10585 | vector signed short vec_vsubshs (vector bool short, | |
10586 | vector signed short); | |
10587 | vector signed short vec_vsubshs (vector signed short, | |
10588 | vector bool short); | |
10589 | vector signed short vec_vsubshs (vector signed short, | |
10590 | vector signed short); | |
10591 | ||
10592 | vector unsigned short vec_vsubuhs (vector bool short, | |
10593 | vector unsigned short); | |
10594 | vector unsigned short vec_vsubuhs (vector unsigned short, | |
10595 | vector bool short); | |
10596 | vector unsigned short vec_vsubuhs (vector unsigned short, | |
10597 | vector unsigned short); | |
10598 | ||
10599 | vector signed char vec_vsubsbs (vector bool char, vector signed char); | |
10600 | vector signed char vec_vsubsbs (vector signed char, vector bool char); | |
10601 | vector signed char vec_vsubsbs (vector signed char, vector signed char); | |
10602 | ||
10603 | vector unsigned char vec_vsububs (vector bool char, | |
10604 | vector unsigned char); | |
10605 | vector unsigned char vec_vsububs (vector unsigned char, | |
10606 | vector bool char); | |
10607 | vector unsigned char vec_vsububs (vector unsigned char, | |
10608 | vector unsigned char); | |
10609 | ||
924fcc4e JM |
10610 | vector unsigned int vec_sum4s (vector unsigned char, |
10611 | vector unsigned int); | |
333c8841 AH |
10612 | vector signed int vec_sum4s (vector signed char, vector signed int); |
10613 | vector signed int vec_sum4s (vector signed short, vector signed int); | |
10614 | ||
b0b343db JJ |
10615 | vector signed int vec_vsum4shs (vector signed short, vector signed int); |
10616 | ||
10617 | vector signed int vec_vsum4sbs (vector signed char, vector signed int); | |
10618 | ||
10619 | vector unsigned int vec_vsum4ubs (vector unsigned char, | |
10620 | vector unsigned int); | |
10621 | ||
333c8841 AH |
10622 | vector signed int vec_sum2s (vector signed int, vector signed int); |
10623 | ||
10624 | vector signed int vec_sums (vector signed int, vector signed int); | |
10625 | ||
10626 | vector float vec_trunc (vector float); | |
10627 | ||
10628 | vector signed short vec_unpackh (vector signed char); | |
b0b343db | 10629 | vector bool short vec_unpackh (vector bool char); |
333c8841 | 10630 | vector signed int vec_unpackh (vector signed short); |
b0b343db JJ |
10631 | vector bool int vec_unpackh (vector bool short); |
10632 | vector unsigned int vec_unpackh (vector pixel); | |
10633 | ||
10634 | vector bool int vec_vupkhsh (vector bool short); | |
10635 | vector signed int vec_vupkhsh (vector signed short); | |
10636 | ||
10637 | vector unsigned int vec_vupkhpx (vector pixel); | |
10638 | ||
10639 | vector bool short vec_vupkhsb (vector bool char); | |
10640 | vector signed short vec_vupkhsb (vector signed char); | |
333c8841 AH |
10641 | |
10642 | vector signed short vec_unpackl (vector signed char); | |
b0b343db JJ |
10643 | vector bool short vec_unpackl (vector bool char); |
10644 | vector unsigned int vec_unpackl (vector pixel); | |
333c8841 | 10645 | vector signed int vec_unpackl (vector signed short); |
b0b343db JJ |
10646 | vector bool int vec_unpackl (vector bool short); |
10647 | ||
10648 | vector unsigned int vec_vupklpx (vector pixel); | |
10649 | ||
10650 | vector bool int vec_vupklsh (vector bool short); | |
10651 | vector signed int vec_vupklsh (vector signed short); | |
10652 | ||
10653 | vector bool short vec_vupklsb (vector bool char); | |
10654 | vector signed short vec_vupklsb (vector signed char); | |
333c8841 AH |
10655 | |
10656 | vector float vec_xor (vector float, vector float); | |
b0b343db JJ |
10657 | vector float vec_xor (vector float, vector bool int); |
10658 | vector float vec_xor (vector bool int, vector float); | |
10659 | vector bool int vec_xor (vector bool int, vector bool int); | |
10660 | vector signed int vec_xor (vector bool int, vector signed int); | |
10661 | vector signed int vec_xor (vector signed int, vector bool int); | |
333c8841 | 10662 | vector signed int vec_xor (vector signed int, vector signed int); |
b0b343db JJ |
10663 | vector unsigned int vec_xor (vector bool int, vector unsigned int); |
10664 | vector unsigned int vec_xor (vector unsigned int, vector bool int); | |
333c8841 | 10665 | vector unsigned int vec_xor (vector unsigned int, vector unsigned int); |
b0b343db JJ |
10666 | vector bool short vec_xor (vector bool short, vector bool short); |
10667 | vector signed short vec_xor (vector bool short, vector signed short); | |
10668 | vector signed short vec_xor (vector signed short, vector bool short); | |
333c8841 | 10669 | vector signed short vec_xor (vector signed short, vector signed short); |
b0b343db | 10670 | vector unsigned short vec_xor (vector bool short, |
924fcc4e JM |
10671 | vector unsigned short); |
10672 | vector unsigned short vec_xor (vector unsigned short, | |
b0b343db | 10673 | vector bool short); |
6e5bb5ad JM |
10674 | vector unsigned short vec_xor (vector unsigned short, |
10675 | vector unsigned short); | |
b0b343db JJ |
10676 | vector signed char vec_xor (vector bool char, vector signed char); |
10677 | vector bool char vec_xor (vector bool char, vector bool char); | |
10678 | vector signed char vec_xor (vector signed char, vector bool char); | |
333c8841 | 10679 | vector signed char vec_xor (vector signed char, vector signed char); |
b0b343db JJ |
10680 | vector unsigned char vec_xor (vector bool char, vector unsigned char); |
10681 | vector unsigned char vec_xor (vector unsigned char, vector bool char); | |
924fcc4e JM |
10682 | vector unsigned char vec_xor (vector unsigned char, |
10683 | vector unsigned char); | |
333c8841 | 10684 | |
b0b343db JJ |
10685 | int vec_all_eq (vector signed char, vector bool char); |
10686 | int vec_all_eq (vector signed char, vector signed char); | |
10687 | int vec_all_eq (vector unsigned char, vector bool char); | |
10688 | int vec_all_eq (vector unsigned char, vector unsigned char); | |
10689 | int vec_all_eq (vector bool char, vector bool char); | |
10690 | int vec_all_eq (vector bool char, vector unsigned char); | |
10691 | int vec_all_eq (vector bool char, vector signed char); | |
10692 | int vec_all_eq (vector signed short, vector bool short); | |
10693 | int vec_all_eq (vector signed short, vector signed short); | |
10694 | int vec_all_eq (vector unsigned short, vector bool short); | |
10695 | int vec_all_eq (vector unsigned short, vector unsigned short); | |
10696 | int vec_all_eq (vector bool short, vector bool short); | |
10697 | int vec_all_eq (vector bool short, vector unsigned short); | |
10698 | int vec_all_eq (vector bool short, vector signed short); | |
10699 | int vec_all_eq (vector pixel, vector pixel); | |
10700 | int vec_all_eq (vector signed int, vector bool int); | |
10701 | int vec_all_eq (vector signed int, vector signed int); | |
10702 | int vec_all_eq (vector unsigned int, vector bool int); | |
10703 | int vec_all_eq (vector unsigned int, vector unsigned int); | |
10704 | int vec_all_eq (vector bool int, vector bool int); | |
10705 | int vec_all_eq (vector bool int, vector unsigned int); | |
10706 | int vec_all_eq (vector bool int, vector signed int); | |
10707 | int vec_all_eq (vector float, vector float); | |
10708 | ||
10709 | int vec_all_ge (vector bool char, vector unsigned char); | |
10710 | int vec_all_ge (vector unsigned char, vector bool char); | |
10711 | int vec_all_ge (vector unsigned char, vector unsigned char); | |
10712 | int vec_all_ge (vector bool char, vector signed char); | |
10713 | int vec_all_ge (vector signed char, vector bool char); | |
10714 | int vec_all_ge (vector signed char, vector signed char); | |
10715 | int vec_all_ge (vector bool short, vector unsigned short); | |
10716 | int vec_all_ge (vector unsigned short, vector bool short); | |
10717 | int vec_all_ge (vector unsigned short, vector unsigned short); | |
10718 | int vec_all_ge (vector signed short, vector signed short); | |
10719 | int vec_all_ge (vector bool short, vector signed short); | |
10720 | int vec_all_ge (vector signed short, vector bool short); | |
10721 | int vec_all_ge (vector bool int, vector unsigned int); | |
10722 | int vec_all_ge (vector unsigned int, vector bool int); | |
10723 | int vec_all_ge (vector unsigned int, vector unsigned int); | |
10724 | int vec_all_ge (vector bool int, vector signed int); | |
10725 | int vec_all_ge (vector signed int, vector bool int); | |
10726 | int vec_all_ge (vector signed int, vector signed int); | |
10727 | int vec_all_ge (vector float, vector float); | |
10728 | ||
10729 | int vec_all_gt (vector bool char, vector unsigned char); | |
10730 | int vec_all_gt (vector unsigned char, vector bool char); | |
10731 | int vec_all_gt (vector unsigned char, vector unsigned char); | |
10732 | int vec_all_gt (vector bool char, vector signed char); | |
10733 | int vec_all_gt (vector signed char, vector bool char); | |
10734 | int vec_all_gt (vector signed char, vector signed char); | |
10735 | int vec_all_gt (vector bool short, vector unsigned short); | |
10736 | int vec_all_gt (vector unsigned short, vector bool short); | |
10737 | int vec_all_gt (vector unsigned short, vector unsigned short); | |
10738 | int vec_all_gt (vector bool short, vector signed short); | |
10739 | int vec_all_gt (vector signed short, vector bool short); | |
10740 | int vec_all_gt (vector signed short, vector signed short); | |
10741 | int vec_all_gt (vector bool int, vector unsigned int); | |
10742 | int vec_all_gt (vector unsigned int, vector bool int); | |
10743 | int vec_all_gt (vector unsigned int, vector unsigned int); | |
10744 | int vec_all_gt (vector bool int, vector signed int); | |
10745 | int vec_all_gt (vector signed int, vector bool int); | |
10746 | int vec_all_gt (vector signed int, vector signed int); | |
10747 | int vec_all_gt (vector float, vector float); | |
10748 | ||
10749 | int vec_all_in (vector float, vector float); | |
10750 | ||
10751 | int vec_all_le (vector bool char, vector unsigned char); | |
10752 | int vec_all_le (vector unsigned char, vector bool char); | |
10753 | int vec_all_le (vector unsigned char, vector unsigned char); | |
10754 | int vec_all_le (vector bool char, vector signed char); | |
10755 | int vec_all_le (vector signed char, vector bool char); | |
10756 | int vec_all_le (vector signed char, vector signed char); | |
10757 | int vec_all_le (vector bool short, vector unsigned short); | |
10758 | int vec_all_le (vector unsigned short, vector bool short); | |
10759 | int vec_all_le (vector unsigned short, vector unsigned short); | |
10760 | int vec_all_le (vector bool short, vector signed short); | |
10761 | int vec_all_le (vector signed short, vector bool short); | |
10762 | int vec_all_le (vector signed short, vector signed short); | |
10763 | int vec_all_le (vector bool int, vector unsigned int); | |
10764 | int vec_all_le (vector unsigned int, vector bool int); | |
10765 | int vec_all_le (vector unsigned int, vector unsigned int); | |
10766 | int vec_all_le (vector bool int, vector signed int); | |
10767 | int vec_all_le (vector signed int, vector bool int); | |
10768 | int vec_all_le (vector signed int, vector signed int); | |
10769 | int vec_all_le (vector float, vector float); | |
10770 | ||
10771 | int vec_all_lt (vector bool char, vector unsigned char); | |
10772 | int vec_all_lt (vector unsigned char, vector bool char); | |
10773 | int vec_all_lt (vector unsigned char, vector unsigned char); | |
10774 | int vec_all_lt (vector bool char, vector signed char); | |
10775 | int vec_all_lt (vector signed char, vector bool char); | |
10776 | int vec_all_lt (vector signed char, vector signed char); | |
10777 | int vec_all_lt (vector bool short, vector unsigned short); | |
10778 | int vec_all_lt (vector unsigned short, vector bool short); | |
10779 | int vec_all_lt (vector unsigned short, vector unsigned short); | |
10780 | int vec_all_lt (vector bool short, vector signed short); | |
10781 | int vec_all_lt (vector signed short, vector bool short); | |
10782 | int vec_all_lt (vector signed short, vector signed short); | |
10783 | int vec_all_lt (vector bool int, vector unsigned int); | |
10784 | int vec_all_lt (vector unsigned int, vector bool int); | |
10785 | int vec_all_lt (vector unsigned int, vector unsigned int); | |
10786 | int vec_all_lt (vector bool int, vector signed int); | |
10787 | int vec_all_lt (vector signed int, vector bool int); | |
10788 | int vec_all_lt (vector signed int, vector signed int); | |
10789 | int vec_all_lt (vector float, vector float); | |
10790 | ||
10791 | int vec_all_nan (vector float); | |
10792 | ||
10793 | int vec_all_ne (vector signed char, vector bool char); | |
10794 | int vec_all_ne (vector signed char, vector signed char); | |
10795 | int vec_all_ne (vector unsigned char, vector bool char); | |
10796 | int vec_all_ne (vector unsigned char, vector unsigned char); | |
10797 | int vec_all_ne (vector bool char, vector bool char); | |
10798 | int vec_all_ne (vector bool char, vector unsigned char); | |
10799 | int vec_all_ne (vector bool char, vector signed char); | |
10800 | int vec_all_ne (vector signed short, vector bool short); | |
10801 | int vec_all_ne (vector signed short, vector signed short); | |
10802 | int vec_all_ne (vector unsigned short, vector bool short); | |
10803 | int vec_all_ne (vector unsigned short, vector unsigned short); | |
10804 | int vec_all_ne (vector bool short, vector bool short); | |
10805 | int vec_all_ne (vector bool short, vector unsigned short); | |
10806 | int vec_all_ne (vector bool short, vector signed short); | |
10807 | int vec_all_ne (vector pixel, vector pixel); | |
10808 | int vec_all_ne (vector signed int, vector bool int); | |
10809 | int vec_all_ne (vector signed int, vector signed int); | |
10810 | int vec_all_ne (vector unsigned int, vector bool int); | |
10811 | int vec_all_ne (vector unsigned int, vector unsigned int); | |
10812 | int vec_all_ne (vector bool int, vector bool int); | |
10813 | int vec_all_ne (vector bool int, vector unsigned int); | |
10814 | int vec_all_ne (vector bool int, vector signed int); | |
10815 | int vec_all_ne (vector float, vector float); | |
10816 | ||
10817 | int vec_all_nge (vector float, vector float); | |
10818 | ||
10819 | int vec_all_ngt (vector float, vector float); | |
10820 | ||
10821 | int vec_all_nle (vector float, vector float); | |
10822 | ||
10823 | int vec_all_nlt (vector float, vector float); | |
10824 | ||
10825 | int vec_all_numeric (vector float); | |
10826 | ||
10827 | int vec_any_eq (vector signed char, vector bool char); | |
10828 | int vec_any_eq (vector signed char, vector signed char); | |
10829 | int vec_any_eq (vector unsigned char, vector bool char); | |
10830 | int vec_any_eq (vector unsigned char, vector unsigned char); | |
10831 | int vec_any_eq (vector bool char, vector bool char); | |
10832 | int vec_any_eq (vector bool char, vector unsigned char); | |
10833 | int vec_any_eq (vector bool char, vector signed char); | |
10834 | int vec_any_eq (vector signed short, vector bool short); | |
10835 | int vec_any_eq (vector signed short, vector signed short); | |
10836 | int vec_any_eq (vector unsigned short, vector bool short); | |
10837 | int vec_any_eq (vector unsigned short, vector unsigned short); | |
10838 | int vec_any_eq (vector bool short, vector bool short); | |
10839 | int vec_any_eq (vector bool short, vector unsigned short); | |
10840 | int vec_any_eq (vector bool short, vector signed short); | |
10841 | int vec_any_eq (vector pixel, vector pixel); | |
10842 | int vec_any_eq (vector signed int, vector bool int); | |
10843 | int vec_any_eq (vector signed int, vector signed int); | |
10844 | int vec_any_eq (vector unsigned int, vector bool int); | |
10845 | int vec_any_eq (vector unsigned int, vector unsigned int); | |
10846 | int vec_any_eq (vector bool int, vector bool int); | |
10847 | int vec_any_eq (vector bool int, vector unsigned int); | |
10848 | int vec_any_eq (vector bool int, vector signed int); | |
10849 | int vec_any_eq (vector float, vector float); | |
10850 | ||
10851 | int vec_any_ge (vector signed char, vector bool char); | |
10852 | int vec_any_ge (vector unsigned char, vector bool char); | |
10853 | int vec_any_ge (vector unsigned char, vector unsigned char); | |
10854 | int vec_any_ge (vector signed char, vector signed char); | |
10855 | int vec_any_ge (vector bool char, vector unsigned char); | |
10856 | int vec_any_ge (vector bool char, vector signed char); | |
10857 | int vec_any_ge (vector unsigned short, vector bool short); | |
10858 | int vec_any_ge (vector unsigned short, vector unsigned short); | |
10859 | int vec_any_ge (vector signed short, vector signed short); | |
10860 | int vec_any_ge (vector signed short, vector bool short); | |
10861 | int vec_any_ge (vector bool short, vector unsigned short); | |
10862 | int vec_any_ge (vector bool short, vector signed short); | |
10863 | int vec_any_ge (vector signed int, vector bool int); | |
10864 | int vec_any_ge (vector unsigned int, vector bool int); | |
10865 | int vec_any_ge (vector unsigned int, vector unsigned int); | |
10866 | int vec_any_ge (vector signed int, vector signed int); | |
10867 | int vec_any_ge (vector bool int, vector unsigned int); | |
10868 | int vec_any_ge (vector bool int, vector signed int); | |
10869 | int vec_any_ge (vector float, vector float); | |
10870 | ||
10871 | int vec_any_gt (vector bool char, vector unsigned char); | |
10872 | int vec_any_gt (vector unsigned char, vector bool char); | |
10873 | int vec_any_gt (vector unsigned char, vector unsigned char); | |
10874 | int vec_any_gt (vector bool char, vector signed char); | |
10875 | int vec_any_gt (vector signed char, vector bool char); | |
10876 | int vec_any_gt (vector signed char, vector signed char); | |
10877 | int vec_any_gt (vector bool short, vector unsigned short); | |
10878 | int vec_any_gt (vector unsigned short, vector bool short); | |
10879 | int vec_any_gt (vector unsigned short, vector unsigned short); | |
10880 | int vec_any_gt (vector bool short, vector signed short); | |
10881 | int vec_any_gt (vector signed short, vector bool short); | |
10882 | int vec_any_gt (vector signed short, vector signed short); | |
10883 | int vec_any_gt (vector bool int, vector unsigned int); | |
10884 | int vec_any_gt (vector unsigned int, vector bool int); | |
10885 | int vec_any_gt (vector unsigned int, vector unsigned int); | |
10886 | int vec_any_gt (vector bool int, vector signed int); | |
10887 | int vec_any_gt (vector signed int, vector bool int); | |
10888 | int vec_any_gt (vector signed int, vector signed int); | |
10889 | int vec_any_gt (vector float, vector float); | |
10890 | ||
10891 | int vec_any_le (vector bool char, vector unsigned char); | |
10892 | int vec_any_le (vector unsigned char, vector bool char); | |
10893 | int vec_any_le (vector unsigned char, vector unsigned char); | |
10894 | int vec_any_le (vector bool char, vector signed char); | |
10895 | int vec_any_le (vector signed char, vector bool char); | |
10896 | int vec_any_le (vector signed char, vector signed char); | |
10897 | int vec_any_le (vector bool short, vector unsigned short); | |
10898 | int vec_any_le (vector unsigned short, vector bool short); | |
10899 | int vec_any_le (vector unsigned short, vector unsigned short); | |
10900 | int vec_any_le (vector bool short, vector signed short); | |
10901 | int vec_any_le (vector signed short, vector bool short); | |
10902 | int vec_any_le (vector signed short, vector signed short); | |
10903 | int vec_any_le (vector bool int, vector unsigned int); | |
10904 | int vec_any_le (vector unsigned int, vector bool int); | |
10905 | int vec_any_le (vector unsigned int, vector unsigned int); | |
10906 | int vec_any_le (vector bool int, vector signed int); | |
10907 | int vec_any_le (vector signed int, vector bool int); | |
10908 | int vec_any_le (vector signed int, vector signed int); | |
10909 | int vec_any_le (vector float, vector float); | |
10910 | ||
10911 | int vec_any_lt (vector bool char, vector unsigned char); | |
10912 | int vec_any_lt (vector unsigned char, vector bool char); | |
10913 | int vec_any_lt (vector unsigned char, vector unsigned char); | |
10914 | int vec_any_lt (vector bool char, vector signed char); | |
10915 | int vec_any_lt (vector signed char, vector bool char); | |
10916 | int vec_any_lt (vector signed char, vector signed char); | |
10917 | int vec_any_lt (vector bool short, vector unsigned short); | |
10918 | int vec_any_lt (vector unsigned short, vector bool short); | |
10919 | int vec_any_lt (vector unsigned short, vector unsigned short); | |
10920 | int vec_any_lt (vector bool short, vector signed short); | |
10921 | int vec_any_lt (vector signed short, vector bool short); | |
10922 | int vec_any_lt (vector signed short, vector signed short); | |
10923 | int vec_any_lt (vector bool int, vector unsigned int); | |
10924 | int vec_any_lt (vector unsigned int, vector bool int); | |
10925 | int vec_any_lt (vector unsigned int, vector unsigned int); | |
10926 | int vec_any_lt (vector bool int, vector signed int); | |
10927 | int vec_any_lt (vector signed int, vector bool int); | |
10928 | int vec_any_lt (vector signed int, vector signed int); | |
10929 | int vec_any_lt (vector float, vector float); | |
10930 | ||
10931 | int vec_any_nan (vector float); | |
10932 | ||
10933 | int vec_any_ne (vector signed char, vector bool char); | |
10934 | int vec_any_ne (vector signed char, vector signed char); | |
10935 | int vec_any_ne (vector unsigned char, vector bool char); | |
10936 | int vec_any_ne (vector unsigned char, vector unsigned char); | |
10937 | int vec_any_ne (vector bool char, vector bool char); | |
10938 | int vec_any_ne (vector bool char, vector unsigned char); | |
10939 | int vec_any_ne (vector bool char, vector signed char); | |
10940 | int vec_any_ne (vector signed short, vector bool short); | |
10941 | int vec_any_ne (vector signed short, vector signed short); | |
10942 | int vec_any_ne (vector unsigned short, vector bool short); | |
10943 | int vec_any_ne (vector unsigned short, vector unsigned short); | |
10944 | int vec_any_ne (vector bool short, vector bool short); | |
10945 | int vec_any_ne (vector bool short, vector unsigned short); | |
10946 | int vec_any_ne (vector bool short, vector signed short); | |
10947 | int vec_any_ne (vector pixel, vector pixel); | |
10948 | int vec_any_ne (vector signed int, vector bool int); | |
10949 | int vec_any_ne (vector signed int, vector signed int); | |
10950 | int vec_any_ne (vector unsigned int, vector bool int); | |
10951 | int vec_any_ne (vector unsigned int, vector unsigned int); | |
10952 | int vec_any_ne (vector bool int, vector bool int); | |
10953 | int vec_any_ne (vector bool int, vector unsigned int); | |
10954 | int vec_any_ne (vector bool int, vector signed int); | |
10955 | int vec_any_ne (vector float, vector float); | |
10956 | ||
10957 | int vec_any_nge (vector float, vector float); | |
10958 | ||
10959 | int vec_any_ngt (vector float, vector float); | |
10960 | ||
10961 | int vec_any_nle (vector float, vector float); | |
10962 | ||
10963 | int vec_any_nlt (vector float, vector float); | |
10964 | ||
10965 | int vec_any_numeric (vector float); | |
10966 | ||
10967 | int vec_any_out (vector float, vector float); | |
333c8841 AH |
10968 | @end smallexample |
10969 | ||
c5145ceb JM |
10970 | @node SPARC VIS Built-in Functions |
10971 | @subsection SPARC VIS Built-in Functions | |
10972 | ||
10973 | GCC supports SIMD operations on the SPARC using both the generic vector | |
2fd13506 | 10974 | extensions (@pxref{Vector Extensions}) as well as built-in functions for |
c5145ceb JM |
10975 | the SPARC Visual Instruction Set (VIS). When you use the @option{-mvis} |
10976 | switch, the VIS extension is exposed as the following built-in functions: | |
10977 | ||
10978 | @smallexample | |
10979 | typedef int v2si __attribute__ ((vector_size (8))); | |
10980 | typedef short v4hi __attribute__ ((vector_size (8))); | |
10981 | typedef short v2hi __attribute__ ((vector_size (4))); | |
10982 | typedef char v8qi __attribute__ ((vector_size (8))); | |
10983 | typedef char v4qi __attribute__ ((vector_size (4))); | |
10984 | ||
10985 | void * __builtin_vis_alignaddr (void *, long); | |
10986 | int64_t __builtin_vis_faligndatadi (int64_t, int64_t); | |
10987 | v2si __builtin_vis_faligndatav2si (v2si, v2si); | |
10988 | v4hi __builtin_vis_faligndatav4hi (v4si, v4si); | |
10989 | v8qi __builtin_vis_faligndatav8qi (v8qi, v8qi); | |
10990 | ||
10991 | v4hi __builtin_vis_fexpand (v4qi); | |
10992 | ||
10993 | v4hi __builtin_vis_fmul8x16 (v4qi, v4hi); | |
10994 | v4hi __builtin_vis_fmul8x16au (v4qi, v4hi); | |
10995 | v4hi __builtin_vis_fmul8x16al (v4qi, v4hi); | |
10996 | v4hi __builtin_vis_fmul8sux16 (v8qi, v4hi); | |
10997 | v4hi __builtin_vis_fmul8ulx16 (v8qi, v4hi); | |
10998 | v2si __builtin_vis_fmuld8sux16 (v4qi, v2hi); | |
10999 | v2si __builtin_vis_fmuld8ulx16 (v4qi, v2hi); | |
11000 | ||
11001 | v4qi __builtin_vis_fpack16 (v4hi); | |
11002 | v8qi __builtin_vis_fpack32 (v2si, v2si); | |
11003 | v2hi __builtin_vis_fpackfix (v2si); | |
11004 | v8qi __builtin_vis_fpmerge (v4qi, v4qi); | |
11005 | ||
11006 | int64_t __builtin_vis_pdist (v8qi, v8qi, int64_t); | |
11007 | @end smallexample | |
11008 | ||
85d9c13c TS |
11009 | @node SPU Built-in Functions |
11010 | @subsection SPU Built-in Functions | |
11011 | ||
11012 | GCC provides extensions for the SPU processor as described in the | |
11013 | Sony/Toshiba/IBM SPU Language Extensions Specification, which can be | |
11014 | found at @uref{http://cell.scei.co.jp/} or | |
11015 | @uref{http://www.ibm.com/developerworks/power/cell/}. GCC's | |
11016 | implementation differs in several ways. | |
11017 | ||
11018 | @itemize @bullet | |
11019 | ||
11020 | @item | |
11021 | The optional extension of specifying vector constants in parentheses is | |
11022 | not supported. | |
11023 | ||
11024 | @item | |
11025 | A vector initializer requires no cast if the vector constant is of the | |
11026 | same type as the variable it is initializing. | |
11027 | ||
11028 | @item | |
11029 | If @code{signed} or @code{unsigned} is omitted, the signedness of the | |
11030 | vector type is the default signedness of the base type. The default | |
11031 | varies depending on the operating system, so a portable program should | |
11032 | always specify the signedness. | |
11033 | ||
11034 | @item | |
11035 | By default, the keyword @code{__vector} is added. The macro | |
11036 | @code{vector} is defined in @code{<spu_intrinsics.h>} and can be | |
11037 | undefined. | |
11038 | ||
11039 | @item | |
11040 | GCC allows using a @code{typedef} name as the type specifier for a | |
11041 | vector type. | |
11042 | ||
11043 | @item | |
11044 | For C, overloaded functions are implemented with macros so the following | |
11045 | does not work: | |
11046 | ||
11047 | @smallexample | |
11048 | spu_add ((vector signed int)@{1, 2, 3, 4@}, foo); | |
11049 | @end smallexample | |
11050 | ||
11051 | Since @code{spu_add} is a macro, the vector constant in the example | |
11052 | is treated as four separate arguments. Wrap the entire argument in | |
11053 | parentheses for this to work. | |
11054 | ||
11055 | @item | |
11056 | The extended version of @code{__builtin_expect} is not supported. | |
11057 | ||
11058 | @end itemize | |
11059 | ||
5681c208 | 11060 | @emph{Note:} Only the interface described in the aforementioned |
85d9c13c TS |
11061 | specification is supported. Internally, GCC uses built-in functions to |
11062 | implement the required functionality, but these are not supported and | |
11063 | are subject to change without notice. | |
11064 | ||
a2bec818 DJ |
11065 | @node Target Format Checks |
11066 | @section Format Checks Specific to Particular Target Machines | |
11067 | ||
11068 | For some target machines, GCC supports additional options to the | |
11069 | format attribute | |
11070 | (@pxref{Function Attributes,,Declaring Attributes of Functions}). | |
11071 | ||
11072 | @menu | |
11073 | * Solaris Format Checks:: | |
11074 | @end menu | |
11075 | ||
11076 | @node Solaris Format Checks | |
11077 | @subsection Solaris Format Checks | |
11078 | ||
11079 | Solaris targets support the @code{cmn_err} (or @code{__cmn_err__}) format | |
11080 | check. @code{cmn_err} accepts a subset of the standard @code{printf} | |
11081 | conversions, and the two-argument @code{%b} conversion for displaying | |
11082 | bit-fields. See the Solaris man page for @code{cmn_err} for more information. | |
11083 | ||
0168a849 SS |
11084 | @node Pragmas |
11085 | @section Pragmas Accepted by GCC | |
11086 | @cindex pragmas | |
11087 | @cindex #pragma | |
11088 | ||
11089 | GCC supports several types of pragmas, primarily in order to compile | |
11090 | code originally written for other compilers. Note that in general | |
11091 | we do not recommend the use of pragmas; @xref{Function Attributes}, | |
11092 | for further explanation. | |
11093 | ||
11094 | @menu | |
11095 | * ARM Pragmas:: | |
38b2d076 | 11096 | * M32C Pragmas:: |
a5c76ee6 | 11097 | * RS/6000 and PowerPC Pragmas:: |
0168a849 | 11098 | * Darwin Pragmas:: |
07a43492 | 11099 | * Solaris Pragmas:: |
84b8b0e0 | 11100 | * Symbol-Renaming Pragmas:: |
467cecf3 | 11101 | * Structure-Packing Pragmas:: |
52eb57df | 11102 | * Weak Pragmas:: |
79cf5994 | 11103 | * Diagnostic Pragmas:: |
b9e75696 | 11104 | * Visibility Pragmas:: |
20cef83a | 11105 | * Push/Pop Macro Pragmas:: |
0168a849 SS |
11106 | @end menu |
11107 | ||
11108 | @node ARM Pragmas | |
11109 | @subsection ARM Pragmas | |
11110 | ||
11111 | The ARM target defines pragmas for controlling the default addition of | |
11112 | @code{long_call} and @code{short_call} attributes to functions. | |
11113 | @xref{Function Attributes}, for information about the effects of these | |
11114 | attributes. | |
11115 | ||
11116 | @table @code | |
11117 | @item long_calls | |
11118 | @cindex pragma, long_calls | |
11119 | Set all subsequent functions to have the @code{long_call} attribute. | |
11120 | ||
11121 | @item no_long_calls | |
11122 | @cindex pragma, no_long_calls | |
11123 | Set all subsequent functions to have the @code{short_call} attribute. | |
11124 | ||
11125 | @item long_calls_off | |
11126 | @cindex pragma, long_calls_off | |
11127 | Do not affect the @code{long_call} or @code{short_call} attributes of | |
11128 | subsequent functions. | |
11129 | @end table | |
11130 | ||
38b2d076 DD |
11131 | @node M32C Pragmas |
11132 | @subsection M32C Pragmas | |
11133 | ||
11134 | @table @code | |
11135 | @item memregs @var{number} | |
11136 | @cindex pragma, memregs | |
11137 | Overrides the command line option @code{-memregs=} for the current | |
11138 | file. Use with care! This pragma must be before any function in the | |
11139 | file, and mixing different memregs values in different objects may | |
11140 | make them incompatible. This pragma is useful when a | |
11141 | performance-critical function uses a memreg for temporary values, | |
11142 | as it may allow you to reduce the number of memregs used. | |
11143 | ||
11144 | @end table | |
11145 | ||
a5c76ee6 ZW |
11146 | @node RS/6000 and PowerPC Pragmas |
11147 | @subsection RS/6000 and PowerPC Pragmas | |
11148 | ||
11149 | The RS/6000 and PowerPC targets define one pragma for controlling | |
11150 | whether or not the @code{longcall} attribute is added to function | |
11151 | declarations by default. This pragma overrides the @option{-mlongcall} | |
95b1627e | 11152 | option, but not the @code{longcall} and @code{shortcall} attributes. |
a5c76ee6 ZW |
11153 | @xref{RS/6000 and PowerPC Options}, for more information about when long |
11154 | calls are and are not necessary. | |
11155 | ||
11156 | @table @code | |
11157 | @item longcall (1) | |
11158 | @cindex pragma, longcall | |
11159 | Apply the @code{longcall} attribute to all subsequent function | |
11160 | declarations. | |
11161 | ||
11162 | @item longcall (0) | |
11163 | Do not apply the @code{longcall} attribute to subsequent function | |
11164 | declarations. | |
11165 | @end table | |
11166 | ||
0168a849 | 11167 | @c Describe h8300 pragmas here. |
0168a849 SS |
11168 | @c Describe sh pragmas here. |
11169 | @c Describe v850 pragmas here. | |
11170 | ||
11171 | @node Darwin Pragmas | |
11172 | @subsection Darwin Pragmas | |
11173 | ||
11174 | The following pragmas are available for all architectures running the | |
11175 | Darwin operating system. These are useful for compatibility with other | |
85ebf0c6 | 11176 | Mac OS compilers. |
0168a849 SS |
11177 | |
11178 | @table @code | |
11179 | @item mark @var{tokens}@dots{} | |
11180 | @cindex pragma, mark | |
11181 | This pragma is accepted, but has no effect. | |
11182 | ||
11183 | @item options align=@var{alignment} | |
11184 | @cindex pragma, options align | |
11185 | This pragma sets the alignment of fields in structures. The values of | |
11186 | @var{alignment} may be @code{mac68k}, to emulate m68k alignment, or | |
11187 | @code{power}, to emulate PowerPC alignment. Uses of this pragma nest | |
11188 | properly; to restore the previous setting, use @code{reset} for the | |
11189 | @var{alignment}. | |
11190 | ||
11191 | @item segment @var{tokens}@dots{} | |
11192 | @cindex pragma, segment | |
11193 | This pragma is accepted, but has no effect. | |
11194 | ||
11195 | @item unused (@var{var} [, @var{var}]@dots{}) | |
11196 | @cindex pragma, unused | |
11197 | This pragma declares variables to be possibly unused. GCC will not | |
11198 | produce warnings for the listed variables. The effect is similar to | |
11199 | that of the @code{unused} attribute, except that this pragma may appear | |
11200 | anywhere within the variables' scopes. | |
11201 | @end table | |
11202 | ||
07a43492 DJ |
11203 | @node Solaris Pragmas |
11204 | @subsection Solaris Pragmas | |
11205 | ||
11206 | The Solaris target supports @code{#pragma redefine_extname} | |
11207 | (@pxref{Symbol-Renaming Pragmas}). It also supports additional | |
11208 | @code{#pragma} directives for compatibility with the system compiler. | |
11209 | ||
11210 | @table @code | |
11211 | @item align @var{alignment} (@var{variable} [, @var{variable}]...) | |
11212 | @cindex pragma, align | |
11213 | ||
11214 | Increase the minimum alignment of each @var{variable} to @var{alignment}. | |
11215 | This is the same as GCC's @code{aligned} attribute @pxref{Variable | |
b5b3e36a | 11216 | Attributes}). Macro expansion occurs on the arguments to this pragma |
0ee2ea09 | 11217 | when compiling C and Objective-C@. It does not currently occur when |
b5b3e36a DJ |
11218 | compiling C++, but this is a bug which may be fixed in a future |
11219 | release. | |
07a43492 DJ |
11220 | |
11221 | @item fini (@var{function} [, @var{function}]...) | |
11222 | @cindex pragma, fini | |
11223 | ||
11224 | This pragma causes each listed @var{function} to be called after | |
11225 | main, or during shared module unloading, by adding a call to the | |
11226 | @code{.fini} section. | |
11227 | ||
11228 | @item init (@var{function} [, @var{function}]...) | |
11229 | @cindex pragma, init | |
11230 | ||
11231 | This pragma causes each listed @var{function} to be called during | |
11232 | initialization (before @code{main}) or during shared module loading, by | |
11233 | adding a call to the @code{.init} section. | |
11234 | ||
11235 | @end table | |
11236 | ||
84b8b0e0 ZW |
11237 | @node Symbol-Renaming Pragmas |
11238 | @subsection Symbol-Renaming Pragmas | |
41c64394 | 11239 | |
84b8b0e0 ZW |
11240 | For compatibility with the Solaris and Tru64 UNIX system headers, GCC |
11241 | supports two @code{#pragma} directives which change the name used in | |
11242 | assembly for a given declaration. These pragmas are only available on | |
11243 | platforms whose system headers need them. To get this effect on all | |
11244 | platforms supported by GCC, use the asm labels extension (@pxref{Asm | |
11245 | Labels}). | |
41c64394 RH |
11246 | |
11247 | @table @code | |
11248 | @item redefine_extname @var{oldname} @var{newname} | |
11249 | @cindex pragma, redefine_extname | |
11250 | ||
84b8b0e0 ZW |
11251 | This pragma gives the C function @var{oldname} the assembly symbol |
11252 | @var{newname}. The preprocessor macro @code{__PRAGMA_REDEFINE_EXTNAME} | |
11253 | will be defined if this pragma is available (currently only on | |
11254 | Solaris). | |
41c64394 | 11255 | |
41c64394 RH |
11256 | @item extern_prefix @var{string} |
11257 | @cindex pragma, extern_prefix | |
11258 | ||
84b8b0e0 ZW |
11259 | This pragma causes all subsequent external function and variable |
11260 | declarations to have @var{string} prepended to their assembly symbols. | |
11261 | This effect may be terminated with another @code{extern_prefix} pragma | |
11262 | whose argument is an empty string. The preprocessor macro | |
11263 | @code{__PRAGMA_EXTERN_PREFIX} will be defined if this pragma is | |
8a36672b | 11264 | available (currently only on Tru64 UNIX)@. |
41c64394 RH |
11265 | @end table |
11266 | ||
84b8b0e0 ZW |
11267 | These pragmas and the asm labels extension interact in a complicated |
11268 | manner. Here are some corner cases you may want to be aware of. | |
11269 | ||
11270 | @enumerate | |
11271 | @item Both pragmas silently apply only to declarations with external | |
11272 | linkage. Asm labels do not have this restriction. | |
11273 | ||
11274 | @item In C++, both pragmas silently apply only to declarations with | |
11275 | ``C'' linkage. Again, asm labels do not have this restriction. | |
11276 | ||
11277 | @item If any of the three ways of changing the assembly name of a | |
11278 | declaration is applied to a declaration whose assembly name has | |
11279 | already been determined (either by a previous use of one of these | |
11280 | features, or because the compiler needed the assembly name in order to | |
11281 | generate code), and the new name is different, a warning issues and | |
11282 | the name does not change. | |
11283 | ||
11284 | @item The @var{oldname} used by @code{#pragma redefine_extname} is | |
11285 | always the C-language name. | |
11286 | ||
11287 | @item If @code{#pragma extern_prefix} is in effect, and a declaration | |
11288 | occurs with an asm label attached, the prefix is silently ignored for | |
11289 | that declaration. | |
11290 | ||
11291 | @item If @code{#pragma extern_prefix} and @code{#pragma redefine_extname} | |
11292 | apply to the same declaration, whichever triggered first wins, and a | |
11293 | warning issues if they contradict each other. (We would like to have | |
11294 | @code{#pragma redefine_extname} always win, for consistency with asm | |
11295 | labels, but if @code{#pragma extern_prefix} triggers first we have no | |
11296 | way of knowing that that happened.) | |
11297 | @end enumerate | |
11298 | ||
467cecf3 JB |
11299 | @node Structure-Packing Pragmas |
11300 | @subsection Structure-Packing Pragmas | |
11301 | ||
20cef83a DS |
11302 | For compatibility with Microsoft Windows compilers, GCC supports a |
11303 | set of @code{#pragma} directives which change the maximum alignment of | |
11304 | members of structures (other than zero-width bitfields), unions, and | |
11305 | classes subsequently defined. The @var{n} value below always is required | |
11306 | to be a small power of two and specifies the new alignment in bytes. | |
467cecf3 JB |
11307 | |
11308 | @enumerate | |
11309 | @item @code{#pragma pack(@var{n})} simply sets the new alignment. | |
11310 | @item @code{#pragma pack()} sets the alignment to the one that was in | |
11311 | effect when compilation started (see also command line option | |
11312 | @option{-fpack-struct[=<n>]} @pxref{Code Gen Options}). | |
11313 | @item @code{#pragma pack(push[,@var{n}])} pushes the current alignment | |
11314 | setting on an internal stack and then optionally sets the new alignment. | |
11315 | @item @code{#pragma pack(pop)} restores the alignment setting to the one | |
11316 | saved at the top of the internal stack (and removes that stack entry). | |
11317 | Note that @code{#pragma pack([@var{n}])} does not influence this internal | |
11318 | stack; thus it is possible to have @code{#pragma pack(push)} followed by | |
11319 | multiple @code{#pragma pack(@var{n})} instances and finalized by a single | |
11320 | @code{#pragma pack(pop)}. | |
11321 | @end enumerate | |
11322 | ||
021efafc | 11323 | Some targets, e.g.@: i386 and powerpc, support the @code{ms_struct} |
6bb7beac EC |
11324 | @code{#pragma} which lays out a structure as the documented |
11325 | @code{__attribute__ ((ms_struct))}. | |
11326 | @enumerate | |
11327 | @item @code{#pragma ms_struct on} turns on the layout for structures | |
11328 | declared. | |
11329 | @item @code{#pragma ms_struct off} turns off the layout for structures | |
11330 | declared. | |
11331 | @item @code{#pragma ms_struct reset} goes back to the default layout. | |
11332 | @end enumerate | |
11333 | ||
52eb57df RH |
11334 | @node Weak Pragmas |
11335 | @subsection Weak Pragmas | |
11336 | ||
11337 | For compatibility with SVR4, GCC supports a set of @code{#pragma} | |
11338 | directives for declaring symbols to be weak, and defining weak | |
11339 | aliases. | |
11340 | ||
11341 | @table @code | |
11342 | @item #pragma weak @var{symbol} | |
11343 | @cindex pragma, weak | |
11344 | This pragma declares @var{symbol} to be weak, as if the declaration | |
11345 | had the attribute of the same name. The pragma may appear before | |
0ac11108 | 11346 | or after the declaration of @var{symbol}, but must appear before |
52eb57df RH |
11347 | either its first use or its definition. It is not an error for |
11348 | @var{symbol} to never be defined at all. | |
11349 | ||
11350 | @item #pragma weak @var{symbol1} = @var{symbol2} | |
11351 | This pragma declares @var{symbol1} to be a weak alias of @var{symbol2}. | |
11352 | It is an error if @var{symbol2} is not defined in the current | |
11353 | translation unit. | |
11354 | @end table | |
11355 | ||
79cf5994 DD |
11356 | @node Diagnostic Pragmas |
11357 | @subsection Diagnostic Pragmas | |
11358 | ||
11359 | GCC allows the user to selectively enable or disable certain types of | |
11360 | diagnostics, and change the kind of the diagnostic. For example, a | |
11361 | project's policy might require that all sources compile with | |
11362 | @option{-Werror} but certain files might have exceptions allowing | |
11363 | specific types of warnings. Or, a project might selectively enable | |
11364 | diagnostics and treat them as errors depending on which preprocessor | |
11365 | macros are defined. | |
11366 | ||
11367 | @table @code | |
11368 | @item #pragma GCC diagnostic @var{kind} @var{option} | |
11369 | @cindex pragma, diagnostic | |
11370 | ||
11371 | Modifies the disposition of a diagnostic. Note that not all | |
1eaf20ec | 11372 | diagnostics are modifiable; at the moment only warnings (normally |
923158be | 11373 | controlled by @samp{-W@dots{}}) can be controlled, and not all of them. |
79cf5994 DD |
11374 | Use @option{-fdiagnostics-show-option} to determine which diagnostics |
11375 | are controllable and which option controls them. | |
11376 | ||
11377 | @var{kind} is @samp{error} to treat this diagnostic as an error, | |
11378 | @samp{warning} to treat it like a warning (even if @option{-Werror} is | |
11379 | in effect), or @samp{ignored} if the diagnostic is to be ignored. | |
11380 | @var{option} is a double quoted string which matches the command line | |
11381 | option. | |
11382 | ||
11383 | @example | |
11384 | #pragma GCC diagnostic warning "-Wformat" | |
c116cd05 MLI |
11385 | #pragma GCC diagnostic error "-Wformat" |
11386 | #pragma GCC diagnostic ignored "-Wformat" | |
79cf5994 DD |
11387 | @end example |
11388 | ||
11389 | Note that these pragmas override any command line options. Also, | |
11390 | while it is syntactically valid to put these pragmas anywhere in your | |
11391 | sources, the only supported location for them is before any data or | |
11392 | functions are defined. Doing otherwise may result in unpredictable | |
11393 | results depending on how the optimizer manages your sources. If the | |
11394 | same option is listed multiple times, the last one specified is the | |
11395 | one that is in effect. This pragma is not intended to be a general | |
11396 | purpose replacement for command line options, but for implementing | |
11397 | strict control over project policies. | |
11398 | ||
11399 | @end table | |
11400 | ||
b9e75696 JM |
11401 | @node Visibility Pragmas |
11402 | @subsection Visibility Pragmas | |
11403 | ||
11404 | @table @code | |
11405 | @item #pragma GCC visibility push(@var{visibility}) | |
11406 | @itemx #pragma GCC visibility pop | |
11407 | @cindex pragma, visibility | |
11408 | ||
11409 | This pragma allows the user to set the visibility for multiple | |
11410 | declarations without having to give each a visibility attribute | |
11411 | @xref{Function Attributes}, for more information about visibility and | |
11412 | the attribute syntax. | |
11413 | ||
11414 | In C++, @samp{#pragma GCC visibility} affects only namespace-scope | |
11415 | declarations. Class members and template specializations are not | |
11416 | affected; if you want to override the visibility for a particular | |
11417 | member or instantiation, you must use an attribute. | |
11418 | ||
11419 | @end table | |
11420 | ||
20cef83a DS |
11421 | |
11422 | @node Push/Pop Macro Pragmas | |
11423 | @subsection Push/Pop Macro Pragmas | |
11424 | ||
11425 | For compatibility with Microsoft Windows compilers, GCC supports | |
11426 | @samp{#pragma push_macro(@var{"macro_name"})} | |
11427 | and @samp{#pragma pop_macro(@var{"macro_name"})}. | |
11428 | ||
11429 | @table @code | |
11430 | @item #pragma push_macro(@var{"macro_name"}) | |
11431 | @cindex pragma, push_macro | |
11432 | This pragma saves the value of the macro named as @var{macro_name} to | |
11433 | the top of the stack for this macro. | |
11434 | ||
11435 | @item #pragma pop_macro(@var{"macro_name"}) | |
11436 | @cindex pragma, pop_macro | |
11437 | This pragma sets the value of the macro named as @var{macro_name} to | |
11438 | the value on top of the stack for this macro. If the stack for | |
11439 | @var{macro_name} is empty, the value of the macro remains unchanged. | |
11440 | @end table | |
11441 | ||
11442 | For example: | |
11443 | ||
11444 | @smallexample | |
11445 | #define X 1 | |
11446 | #pragma push_macro("X") | |
11447 | #undef X | |
11448 | #define X -1 | |
11449 | #pragma pop_macro("X") | |
11450 | int x [X]; | |
11451 | @end smallexample | |
11452 | ||
11453 | In this example, the definition of X as 1 is saved by @code{#pragma | |
11454 | push_macro} and restored by @code{#pragma pop_macro}. | |
11455 | ||
3e96a2fd | 11456 | @node Unnamed Fields |
2fbebc71 | 11457 | @section Unnamed struct/union fields within structs/unions |
3e96a2fd DD |
11458 | @cindex struct |
11459 | @cindex union | |
11460 | ||
11461 | For compatibility with other compilers, GCC allows you to define | |
11462 | a structure or union that contains, as fields, structures and unions | |
11463 | without names. For example: | |
11464 | ||
3ab51846 | 11465 | @smallexample |
3e96a2fd DD |
11466 | struct @{ |
11467 | int a; | |
11468 | union @{ | |
11469 | int b; | |
11470 | float c; | |
11471 | @}; | |
11472 | int d; | |
11473 | @} foo; | |
3ab51846 | 11474 | @end smallexample |
3e96a2fd DD |
11475 | |
11476 | In this example, the user would be able to access members of the unnamed | |
11477 | union with code like @samp{foo.b}. Note that only unnamed structs and | |
11478 | unions are allowed, you may not have, for example, an unnamed | |
11479 | @code{int}. | |
11480 | ||
11481 | You must never create such structures that cause ambiguous field definitions. | |
11482 | For example, this structure: | |
11483 | ||
3ab51846 | 11484 | @smallexample |
3e96a2fd DD |
11485 | struct @{ |
11486 | int a; | |
11487 | struct @{ | |
11488 | int a; | |
11489 | @}; | |
11490 | @} foo; | |
3ab51846 | 11491 | @end smallexample |
3e96a2fd DD |
11492 | |
11493 | It is ambiguous which @code{a} is being referred to with @samp{foo.a}. | |
11494 | Such constructs are not supported and must be avoided. In the future, | |
11495 | such constructs may be detected and treated as compilation errors. | |
11496 | ||
2fbebc71 JM |
11497 | @opindex fms-extensions |
11498 | Unless @option{-fms-extensions} is used, the unnamed field must be a | |
11499 | structure or union definition without a tag (for example, @samp{struct | |
11500 | @{ int a; @};}). If @option{-fms-extensions} is used, the field may | |
11501 | also be a definition with a tag such as @samp{struct foo @{ int a; | |
11502 | @};}, a reference to a previously defined structure or union such as | |
11503 | @samp{struct foo;}, or a reference to a @code{typedef} name for a | |
11504 | previously defined structure or union type. | |
11505 | ||
3d78f2e9 RH |
11506 | @node Thread-Local |
11507 | @section Thread-Local Storage | |
11508 | @cindex Thread-Local Storage | |
9217ef40 | 11509 | @cindex @acronym{TLS} |
3d78f2e9 RH |
11510 | @cindex __thread |
11511 | ||
9217ef40 RH |
11512 | Thread-local storage (@acronym{TLS}) is a mechanism by which variables |
11513 | are allocated such that there is one instance of the variable per extant | |
3d78f2e9 RH |
11514 | thread. The run-time model GCC uses to implement this originates |
11515 | in the IA-64 processor-specific ABI, but has since been migrated | |
11516 | to other processors as well. It requires significant support from | |
11517 | the linker (@command{ld}), dynamic linker (@command{ld.so}), and | |
11518 | system libraries (@file{libc.so} and @file{libpthread.so}), so it | |
9217ef40 | 11519 | is not available everywhere. |
3d78f2e9 RH |
11520 | |
11521 | At the user level, the extension is visible with a new storage | |
11522 | class keyword: @code{__thread}. For example: | |
11523 | ||
3ab51846 | 11524 | @smallexample |
3d78f2e9 RH |
11525 | __thread int i; |
11526 | extern __thread struct state s; | |
11527 | static __thread char *p; | |
3ab51846 | 11528 | @end smallexample |
3d78f2e9 RH |
11529 | |
11530 | The @code{__thread} specifier may be used alone, with the @code{extern} | |
11531 | or @code{static} specifiers, but with no other storage class specifier. | |
11532 | When used with @code{extern} or @code{static}, @code{__thread} must appear | |
11533 | immediately after the other storage class specifier. | |
11534 | ||
11535 | The @code{__thread} specifier may be applied to any global, file-scoped | |
244c2241 RH |
11536 | static, function-scoped static, or static data member of a class. It may |
11537 | not be applied to block-scoped automatic or non-static data member. | |
3d78f2e9 RH |
11538 | |
11539 | When the address-of operator is applied to a thread-local variable, it is | |
11540 | evaluated at run-time and returns the address of the current thread's | |
11541 | instance of that variable. An address so obtained may be used by any | |
11542 | thread. When a thread terminates, any pointers to thread-local variables | |
11543 | in that thread become invalid. | |
11544 | ||
11545 | No static initialization may refer to the address of a thread-local variable. | |
11546 | ||
244c2241 RH |
11547 | In C++, if an initializer is present for a thread-local variable, it must |
11548 | be a @var{constant-expression}, as defined in 5.19.2 of the ANSI/ISO C++ | |
11549 | standard. | |
3d78f2e9 RH |
11550 | |
11551 | See @uref{http://people.redhat.com/drepper/tls.pdf, | |
11552 | ELF Handling For Thread-Local Storage} for a detailed explanation of | |
11553 | the four thread-local storage addressing models, and how the run-time | |
11554 | is expected to function. | |
11555 | ||
9217ef40 RH |
11556 | @menu |
11557 | * C99 Thread-Local Edits:: | |
11558 | * C++98 Thread-Local Edits:: | |
11559 | @end menu | |
11560 | ||
11561 | @node C99 Thread-Local Edits | |
11562 | @subsection ISO/IEC 9899:1999 Edits for Thread-Local Storage | |
11563 | ||
11564 | The following are a set of changes to ISO/IEC 9899:1999 (aka C99) | |
11565 | that document the exact semantics of the language extension. | |
11566 | ||
11567 | @itemize @bullet | |
11568 | @item | |
11569 | @cite{5.1.2 Execution environments} | |
11570 | ||
11571 | Add new text after paragraph 1 | |
11572 | ||
11573 | @quotation | |
11574 | Within either execution environment, a @dfn{thread} is a flow of | |
11575 | control within a program. It is implementation defined whether | |
11576 | or not there may be more than one thread associated with a program. | |
11577 | It is implementation defined how threads beyond the first are | |
11578 | created, the name and type of the function called at thread | |
11579 | startup, and how threads may be terminated. However, objects | |
11580 | with thread storage duration shall be initialized before thread | |
11581 | startup. | |
11582 | @end quotation | |
11583 | ||
11584 | @item | |
11585 | @cite{6.2.4 Storage durations of objects} | |
11586 | ||
11587 | Add new text before paragraph 3 | |
11588 | ||
11589 | @quotation | |
11590 | An object whose identifier is declared with the storage-class | |
11591 | specifier @w{@code{__thread}} has @dfn{thread storage duration}. | |
11592 | Its lifetime is the entire execution of the thread, and its | |
11593 | stored value is initialized only once, prior to thread startup. | |
11594 | @end quotation | |
11595 | ||
11596 | @item | |
11597 | @cite{6.4.1 Keywords} | |
11598 | ||
11599 | Add @code{__thread}. | |
11600 | ||
11601 | @item | |
11602 | @cite{6.7.1 Storage-class specifiers} | |
11603 | ||
11604 | Add @code{__thread} to the list of storage class specifiers in | |
11605 | paragraph 1. | |
11606 | ||
11607 | Change paragraph 2 to | |
11608 | ||
11609 | @quotation | |
11610 | With the exception of @code{__thread}, at most one storage-class | |
11611 | specifier may be given [@dots{}]. The @code{__thread} specifier may | |
11612 | be used alone, or immediately following @code{extern} or | |
11613 | @code{static}. | |
11614 | @end quotation | |
11615 | ||
11616 | Add new text after paragraph 6 | |
11617 | ||
11618 | @quotation | |
11619 | The declaration of an identifier for a variable that has | |
11620 | block scope that specifies @code{__thread} shall also | |
11621 | specify either @code{extern} or @code{static}. | |
11622 | ||
11623 | The @code{__thread} specifier shall be used only with | |
11624 | variables. | |
11625 | @end quotation | |
11626 | @end itemize | |
11627 | ||
11628 | @node C++98 Thread-Local Edits | |
11629 | @subsection ISO/IEC 14882:1998 Edits for Thread-Local Storage | |
11630 | ||
11631 | The following are a set of changes to ISO/IEC 14882:1998 (aka C++98) | |
11632 | that document the exact semantics of the language extension. | |
11633 | ||
11634 | @itemize @bullet | |
8d23a2c8 | 11635 | @item |
9217ef40 RH |
11636 | @b{[intro.execution]} |
11637 | ||
11638 | New text after paragraph 4 | |
11639 | ||
11640 | @quotation | |
11641 | A @dfn{thread} is a flow of control within the abstract machine. | |
11642 | It is implementation defined whether or not there may be more than | |
11643 | one thread. | |
11644 | @end quotation | |
11645 | ||
11646 | New text after paragraph 7 | |
11647 | ||
11648 | @quotation | |
95b1627e | 11649 | It is unspecified whether additional action must be taken to |
9217ef40 RH |
11650 | ensure when and whether side effects are visible to other threads. |
11651 | @end quotation | |
11652 | ||
11653 | @item | |
11654 | @b{[lex.key]} | |
11655 | ||
11656 | Add @code{__thread}. | |
11657 | ||
11658 | @item | |
11659 | @b{[basic.start.main]} | |
11660 | ||
11661 | Add after paragraph 5 | |
11662 | ||
11663 | @quotation | |
11664 | The thread that begins execution at the @code{main} function is called | |
95b1627e | 11665 | the @dfn{main thread}. It is implementation defined how functions |
9217ef40 RH |
11666 | beginning threads other than the main thread are designated or typed. |
11667 | A function so designated, as well as the @code{main} function, is called | |
11668 | a @dfn{thread startup function}. It is implementation defined what | |
11669 | happens if a thread startup function returns. It is implementation | |
11670 | defined what happens to other threads when any thread calls @code{exit}. | |
11671 | @end quotation | |
11672 | ||
11673 | @item | |
11674 | @b{[basic.start.init]} | |
11675 | ||
11676 | Add after paragraph 4 | |
11677 | ||
11678 | @quotation | |
11679 | The storage for an object of thread storage duration shall be | |
c0478a66 | 11680 | statically initialized before the first statement of the thread startup |
9217ef40 RH |
11681 | function. An object of thread storage duration shall not require |
11682 | dynamic initialization. | |
11683 | @end quotation | |
11684 | ||
11685 | @item | |
11686 | @b{[basic.start.term]} | |
11687 | ||
11688 | Add after paragraph 3 | |
11689 | ||
11690 | @quotation | |
244c2241 RH |
11691 | The type of an object with thread storage duration shall not have a |
11692 | non-trivial destructor, nor shall it be an array type whose elements | |
11693 | (directly or indirectly) have non-trivial destructors. | |
9217ef40 RH |
11694 | @end quotation |
11695 | ||
11696 | @item | |
11697 | @b{[basic.stc]} | |
11698 | ||
11699 | Add ``thread storage duration'' to the list in paragraph 1. | |
11700 | ||
11701 | Change paragraph 2 | |
11702 | ||
11703 | @quotation | |
11704 | Thread, static, and automatic storage durations are associated with | |
11705 | objects introduced by declarations [@dots{}]. | |
11706 | @end quotation | |
11707 | ||
11708 | Add @code{__thread} to the list of specifiers in paragraph 3. | |
11709 | ||
11710 | @item | |
11711 | @b{[basic.stc.thread]} | |
11712 | ||
11713 | New section before @b{[basic.stc.static]} | |
11714 | ||
11715 | @quotation | |
63519d23 | 11716 | The keyword @code{__thread} applied to a non-local object gives the |
9217ef40 RH |
11717 | object thread storage duration. |
11718 | ||
11719 | A local variable or class data member declared both @code{static} | |
11720 | and @code{__thread} gives the variable or member thread storage | |
11721 | duration. | |
11722 | @end quotation | |
11723 | ||
11724 | @item | |
11725 | @b{[basic.stc.static]} | |
11726 | ||
11727 | Change paragraph 1 | |
11728 | ||
11729 | @quotation | |
11730 | All objects which have neither thread storage duration, dynamic | |
11731 | storage duration nor are local [@dots{}]. | |
11732 | @end quotation | |
11733 | ||
11734 | @item | |
11735 | @b{[dcl.stc]} | |
11736 | ||
11737 | Add @code{__thread} to the list in paragraph 1. | |
11738 | ||
11739 | Change paragraph 1 | |
11740 | ||
11741 | @quotation | |
11742 | With the exception of @code{__thread}, at most one | |
11743 | @var{storage-class-specifier} shall appear in a given | |
11744 | @var{decl-specifier-seq}. The @code{__thread} specifier may | |
11745 | be used alone, or immediately following the @code{extern} or | |
11746 | @code{static} specifiers. [@dots{}] | |
11747 | @end quotation | |
11748 | ||
11749 | Add after paragraph 5 | |
11750 | ||
11751 | @quotation | |
11752 | The @code{__thread} specifier can be applied only to the names of objects | |
11753 | and to anonymous unions. | |
11754 | @end quotation | |
11755 | ||
11756 | @item | |
11757 | @b{[class.mem]} | |
11758 | ||
11759 | Add after paragraph 6 | |
11760 | ||
11761 | @quotation | |
11762 | Non-@code{static} members shall not be @code{__thread}. | |
11763 | @end quotation | |
11764 | @end itemize | |
11765 | ||
f7fd775f JW |
11766 | @node Binary constants |
11767 | @section Binary constants using the @samp{0b} prefix | |
11768 | @cindex Binary constants using the @samp{0b} prefix | |
11769 | ||
11770 | Integer constants can be written as binary constants, consisting of a | |
11771 | sequence of @samp{0} and @samp{1} digits, prefixed by @samp{0b} or | |
11772 | @samp{0B}. This is particularly useful in environments that operate a | |
11773 | lot on the bit-level (like microcontrollers). | |
11774 | ||
11775 | The following statements are identical: | |
11776 | ||
11777 | @smallexample | |
11778 | i = 42; | |
11779 | i = 0x2a; | |
11780 | i = 052; | |
11781 | i = 0b101010; | |
11782 | @end smallexample | |
11783 | ||
11784 | The type of these constants follows the same rules as for octal or | |
11785 | hexadecimal integer constants, so suffixes like @samp{L} or @samp{UL} | |
11786 | can be applied. | |
11787 | ||
c1f7febf RK |
11788 | @node C++ Extensions |
11789 | @chapter Extensions to the C++ Language | |
11790 | @cindex extensions, C++ language | |
11791 | @cindex C++ language extensions | |
11792 | ||
11793 | The GNU compiler provides these extensions to the C++ language (and you | |
11794 | can also use most of the C language extensions in your C++ programs). If you | |
11795 | want to write code that checks whether these features are available, you can | |
11796 | test for the GNU compiler the same way as for C programs: check for a | |
11797 | predefined macro @code{__GNUC__}. You can also use @code{__GNUG__} to | |
48795525 GP |
11798 | test specifically for GNU C++ (@pxref{Common Predefined Macros,, |
11799 | Predefined Macros,cpp,The GNU C Preprocessor}). | |
c1f7febf RK |
11800 | |
11801 | @menu | |
6ccde948 | 11802 | * Volatiles:: What constitutes an access to a volatile object. |
49419c8f | 11803 | * Restricted Pointers:: C99 restricted pointers and references. |
7a81cf7f | 11804 | * Vague Linkage:: Where G++ puts inlines, vtables and such. |
c1f7febf | 11805 | * C++ Interface:: You can use a single C++ header file for both |
e6f3b89d | 11806 | declarations and definitions. |
c1f7febf | 11807 | * Template Instantiation:: Methods for ensuring that exactly one copy of |
e6f3b89d | 11808 | each needed template instantiation is emitted. |
0ded1f18 JM |
11809 | * Bound member functions:: You can extract a function pointer to the |
11810 | method denoted by a @samp{->*} or @samp{.*} expression. | |
e6f3b89d | 11811 | * C++ Attributes:: Variable, function, and type attributes for C++ only. |
664a90c0 | 11812 | * Namespace Association:: Strong using-directives for namespace association. |
cb68ec50 | 11813 | * Type Traits:: Compiler support for type traits |
1f730ff7 | 11814 | * Java Exceptions:: Tweaking exception handling to work with Java. |
90ea7324 | 11815 | * Deprecated Features:: Things will disappear from g++. |
e6f3b89d | 11816 | * Backwards Compatibility:: Compatibilities with earlier definitions of C++. |
c1f7febf RK |
11817 | @end menu |
11818 | ||
02cac427 NS |
11819 | @node Volatiles |
11820 | @section When is a Volatile Object Accessed? | |
11821 | @cindex accessing volatiles | |
11822 | @cindex volatile read | |
11823 | @cindex volatile write | |
11824 | @cindex volatile access | |
11825 | ||
767094dd JM |
11826 | Both the C and C++ standard have the concept of volatile objects. These |
11827 | are normally accessed by pointers and used for accessing hardware. The | |
a9e64c63 EB |
11828 | standards encourage compilers to refrain from optimizations concerning |
11829 | accesses to volatile objects. The C standard leaves it implementation | |
11830 | defined as to what constitutes a volatile access. The C++ standard omits | |
11831 | to specify this, except to say that C++ should behave in a similar manner | |
767094dd | 11832 | to C with respect to volatiles, where possible. The minimum either |
8117da65 | 11833 | standard specifies is that at a sequence point all previous accesses to |
02cac427 | 11834 | volatile objects have stabilized and no subsequent accesses have |
767094dd | 11835 | occurred. Thus an implementation is free to reorder and combine |
02cac427 | 11836 | volatile accesses which occur between sequence points, but cannot do so |
767094dd | 11837 | for accesses across a sequence point. The use of volatiles does not |
02cac427 NS |
11838 | allow you to violate the restriction on updating objects multiple times |
11839 | within a sequence point. | |
11840 | ||
a9e64c63 | 11841 | @xref{Qualifiers implementation, , Volatile qualifier and the C compiler}. |
02cac427 | 11842 | |
a9e64c63 | 11843 | The behavior differs slightly between C and C++ in the non-obvious cases: |
02cac427 | 11844 | |
3ab51846 | 11845 | @smallexample |
c771326b | 11846 | volatile int *src = @var{somevalue}; |
02cac427 | 11847 | *src; |
3ab51846 | 11848 | @end smallexample |
02cac427 | 11849 | |
a9e64c63 EB |
11850 | With C, such expressions are rvalues, and GCC interprets this either as a |
11851 | read of the volatile object being pointed to or only as request to evaluate | |
11852 | the side-effects. The C++ standard specifies that such expressions do not | |
11853 | undergo lvalue to rvalue conversion, and that the type of the dereferenced | |
767094dd | 11854 | object may be incomplete. The C++ standard does not specify explicitly |
a9e64c63 | 11855 | that it is this lvalue to rvalue conversion which may be responsible for |
767094dd JM |
11856 | causing an access. However, there is reason to believe that it is, |
11857 | because otherwise certain simple expressions become undefined. However, | |
f0523f02 | 11858 | because it would surprise most programmers, G++ treats dereferencing a |
a9e64c63 | 11859 | pointer to volatile object of complete type when the value is unused as |
0ee2ea09 | 11860 | GCC would do for an equivalent type in C@. When the object has incomplete |
a9e64c63 EB |
11861 | type, G++ issues a warning; if you wish to force an error, you must |
11862 | force a conversion to rvalue with, for instance, a static cast. | |
02cac427 | 11863 | |
f0523f02 | 11864 | When using a reference to volatile, G++ does not treat equivalent |
02cac427 | 11865 | expressions as accesses to volatiles, but instead issues a warning that |
767094dd | 11866 | no volatile is accessed. The rationale for this is that otherwise it |
02cac427 NS |
11867 | becomes difficult to determine where volatile access occur, and not |
11868 | possible to ignore the return value from functions returning volatile | |
767094dd | 11869 | references. Again, if you wish to force a read, cast the reference to |
02cac427 NS |
11870 | an rvalue. |
11871 | ||
535233a8 NS |
11872 | @node Restricted Pointers |
11873 | @section Restricting Pointer Aliasing | |
11874 | @cindex restricted pointers | |
11875 | @cindex restricted references | |
11876 | @cindex restricted this pointer | |
11877 | ||
2dd76960 | 11878 | As with the C front end, G++ understands the C99 feature of restricted pointers, |
535233a8 | 11879 | specified with the @code{__restrict__}, or @code{__restrict} type |
767094dd | 11880 | qualifier. Because you cannot compile C++ by specifying the @option{-std=c99} |
535233a8 NS |
11881 | language flag, @code{restrict} is not a keyword in C++. |
11882 | ||
11883 | In addition to allowing restricted pointers, you can specify restricted | |
11884 | references, which indicate that the reference is not aliased in the local | |
11885 | context. | |
11886 | ||
3ab51846 | 11887 | @smallexample |
535233a8 NS |
11888 | void fn (int *__restrict__ rptr, int &__restrict__ rref) |
11889 | @{ | |
0d893a63 | 11890 | /* @r{@dots{}} */ |
535233a8 | 11891 | @} |
3ab51846 | 11892 | @end smallexample |
535233a8 NS |
11893 | |
11894 | @noindent | |
11895 | In the body of @code{fn}, @var{rptr} points to an unaliased integer and | |
11896 | @var{rref} refers to a (different) unaliased integer. | |
11897 | ||
11898 | You may also specify whether a member function's @var{this} pointer is | |
11899 | unaliased by using @code{__restrict__} as a member function qualifier. | |
11900 | ||
3ab51846 | 11901 | @smallexample |
535233a8 NS |
11902 | void T::fn () __restrict__ |
11903 | @{ | |
0d893a63 | 11904 | /* @r{@dots{}} */ |
535233a8 | 11905 | @} |
3ab51846 | 11906 | @end smallexample |
535233a8 NS |
11907 | |
11908 | @noindent | |
11909 | Within the body of @code{T::fn}, @var{this} will have the effective | |
767094dd | 11910 | definition @code{T *__restrict__ const this}. Notice that the |
535233a8 NS |
11911 | interpretation of a @code{__restrict__} member function qualifier is |
11912 | different to that of @code{const} or @code{volatile} qualifier, in that it | |
767094dd | 11913 | is applied to the pointer rather than the object. This is consistent with |
535233a8 NS |
11914 | other compilers which implement restricted pointers. |
11915 | ||
11916 | As with all outermost parameter qualifiers, @code{__restrict__} is | |
767094dd | 11917 | ignored in function definition matching. This means you only need to |
535233a8 NS |
11918 | specify @code{__restrict__} in a function definition, rather than |
11919 | in a function prototype as well. | |
11920 | ||
7a81cf7f JM |
11921 | @node Vague Linkage |
11922 | @section Vague Linkage | |
11923 | @cindex vague linkage | |
11924 | ||
11925 | There are several constructs in C++ which require space in the object | |
11926 | file but are not clearly tied to a single translation unit. We say that | |
11927 | these constructs have ``vague linkage''. Typically such constructs are | |
11928 | emitted wherever they are needed, though sometimes we can be more | |
11929 | clever. | |
11930 | ||
11931 | @table @asis | |
11932 | @item Inline Functions | |
11933 | Inline functions are typically defined in a header file which can be | |
11934 | included in many different compilations. Hopefully they can usually be | |
11935 | inlined, but sometimes an out-of-line copy is necessary, if the address | |
11936 | of the function is taken or if inlining fails. In general, we emit an | |
11937 | out-of-line copy in all translation units where one is needed. As an | |
11938 | exception, we only emit inline virtual functions with the vtable, since | |
11939 | it will always require a copy. | |
11940 | ||
11941 | Local static variables and string constants used in an inline function | |
11942 | are also considered to have vague linkage, since they must be shared | |
11943 | between all inlined and out-of-line instances of the function. | |
11944 | ||
11945 | @item VTables | |
11946 | @cindex vtable | |
11947 | C++ virtual functions are implemented in most compilers using a lookup | |
11948 | table, known as a vtable. The vtable contains pointers to the virtual | |
11949 | functions provided by a class, and each object of the class contains a | |
11950 | pointer to its vtable (or vtables, in some multiple-inheritance | |
11951 | situations). If the class declares any non-inline, non-pure virtual | |
11952 | functions, the first one is chosen as the ``key method'' for the class, | |
11953 | and the vtable is only emitted in the translation unit where the key | |
11954 | method is defined. | |
11955 | ||
11956 | @emph{Note:} If the chosen key method is later defined as inline, the | |
11957 | vtable will still be emitted in every translation unit which defines it. | |
11958 | Make sure that any inline virtuals are declared inline in the class | |
11959 | body, even if they are not defined there. | |
11960 | ||
11961 | @item type_info objects | |
11962 | @cindex type_info | |
11963 | @cindex RTTI | |
11964 | C++ requires information about types to be written out in order to | |
11965 | implement @samp{dynamic_cast}, @samp{typeid} and exception handling. | |
11966 | For polymorphic classes (classes with virtual functions), the type_info | |
11967 | object is written out along with the vtable so that @samp{dynamic_cast} | |
11968 | can determine the dynamic type of a class object at runtime. For all | |
11969 | other types, we write out the type_info object when it is used: when | |
11970 | applying @samp{typeid} to an expression, throwing an object, or | |
11971 | referring to a type in a catch clause or exception specification. | |
11972 | ||
11973 | @item Template Instantiations | |
11974 | Most everything in this section also applies to template instantiations, | |
11975 | but there are other options as well. | |
11976 | @xref{Template Instantiation,,Where's the Template?}. | |
11977 | ||
11978 | @end table | |
11979 | ||
11980 | When used with GNU ld version 2.8 or later on an ELF system such as | |
95fef11f | 11981 | GNU/Linux or Solaris 2, or on Microsoft Windows, duplicate copies of |
7a81cf7f JM |
11982 | these constructs will be discarded at link time. This is known as |
11983 | COMDAT support. | |
11984 | ||
11985 | On targets that don't support COMDAT, but do support weak symbols, GCC | |
11986 | will use them. This way one copy will override all the others, but | |
11987 | the unused copies will still take up space in the executable. | |
11988 | ||
11989 | For targets which do not support either COMDAT or weak symbols, | |
11990 | most entities with vague linkage will be emitted as local symbols to | |
11991 | avoid duplicate definition errors from the linker. This will not happen | |
11992 | for local statics in inlines, however, as having multiple copies will | |
11993 | almost certainly break things. | |
11994 | ||
11995 | @xref{C++ Interface,,Declarations and Definitions in One Header}, for | |
11996 | another way to control placement of these constructs. | |
11997 | ||
c1f7febf | 11998 | @node C++ Interface |
fc72b380 | 11999 | @section #pragma interface and implementation |
c1f7febf RK |
12000 | |
12001 | @cindex interface and implementation headers, C++ | |
12002 | @cindex C++ interface and implementation headers | |
c1f7febf | 12003 | @cindex pragmas, interface and implementation |
c1f7febf | 12004 | |
fc72b380 JM |
12005 | @code{#pragma interface} and @code{#pragma implementation} provide the |
12006 | user with a way of explicitly directing the compiler to emit entities | |
12007 | with vague linkage (and debugging information) in a particular | |
12008 | translation unit. | |
c1f7febf | 12009 | |
fc72b380 JM |
12010 | @emph{Note:} As of GCC 2.7.2, these @code{#pragma}s are not useful in |
12011 | most cases, because of COMDAT support and the ``key method'' heuristic | |
12012 | mentioned in @ref{Vague Linkage}. Using them can actually cause your | |
27ef2cdd | 12013 | program to grow due to unnecessary out-of-line copies of inline |
fc72b380 JM |
12014 | functions. Currently (3.4) the only benefit of these |
12015 | @code{#pragma}s is reduced duplication of debugging information, and | |
12016 | that should be addressed soon on DWARF 2 targets with the use of | |
12017 | COMDAT groups. | |
c1f7febf RK |
12018 | |
12019 | @table @code | |
12020 | @item #pragma interface | |
12021 | @itemx #pragma interface "@var{subdir}/@var{objects}.h" | |
12022 | @kindex #pragma interface | |
12023 | Use this directive in @emph{header files} that define object classes, to save | |
12024 | space in most of the object files that use those classes. Normally, | |
12025 | local copies of certain information (backup copies of inline member | |
12026 | functions, debugging information, and the internal tables that implement | |
12027 | virtual functions) must be kept in each object file that includes class | |
12028 | definitions. You can use this pragma to avoid such duplication. When a | |
12029 | header file containing @samp{#pragma interface} is included in a | |
12030 | compilation, this auxiliary information will not be generated (unless | |
12031 | the main input source file itself uses @samp{#pragma implementation}). | |
12032 | Instead, the object files will contain references to be resolved at link | |
12033 | time. | |
12034 | ||
12035 | The second form of this directive is useful for the case where you have | |
12036 | multiple headers with the same name in different directories. If you | |
12037 | use this form, you must specify the same string to @samp{#pragma | |
12038 | implementation}. | |
12039 | ||
12040 | @item #pragma implementation | |
12041 | @itemx #pragma implementation "@var{objects}.h" | |
12042 | @kindex #pragma implementation | |
12043 | Use this pragma in a @emph{main input file}, when you want full output from | |
12044 | included header files to be generated (and made globally visible). The | |
12045 | included header file, in turn, should use @samp{#pragma interface}. | |
12046 | Backup copies of inline member functions, debugging information, and the | |
12047 | internal tables used to implement virtual functions are all generated in | |
12048 | implementation files. | |
12049 | ||
12050 | @cindex implied @code{#pragma implementation} | |
12051 | @cindex @code{#pragma implementation}, implied | |
12052 | @cindex naming convention, implementation headers | |
12053 | If you use @samp{#pragma implementation} with no argument, it applies to | |
12054 | an include file with the same basename@footnote{A file's @dfn{basename} | |
12055 | was the name stripped of all leading path information and of trailing | |
12056 | suffixes, such as @samp{.h} or @samp{.C} or @samp{.cc}.} as your source | |
12057 | file. For example, in @file{allclass.cc}, giving just | |
12058 | @samp{#pragma implementation} | |
12059 | by itself is equivalent to @samp{#pragma implementation "allclass.h"}. | |
12060 | ||
12061 | In versions of GNU C++ prior to 2.6.0 @file{allclass.h} was treated as | |
12062 | an implementation file whenever you would include it from | |
12063 | @file{allclass.cc} even if you never specified @samp{#pragma | |
12064 | implementation}. This was deemed to be more trouble than it was worth, | |
12065 | however, and disabled. | |
12066 | ||
c1f7febf RK |
12067 | Use the string argument if you want a single implementation file to |
12068 | include code from multiple header files. (You must also use | |
12069 | @samp{#include} to include the header file; @samp{#pragma | |
12070 | implementation} only specifies how to use the file---it doesn't actually | |
12071 | include it.) | |
12072 | ||
12073 | There is no way to split up the contents of a single header file into | |
12074 | multiple implementation files. | |
12075 | @end table | |
12076 | ||
12077 | @cindex inlining and C++ pragmas | |
12078 | @cindex C++ pragmas, effect on inlining | |
12079 | @cindex pragmas in C++, effect on inlining | |
12080 | @samp{#pragma implementation} and @samp{#pragma interface} also have an | |
12081 | effect on function inlining. | |
12082 | ||
12083 | If you define a class in a header file marked with @samp{#pragma | |
fc72b380 JM |
12084 | interface}, the effect on an inline function defined in that class is |
12085 | similar to an explicit @code{extern} declaration---the compiler emits | |
12086 | no code at all to define an independent version of the function. Its | |
12087 | definition is used only for inlining with its callers. | |
c1f7febf | 12088 | |
84330467 | 12089 | @opindex fno-implement-inlines |
c1f7febf RK |
12090 | Conversely, when you include the same header file in a main source file |
12091 | that declares it as @samp{#pragma implementation}, the compiler emits | |
12092 | code for the function itself; this defines a version of the function | |
12093 | that can be found via pointers (or by callers compiled without | |
12094 | inlining). If all calls to the function can be inlined, you can avoid | |
84330467 | 12095 | emitting the function by compiling with @option{-fno-implement-inlines}. |
c1f7febf RK |
12096 | If any calls were not inlined, you will get linker errors. |
12097 | ||
12098 | @node Template Instantiation | |
12099 | @section Where's the Template? | |
c1f7febf RK |
12100 | @cindex template instantiation |
12101 | ||
12102 | C++ templates are the first language feature to require more | |
12103 | intelligence from the environment than one usually finds on a UNIX | |
12104 | system. Somehow the compiler and linker have to make sure that each | |
12105 | template instance occurs exactly once in the executable if it is needed, | |
12106 | and not at all otherwise. There are two basic approaches to this | |
962e6e00 | 12107 | problem, which are referred to as the Borland model and the Cfront model. |
c1f7febf RK |
12108 | |
12109 | @table @asis | |
12110 | @item Borland model | |
12111 | Borland C++ solved the template instantiation problem by adding the code | |
469b759e JM |
12112 | equivalent of common blocks to their linker; the compiler emits template |
12113 | instances in each translation unit that uses them, and the linker | |
12114 | collapses them together. The advantage of this model is that the linker | |
12115 | only has to consider the object files themselves; there is no external | |
12116 | complexity to worry about. This disadvantage is that compilation time | |
12117 | is increased because the template code is being compiled repeatedly. | |
12118 | Code written for this model tends to include definitions of all | |
12119 | templates in the header file, since they must be seen to be | |
12120 | instantiated. | |
c1f7febf RK |
12121 | |
12122 | @item Cfront model | |
12123 | The AT&T C++ translator, Cfront, solved the template instantiation | |
12124 | problem by creating the notion of a template repository, an | |
469b759e JM |
12125 | automatically maintained place where template instances are stored. A |
12126 | more modern version of the repository works as follows: As individual | |
12127 | object files are built, the compiler places any template definitions and | |
12128 | instantiations encountered in the repository. At link time, the link | |
12129 | wrapper adds in the objects in the repository and compiles any needed | |
12130 | instances that were not previously emitted. The advantages of this | |
12131 | model are more optimal compilation speed and the ability to use the | |
12132 | system linker; to implement the Borland model a compiler vendor also | |
c1f7febf | 12133 | needs to replace the linker. The disadvantages are vastly increased |
469b759e JM |
12134 | complexity, and thus potential for error; for some code this can be |
12135 | just as transparent, but in practice it can been very difficult to build | |
c1f7febf | 12136 | multiple programs in one directory and one program in multiple |
469b759e JM |
12137 | directories. Code written for this model tends to separate definitions |
12138 | of non-inline member templates into a separate file, which should be | |
12139 | compiled separately. | |
c1f7febf RK |
12140 | @end table |
12141 | ||
469b759e | 12142 | When used with GNU ld version 2.8 or later on an ELF system such as |
2dd76960 JM |
12143 | GNU/Linux or Solaris 2, or on Microsoft Windows, G++ supports the |
12144 | Borland model. On other systems, G++ implements neither automatic | |
a4b3b54a | 12145 | model. |
469b759e | 12146 | |
2dd76960 | 12147 | A future version of G++ will support a hybrid model whereby the compiler |
469b759e JM |
12148 | will emit any instantiations for which the template definition is |
12149 | included in the compile, and store template definitions and | |
12150 | instantiation context information into the object file for the rest. | |
12151 | The link wrapper will extract that information as necessary and invoke | |
12152 | the compiler to produce the remaining instantiations. The linker will | |
12153 | then combine duplicate instantiations. | |
12154 | ||
12155 | In the mean time, you have the following options for dealing with | |
12156 | template instantiations: | |
c1f7febf RK |
12157 | |
12158 | @enumerate | |
d863830b | 12159 | @item |
84330467 JM |
12160 | @opindex frepo |
12161 | Compile your template-using code with @option{-frepo}. The compiler will | |
d863830b JL |
12162 | generate files with the extension @samp{.rpo} listing all of the |
12163 | template instantiations used in the corresponding object files which | |
12164 | could be instantiated there; the link wrapper, @samp{collect2}, will | |
12165 | then update the @samp{.rpo} files to tell the compiler where to place | |
12166 | those instantiations and rebuild any affected object files. The | |
12167 | link-time overhead is negligible after the first pass, as the compiler | |
12168 | will continue to place the instantiations in the same files. | |
12169 | ||
12170 | This is your best option for application code written for the Borland | |
12171 | model, as it will just work. Code written for the Cfront model will | |
12172 | need to be modified so that the template definitions are available at | |
12173 | one or more points of instantiation; usually this is as simple as adding | |
12174 | @code{#include <tmethods.cc>} to the end of each template header. | |
12175 | ||
12176 | For library code, if you want the library to provide all of the template | |
12177 | instantiations it needs, just try to link all of its object files | |
12178 | together; the link will fail, but cause the instantiations to be | |
12179 | generated as a side effect. Be warned, however, that this may cause | |
12180 | conflicts if multiple libraries try to provide the same instantiations. | |
12181 | For greater control, use explicit instantiation as described in the next | |
12182 | option. | |
12183 | ||
c1f7febf | 12184 | @item |
84330467 JM |
12185 | @opindex fno-implicit-templates |
12186 | Compile your code with @option{-fno-implicit-templates} to disable the | |
c1f7febf RK |
12187 | implicit generation of template instances, and explicitly instantiate |
12188 | all the ones you use. This approach requires more knowledge of exactly | |
12189 | which instances you need than do the others, but it's less | |
12190 | mysterious and allows greater control. You can scatter the explicit | |
12191 | instantiations throughout your program, perhaps putting them in the | |
12192 | translation units where the instances are used or the translation units | |
12193 | that define the templates themselves; you can put all of the explicit | |
12194 | instantiations you need into one big file; or you can create small files | |
12195 | like | |
12196 | ||
3ab51846 | 12197 | @smallexample |
c1f7febf RK |
12198 | #include "Foo.h" |
12199 | #include "Foo.cc" | |
12200 | ||
12201 | template class Foo<int>; | |
12202 | template ostream& operator << | |
12203 | (ostream&, const Foo<int>&); | |
3ab51846 | 12204 | @end smallexample |
c1f7febf RK |
12205 | |
12206 | for each of the instances you need, and create a template instantiation | |
12207 | library from those. | |
12208 | ||
12209 | If you are using Cfront-model code, you can probably get away with not | |
84330467 | 12210 | using @option{-fno-implicit-templates} when compiling files that don't |
c1f7febf RK |
12211 | @samp{#include} the member template definitions. |
12212 | ||
12213 | If you use one big file to do the instantiations, you may want to | |
84330467 | 12214 | compile it without @option{-fno-implicit-templates} so you get all of the |
c1f7febf RK |
12215 | instances required by your explicit instantiations (but not by any |
12216 | other files) without having to specify them as well. | |
12217 | ||
2dd76960 | 12218 | G++ has extended the template instantiation syntax given in the ISO |
6d9c4c83 | 12219 | standard to allow forward declaration of explicit instantiations |
4003d7f9 | 12220 | (with @code{extern}), instantiation of the compiler support data for a |
e979f9e8 | 12221 | template class (i.e.@: the vtable) without instantiating any of its |
4003d7f9 JM |
12222 | members (with @code{inline}), and instantiation of only the static data |
12223 | members of a template class, without the support data or member | |
12224 | functions (with (@code{static}): | |
c1f7febf | 12225 | |
3ab51846 | 12226 | @smallexample |
c1f7febf | 12227 | extern template int max (int, int); |
c1f7febf | 12228 | inline template class Foo<int>; |
4003d7f9 | 12229 | static template class Foo<int>; |
3ab51846 | 12230 | @end smallexample |
c1f7febf RK |
12231 | |
12232 | @item | |
2dd76960 | 12233 | Do nothing. Pretend G++ does implement automatic instantiation |
c1f7febf RK |
12234 | management. Code written for the Borland model will work fine, but |
12235 | each translation unit will contain instances of each of the templates it | |
12236 | uses. In a large program, this can lead to an unacceptable amount of code | |
12237 | duplication. | |
c1f7febf RK |
12238 | @end enumerate |
12239 | ||
0ded1f18 JM |
12240 | @node Bound member functions |
12241 | @section Extracting the function pointer from a bound pointer to member function | |
0ded1f18 JM |
12242 | @cindex pmf |
12243 | @cindex pointer to member function | |
12244 | @cindex bound pointer to member function | |
12245 | ||
12246 | In C++, pointer to member functions (PMFs) are implemented using a wide | |
12247 | pointer of sorts to handle all the possible call mechanisms; the PMF | |
12248 | needs to store information about how to adjust the @samp{this} pointer, | |
12249 | and if the function pointed to is virtual, where to find the vtable, and | |
12250 | where in the vtable to look for the member function. If you are using | |
12251 | PMFs in an inner loop, you should really reconsider that decision. If | |
12252 | that is not an option, you can extract the pointer to the function that | |
12253 | would be called for a given object/PMF pair and call it directly inside | |
12254 | the inner loop, to save a bit of time. | |
12255 | ||
12256 | Note that you will still be paying the penalty for the call through a | |
12257 | function pointer; on most modern architectures, such a call defeats the | |
161d7b59 | 12258 | branch prediction features of the CPU@. This is also true of normal |
0ded1f18 JM |
12259 | virtual function calls. |
12260 | ||
12261 | The syntax for this extension is | |
12262 | ||
3ab51846 | 12263 | @smallexample |
0ded1f18 JM |
12264 | extern A a; |
12265 | extern int (A::*fp)(); | |
12266 | typedef int (*fptr)(A *); | |
12267 | ||
12268 | fptr p = (fptr)(a.*fp); | |
3ab51846 | 12269 | @end smallexample |
0ded1f18 | 12270 | |
e979f9e8 | 12271 | For PMF constants (i.e.@: expressions of the form @samp{&Klasse::Member}), |
767094dd | 12272 | no object is needed to obtain the address of the function. They can be |
0fb6bbf5 ML |
12273 | converted to function pointers directly: |
12274 | ||
3ab51846 | 12275 | @smallexample |
0fb6bbf5 | 12276 | fptr p1 = (fptr)(&A::foo); |
3ab51846 | 12277 | @end smallexample |
0fb6bbf5 | 12278 | |
84330467 JM |
12279 | @opindex Wno-pmf-conversions |
12280 | You must specify @option{-Wno-pmf-conversions} to use this extension. | |
0ded1f18 | 12281 | |
5c25e11d PE |
12282 | @node C++ Attributes |
12283 | @section C++-Specific Variable, Function, and Type Attributes | |
12284 | ||
12285 | Some attributes only make sense for C++ programs. | |
12286 | ||
12287 | @table @code | |
12288 | @item init_priority (@var{priority}) | |
12289 | @cindex init_priority attribute | |
12290 | ||
12291 | ||
12292 | In Standard C++, objects defined at namespace scope are guaranteed to be | |
12293 | initialized in an order in strict accordance with that of their definitions | |
12294 | @emph{in a given translation unit}. No guarantee is made for initializations | |
12295 | across translation units. However, GNU C++ allows users to control the | |
3844cd2e | 12296 | order of initialization of objects defined at namespace scope with the |
5c25e11d PE |
12297 | @code{init_priority} attribute by specifying a relative @var{priority}, |
12298 | a constant integral expression currently bounded between 101 and 65535 | |
12299 | inclusive. Lower numbers indicate a higher priority. | |
12300 | ||
12301 | In the following example, @code{A} would normally be created before | |
12302 | @code{B}, but the @code{init_priority} attribute has reversed that order: | |
12303 | ||
478c9e72 | 12304 | @smallexample |
5c25e11d PE |
12305 | Some_Class A __attribute__ ((init_priority (2000))); |
12306 | Some_Class B __attribute__ ((init_priority (543))); | |
478c9e72 | 12307 | @end smallexample |
5c25e11d PE |
12308 | |
12309 | @noindent | |
12310 | Note that the particular values of @var{priority} do not matter; only their | |
12311 | relative ordering. | |
12312 | ||
60c87482 BM |
12313 | @item java_interface |
12314 | @cindex java_interface attribute | |
12315 | ||
02f52e19 | 12316 | This type attribute informs C++ that the class is a Java interface. It may |
60c87482 | 12317 | only be applied to classes declared within an @code{extern "Java"} block. |
02f52e19 AJ |
12318 | Calls to methods declared in this interface will be dispatched using GCJ's |
12319 | interface table mechanism, instead of regular virtual table dispatch. | |
60c87482 | 12320 | |
5c25e11d PE |
12321 | @end table |
12322 | ||
38bb2b65 | 12323 | See also @ref{Namespace Association}. |
86098eb8 | 12324 | |
664a90c0 JM |
12325 | @node Namespace Association |
12326 | @section Namespace Association | |
86098eb8 | 12327 | |
fea77ed9 MM |
12328 | @strong{Caution:} The semantics of this extension are not fully |
12329 | defined. Users should refrain from using this extension as its | |
12330 | semantics may change subtly over time. It is possible that this | |
664a90c0 | 12331 | extension will be removed in future versions of G++. |
fea77ed9 | 12332 | |
86098eb8 JM |
12333 | A using-directive with @code{__attribute ((strong))} is stronger |
12334 | than a normal using-directive in two ways: | |
12335 | ||
12336 | @itemize @bullet | |
12337 | @item | |
664a90c0 JM |
12338 | Templates from the used namespace can be specialized and explicitly |
12339 | instantiated as though they were members of the using namespace. | |
86098eb8 JM |
12340 | |
12341 | @item | |
12342 | The using namespace is considered an associated namespace of all | |
12343 | templates in the used namespace for purposes of argument-dependent | |
12344 | name lookup. | |
12345 | @end itemize | |
12346 | ||
664a90c0 JM |
12347 | The used namespace must be nested within the using namespace so that |
12348 | normal unqualified lookup works properly. | |
12349 | ||
86098eb8 JM |
12350 | This is useful for composing a namespace transparently from |
12351 | implementation namespaces. For example: | |
12352 | ||
12353 | @smallexample | |
12354 | namespace std @{ | |
12355 | namespace debug @{ | |
12356 | template <class T> struct A @{ @}; | |
12357 | @} | |
12358 | using namespace debug __attribute ((__strong__)); | |
cd1a8088 | 12359 | template <> struct A<int> @{ @}; // @r{ok to specialize} |
86098eb8 JM |
12360 | |
12361 | template <class T> void f (A<T>); | |
12362 | @} | |
12363 | ||
12364 | int main() | |
12365 | @{ | |
cd1a8088 | 12366 | f (std::A<float>()); // @r{lookup finds} std::f |
86098eb8 JM |
12367 | f (std::A<int>()); |
12368 | @} | |
12369 | @end smallexample | |
12370 | ||
cb68ec50 PC |
12371 | @node Type Traits |
12372 | @section Type Traits | |
12373 | ||
12374 | The C++ front-end implements syntactic extensions that allow to | |
12375 | determine at compile time various characteristics of a type (or of a | |
12376 | pair of types). | |
12377 | ||
12378 | @table @code | |
12379 | @item __has_nothrow_assign (type) | |
b29441ec PC |
12380 | If @code{type} is const qualified or is a reference type then the trait is |
12381 | false. Otherwise if @code{__has_trivial_assign (type)} is true then the trait | |
12382 | is true, else if @code{type} is a cv class or union type with copy assignment | |
12383 | operators that are known not to throw an exception then the trait is true, | |
12384 | else it is false. Requires: @code{type} shall be a complete type, an array | |
12385 | type of unknown bound, or is a @code{void} type. | |
cb68ec50 PC |
12386 | |
12387 | @item __has_nothrow_copy (type) | |
12388 | If @code{__has_trivial_copy (type)} is true then the trait is true, else if | |
12389 | @code{type} is a cv class or union type with copy constructors that | |
12390 | are known not to throw an exception then the trait is true, else it is false. | |
12391 | Requires: @code{type} shall be a complete type, an array type of | |
12392 | unknown bound, or is a @code{void} type. | |
12393 | ||
12394 | @item __has_nothrow_constructor (type) | |
12395 | If @code{__has_trivial_constructor (type)} is true then the trait is | |
12396 | true, else if @code{type} is a cv class or union type (or array | |
12397 | thereof) with a default constructor that is known not to throw an | |
12398 | exception then the trait is true, else it is false. Requires: | |
12399 | @code{type} shall be a complete type, an array type of unknown bound, | |
12400 | or is a @code{void} type. | |
12401 | ||
12402 | @item __has_trivial_assign (type) | |
12403 | If @code{type} is const qualified or is a reference type then the trait is | |
12404 | false. Otherwise if @code{__is_pod (type)} is true then the trait is | |
12405 | true, else if @code{type} is a cv class or union type with a trivial | |
12406 | copy assignment ([class.copy]) then the trait is true, else it is | |
12407 | false. Requires: @code{type} shall be a complete type, an array type | |
12408 | of unknown bound, or is a @code{void} type. | |
12409 | ||
12410 | @item __has_trivial_copy (type) | |
12411 | If @code{__is_pod (type)} is true or @code{type} is a reference type | |
12412 | then the trait is true, else if @code{type} is a cv class or union type | |
12413 | with a trivial copy constructor ([class.copy]) then the trait | |
12414 | is true, else it is false. Requires: @code{type} shall be a complete | |
12415 | type, an array type of unknown bound, or is a @code{void} type. | |
12416 | ||
12417 | @item __has_trivial_constructor (type) | |
12418 | If @code{__is_pod (type)} is true then the trait is true, else if | |
12419 | @code{type} is a cv class or union type (or array thereof) with a | |
12420 | trivial default constructor ([class.ctor]) then the trait is true, | |
12421 | else it is false. Requires: @code{type} shall be a complete type, an | |
12422 | array type of unknown bound, or is a @code{void} type. | |
12423 | ||
12424 | @item __has_trivial_destructor (type) | |
12425 | If @code{__is_pod (type)} is true or @code{type} is a reference type then | |
12426 | the trait is true, else if @code{type} is a cv class or union type (or | |
12427 | array thereof) with a trivial destructor ([class.dtor]) then the trait | |
12428 | is true, else it is false. Requires: @code{type} shall be a complete | |
12429 | type, an array type of unknown bound, or is a @code{void} type. | |
12430 | ||
12431 | @item __has_virtual_destructor (type) | |
12432 | If @code{type} is a class type with a virtual destructor | |
12433 | ([class.dtor]) then the trait is true, else it is false. Requires: | |
12434 | @code{type} shall be a complete type, an array type of unknown bound, | |
12435 | or is a @code{void} type. | |
12436 | ||
12437 | @item __is_abstract (type) | |
12438 | If @code{type} is an abstract class ([class.abstract]) then the trait | |
12439 | is true, else it is false. Requires: @code{type} shall be a complete | |
12440 | type, an array type of unknown bound, or is a @code{void} type. | |
12441 | ||
12442 | @item __is_base_of (base_type, derived_type) | |
12443 | If @code{base_type} is a base class of @code{derived_type} | |
12444 | ([class.derived]) then the trait is true, otherwise it is false. | |
12445 | Top-level cv qualifications of @code{base_type} and | |
12446 | @code{derived_type} are ignored. For the purposes of this trait, a | |
12447 | class type is considered is own base. Requires: if @code{__is_class | |
12448 | (base_type)} and @code{__is_class (derived_type)} are true and | |
12449 | @code{base_type} and @code{derived_type} are not the same type | |
12450 | (disregarding cv-qualifiers), @code{derived_type} shall be a complete | |
12451 | type. Diagnostic is produced if this requirement is not met. | |
12452 | ||
12453 | @item __is_class (type) | |
12454 | If @code{type} is a cv class type, and not a union type | |
12455 | ([basic.compound]) the the trait is true, else it is false. | |
12456 | ||
12457 | @item __is_empty (type) | |
12458 | If @code{__is_class (type)} is false then the trait is false. | |
12459 | Otherwise @code{type} is considered empty if and only if: @code{type} | |
12460 | has no non-static data members, or all non-static data members, if | |
12461 | any, are bit-fields of lenght 0, and @code{type} has no virtual | |
12462 | members, and @code{type} has no virtual base classes, and @code{type} | |
12463 | has no base classes @code{base_type} for which | |
12464 | @code{__is_empty (base_type)} is false. Requires: @code{type} shall | |
12465 | be a complete type, an array type of unknown bound, or is a | |
12466 | @code{void} type. | |
12467 | ||
12468 | @item __is_enum (type) | |
12469 | If @code{type} is a cv enumeration type ([basic.compound]) the the trait is | |
12470 | true, else it is false. | |
12471 | ||
12472 | @item __is_pod (type) | |
12473 | If @code{type} is a cv POD type ([basic.types]) then the trait is true, | |
12474 | else it is false. Requires: @code{type} shall be a complete type, | |
12475 | an array type of unknown bound, or is a @code{void} type. | |
12476 | ||
12477 | @item __is_polymorphic (type) | |
12478 | If @code{type} is a polymorphic class ([class.virtual]) then the trait | |
12479 | is true, else it is false. Requires: @code{type} shall be a complete | |
12480 | type, an array type of unknown bound, or is a @code{void} type. | |
12481 | ||
12482 | @item __is_union (type) | |
12483 | If @code{type} is a cv union type ([basic.compound]) the the trait is | |
12484 | true, else it is false. | |
12485 | ||
12486 | @end table | |
12487 | ||
1f730ff7 ZW |
12488 | @node Java Exceptions |
12489 | @section Java Exceptions | |
12490 | ||
12491 | The Java language uses a slightly different exception handling model | |
12492 | from C++. Normally, GNU C++ will automatically detect when you are | |
12493 | writing C++ code that uses Java exceptions, and handle them | |
12494 | appropriately. However, if C++ code only needs to execute destructors | |
12495 | when Java exceptions are thrown through it, GCC will guess incorrectly. | |
9c34dbbf | 12496 | Sample problematic code is: |
1f730ff7 | 12497 | |
478c9e72 | 12498 | @smallexample |
1f730ff7 | 12499 | struct S @{ ~S(); @}; |
cd1a8088 | 12500 | extern void bar(); // @r{is written in Java, and may throw exceptions} |
1f730ff7 ZW |
12501 | void foo() |
12502 | @{ | |
12503 | S s; | |
12504 | bar(); | |
12505 | @} | |
478c9e72 | 12506 | @end smallexample |
1f730ff7 ZW |
12507 | |
12508 | @noindent | |
12509 | The usual effect of an incorrect guess is a link failure, complaining of | |
12510 | a missing routine called @samp{__gxx_personality_v0}. | |
12511 | ||
12512 | You can inform the compiler that Java exceptions are to be used in a | |
12513 | translation unit, irrespective of what it might think, by writing | |
12514 | @samp{@w{#pragma GCC java_exceptions}} at the head of the file. This | |
12515 | @samp{#pragma} must appear before any functions that throw or catch | |
12516 | exceptions, or run destructors when exceptions are thrown through them. | |
12517 | ||
12518 | You cannot mix Java and C++ exceptions in the same translation unit. It | |
12519 | is believed to be safe to throw a C++ exception from one file through | |
9c34dbbf ZW |
12520 | another file compiled for the Java exception model, or vice versa, but |
12521 | there may be bugs in this area. | |
1f730ff7 | 12522 | |
e6f3b89d PE |
12523 | @node Deprecated Features |
12524 | @section Deprecated Features | |
12525 | ||
12526 | In the past, the GNU C++ compiler was extended to experiment with new | |
767094dd | 12527 | features, at a time when the C++ language was still evolving. Now that |
e6f3b89d | 12528 | the C++ standard is complete, some of those features are superseded by |
767094dd JM |
12529 | superior alternatives. Using the old features might cause a warning in |
12530 | some cases that the feature will be dropped in the future. In other | |
e6f3b89d PE |
12531 | cases, the feature might be gone already. |
12532 | ||
12533 | While the list below is not exhaustive, it documents some of the options | |
12534 | that are now deprecated: | |
12535 | ||
12536 | @table @code | |
12537 | @item -fexternal-templates | |
12538 | @itemx -falt-external-templates | |
2dd76960 | 12539 | These are two of the many ways for G++ to implement template |
767094dd | 12540 | instantiation. @xref{Template Instantiation}. The C++ standard clearly |
e6f3b89d | 12541 | defines how template definitions have to be organized across |
2dd76960 | 12542 | implementation units. G++ has an implicit instantiation mechanism that |
e6f3b89d PE |
12543 | should work just fine for standard-conforming code. |
12544 | ||
12545 | @item -fstrict-prototype | |
12546 | @itemx -fno-strict-prototype | |
12547 | Previously it was possible to use an empty prototype parameter list to | |
12548 | indicate an unspecified number of parameters (like C), rather than no | |
767094dd | 12549 | parameters, as C++ demands. This feature has been removed, except where |
38bb2b65 | 12550 | it is required for backwards compatibility. @xref{Backwards Compatibility}. |
e6f3b89d PE |
12551 | @end table |
12552 | ||
ae209f28 NS |
12553 | G++ allows a virtual function returning @samp{void *} to be overridden |
12554 | by one returning a different pointer type. This extension to the | |
12555 | covariant return type rules is now deprecated and will be removed from a | |
12556 | future version. | |
12557 | ||
8ff24a79 MM |
12558 | The G++ minimum and maximum operators (@samp{<?} and @samp{>?}) and |
12559 | their compound forms (@samp{<?=}) and @samp{>?=}) have been deprecated | |
32e26ece GK |
12560 | and are now removed from G++. Code using these operators should be |
12561 | modified to use @code{std::min} and @code{std::max} instead. | |
8ff24a79 | 12562 | |
ad1a6d45 | 12563 | The named return value extension has been deprecated, and is now |
2dd76960 | 12564 | removed from G++. |
e6f3b89d | 12565 | |
82c18d5c | 12566 | The use of initializer lists with new expressions has been deprecated, |
2dd76960 | 12567 | and is now removed from G++. |
ad1a6d45 NS |
12568 | |
12569 | Floating and complex non-type template parameters have been deprecated, | |
2dd76960 | 12570 | and are now removed from G++. |
ad1a6d45 | 12571 | |
90ea7324 | 12572 | The implicit typename extension has been deprecated and is now |
2dd76960 | 12573 | removed from G++. |
90ea7324 | 12574 | |
1eaf20ec | 12575 | The use of default arguments in function pointers, function typedefs |
90ea7324 | 12576 | and other places where they are not permitted by the standard is |
2dd76960 | 12577 | deprecated and will be removed from a future version of G++. |
82c18d5c | 12578 | |
6871294a JW |
12579 | G++ allows floating-point literals to appear in integral constant expressions, |
12580 | e.g. @samp{ enum E @{ e = int(2.2 * 3.7) @} } | |
12581 | This extension is deprecated and will be removed from a future version. | |
12582 | ||
12583 | G++ allows static data members of const floating-point type to be declared | |
12584 | with an initializer in a class definition. The standard only allows | |
12585 | initializers for static members of const integral types and const | |
12586 | enumeration types so this extension has been deprecated and will be removed | |
12587 | from a future version. | |
12588 | ||
e6f3b89d PE |
12589 | @node Backwards Compatibility |
12590 | @section Backwards Compatibility | |
12591 | @cindex Backwards Compatibility | |
12592 | @cindex ARM [Annotated C++ Reference Manual] | |
12593 | ||
aee96fe9 | 12594 | Now that there is a definitive ISO standard C++, G++ has a specification |
767094dd | 12595 | to adhere to. The C++ language evolved over time, and features that |
e6f3b89d | 12596 | used to be acceptable in previous drafts of the standard, such as the ARM |
767094dd | 12597 | [Annotated C++ Reference Manual], are no longer accepted. In order to allow |
aee96fe9 | 12598 | compilation of C++ written to such drafts, G++ contains some backwards |
767094dd | 12599 | compatibilities. @emph{All such backwards compatibility features are |
aee96fe9 | 12600 | liable to disappear in future versions of G++.} They should be considered |
38bb2b65 | 12601 | deprecated. @xref{Deprecated Features}. |
e6f3b89d PE |
12602 | |
12603 | @table @code | |
12604 | @item For scope | |
12605 | If a variable is declared at for scope, it used to remain in scope until | |
12606 | the end of the scope which contained the for statement (rather than just | |
aee96fe9 | 12607 | within the for scope). G++ retains this, but issues a warning, if such a |
e6f3b89d PE |
12608 | variable is accessed outside the for scope. |
12609 | ||
ad1a6d45 | 12610 | @item Implicit C language |
630d3d5a | 12611 | Old C system header files did not contain an @code{extern "C" @{@dots{}@}} |
767094dd JM |
12612 | scope to set the language. On such systems, all header files are |
12613 | implicitly scoped inside a C language scope. Also, an empty prototype | |
e6f3b89d PE |
12614 | @code{()} will be treated as an unspecified number of arguments, rather |
12615 | than no arguments, as C++ demands. | |
12616 | @end table |